The GLR's (Green Line Revisited) proposal for service in the Somerville rail corridor is to add an Orange Line branch that feeds into its trunk line, starkly different from the GLX's seemingly simple proposal to use the corridor for extending the trunk of the Green Line.
A line/track which is operating at capacity could not add branches (service) without a concurrent reduction in service frequency to other branches, thus the ability of the Orange Line to adequately handle the service must be examined.
Getting Perspective
Branching is fundamental to creating an efficient urban rail system, but it requires more sophistication in allocating resources, scheduling and rail traffic coordination. Branching off of the Orange Line will require one of the following:
- increasing train frequency (by decreasing headway) on the OL's trunk line;
- decreasing the current train frequency on the Oak Grove Branch;
- or, some combination of the other two.
While it seems an obvious choice to reject decreasing the train frequency on the Oak Grove Branch, its current frequency should be weighed against demand and compared to the same factors experienced by other patrons within the transit system. In other words, reducing the frequency of trains on that branch is not off the table if doing so, in balance, provides an optimum amount of public service using the resources available.
Alternatively, investments necessary to allow drastic reductions in headway is not only an option for this application, it is mandatory if cost efficiency is to be achieved for the larger system. The cost of the infrastructure for each 'rapid' rail line is far too high to not make maximum use of its potential. Today, the MBTA's 'rapid' rail lines are among the slowest in the world. Technology is available to comfortably reduce headway to under 2 minutes.
First, we must ascertain whether or not there is any available capacity, at current headways, on the northern leg of the Orange Line.
The Numbers
Where is the current highest demand load on the Orange Line?
The MBTA already has plans to add 3 trains to the Orange Line, to alleviate over crowding; that indicator at first glance might doom any suggestion to reduce frequency on the north side. But, we must first analyze the overcrowding. If the service inadequacy is on the north side, diverting the additional trains to a new north side branch is counter productive.
The lack of hourly breakdowns on ridership prevents having a clear understanding of peak demand, but we can make some reasonable assumptions about the available data to approximate the information needed. We are provided with daily station entries, and will use that to assess relative demand on each half of the Orange Line for the AM (morning) peak period.
Outlying stations can be assumed to have virtually all of their daily entries in the AM period. On the Orange Line, the Mass Ave. and Back Bay stations are in mixed residential/business districts, thus only 50% of the daily entries for those two stations will be counted as AM entries.
Entries for the Community College Station on the northern half was not included in the AM totals due to the lack of a residential market around it. The station primarily serves the college which has no on campus housing and there are no MBTA buses routed to the station. The station would also likely continue to be served by all OL trains as the West Medford branch would divert north of the station.
There is a much larger interface with the MBTA's commuter service on the south end of the OL than on the north, which accounts for a large portion of the imbalance between the two sides.
The south end of the Orange Line appears to have almost a 50% greater load in the AM period than the north end, and a similar ratio can be presumed for the PM period. Overcrowding is chiefly a south side problem, while there exists some excess capacity available on the north side that could be used in serving another branch.
Will demand from a new branch over-load the northern Orange Line?
Using numbers provided by GLX officials, the touted ridership in the year 2030 is estimated to be 52,000 for a weekday (ridership is considered the total of the line's station entries and exits). Also specified is that start up ridership is predicted to be 90% of the 2030 figure, or 46,800. Here are the estimates for the initial demand, given in boardings by station (which in total equals to half of the daily ridership), and from these the demand for the OLB can be derived :
GLX OLB
West Medford - 2,035
College Ave 2,178 2,178
Ball Square 1,701 1,701
Lowell Street 1,134 1,134 |
Gilman Square 3,969 3,969 | -- these 3 stations are combined into 2 in the OLB
Brickbottom 2,457 2,457 |
Union Square 2,079 -
Lechmere 9,810 - (Lechmere continues as Green Line station)
23,328 13,474
Notes:
GLX materials do not mention that, as of 2009, Lechmere already experiences weekday station entries of 6,645.
The West Medford estimate is derived from the 2009 data for the three commuter stations being displaced by it; a partial corroboration for that figure is the MBTA's daily boarding estimate of 2,000 for a once planned Mystic Valley Parkway station.
The 13,474 for the OLB to West Medford, added to the 33,720 current entries on the northern leg, totals to 47,194. Thus, the West Medford branch will in fact roughly equalize the demand between the two legs, coming in just short of the 48,007 daily station entries on the southern leg.
 |
| OLB (Orange Line Branch) and a revised GLX |
How will the addition of a branch affect service frequency?
Note: The Green Line's trunk has the shortest headway in the MBTA's system, with an average of only 94 seconds (1.5 min.) during peak hours.
The MBTA has indicated its Orange Line can technically handle a headway of 4 minutes (perhaps lower), but has not had the equipment, i.e. coaches, to implement the increase in service. It has plans to add 3 trains (18 coaches) to the existing operation of 17 trains on the line, but will now wait until completing the total replacement of its fleet in 2015. We know then there currently exists the operational ability to safely handle a 20% reduction in headway on the Orange Line.
The total round trip time for an OL train is 85 minutes, which, at current peak period headway, allows for 17 trains. A reduction in headway to 4 minutes would allow for using 21 trains (85 / 4 min. interval). Thus, the MBTA should add, at least, 1 more train to its upcoming purchase.
The OL branching allocation will differ from the Red Line, which equalizes the frequency on its two southern branches. While the Red Line achieves similar service intervals for each branch, their ridership demand differs starkly. The Ashmont Branch has less than half of the demand of the Braintree branch, producing a great variance in the crowding conditions on each, and because of this the Red Line suffers frequent criticism from its riders.
The GLR proposes that the Orange Line allocate its trains to better achieve passenger comfort, with a nod to frequency. An apportionment based purely on demand only would allocate 15 out of the 21 trains to the Oak Grove Branch, but such an allocation would create too drastic a difference in service intervals. Therefore, 12 trains will be used for Oak Grove, allowing 9 to run on the West Medford Branch.
The ratio of train allocation will also produce a variance in frequency on each branch, thus while most of the Oak Grove trains will arrive at 8 minute and 4 minute intervals. The West Medford branch will have trains arriving at intervals of 12 minute and 8 minute intervals. This variance, even on a branch, is a concession in order to allocate capacity based on passenger demand.
The average peak period service interval for the Oak Grove branch would be 7 minutes and the West Medford branch 9.45 minutes. Again, compare these frequencies to the current Red Line branches which each have 9 minute intervals on all trains during peak period. For non-peak periods, service frequency on the two OL branches would likely be equalized as capacity and comfort factors are less a concern.
Again, keep in mind this scenario is without making signaling improvements that could drastically reduce headways, adding trains and thus doubling the available capacity. Such an enhancement could provide for having 21 trains (with fewer cars if need be) running on each branch during peak service at 4 minute intervals, with 2 minute intervals on the trunk line.
Can the apportioned supply of trains meet the demand of each Orange Line branch?
We've proposed a 13-train allocation to the Oak Grove branch, which when using the current capacity of OL coaches can provide an hourly passenger throughput of 6,681. That is calculated as 51 coaches/cars (8.5 trains) per hour each with 131 passengers (MBTA max capacity for OL cars):
85 min. RoundTrip / 12 trains = 7.08 min. avg. interval
60 min. / 7.08 min. = 8.5 trains x 6 cars each = 51 cars x 131 capacity = 6,681
The demand on the Oak Grove Branch, as calculated above, is 33,720 inbound passengers daily. If we estimate 80% of that figure is for the AM period, and 2/3 of that number occurs during a 3-hour period within the AM, the peak morning demand for that branch is 17,985.
A 3-hour capacity for the branch at peak service would be 20,043 (3 x 6,681), providing a buffer of 2,058. Sufficient, but not a roomy margin to rest on, thus we are fortunate that the OL gives us options for increasing capacity, both in the short term and long term.
The total daily inbound demand estimate for the West Medford Branch is 13,474 (see above), and the capacity available per hour for the branch would be roughly 6.35 trains x 6 cars = 33 x 131 = 4,991. Using the same AM percentage of 80% as before, and figuring a 2/3 allotment to the 3-hour peak period, produces a morning peak demand of 7,187 (13,747 x .80 x .6667). That is less than half of the available capacity for that period (3 x 4,991 = 14,973).
Capacity on the Orange Line can be improved, if necessary, without adding trains or coaches.
The coaches/cars provided on each of the MBTA's lines not only vary from line to line, but frequently within a line. Seating and standing capacities are quite inconsistent, not only in quantity but also in proportion to the size of each coach.
For 100 years engineers and mechanics have customized Boston's transit vehicles to adjust for conditions, going so far as to strip some coaches of all their seats. Though this method is being used on the Red Line today during the most congested hours, nothing that drastic needs to be done for increasing capacity on Orange Line coaches.
The option to adjust the Orange Line's 'crunch' capacity is a reasonable one, given that the Orange Line currently has the highest seating to standing ratio among all of its city rail lines. The percentage of floor space used for seating varies on all three heavy rail lines; the Blue is 8%, Red 9% and Orange 10%.
The removal of only a few double seat benches increases standing room significantly. If this sounds unappealing, note this comparison: Green Line cars have, on average, 13 fewer seats than Orange Line cars.
Thus, the 131 passenger capacity per OL car used above could easily be increased, while maintaining an equal or better seating option for its riders compared to the other lines.
Compare this proposed service with that provided on the Red Line branches.
I think it is safe to state that if a West Medford branch is implemented off of the Orange Line's northern trunk, service levels equal to (or above) those provided on the Red Line can be attained with an addition of only 1 train , beyond the 3 already planned to be added, with or without minor coach modifications.
An explanation of calculations to produce a capacity per mile of service line comparison are too complicated for this overview. The numbers do show
Here, the Red Line is an important reference for the GLRevisited's proposal to branch the Orange Line, as the Red Line is currently the MBTA's only branched 'heavy' urban rail line. Thus, we can turn to service levels experienced by riders on the Red Line's south side for guidance.
The functional length of the new branch is 4.6 miles, from Community College Station to West Medford Station. That also equals the length from CC to Oak Grove, thus a run from Forest Hills to either branch will be 11.16 miles. The average length of the Red Line is 14.675 miles. Running 26 trains, it has 1.77 trains per mile of operation, using a 4.5 minute headway. Applying the 1.77 rate to the Orange Line gives 20 trains, thus with the upcoming acquisition the two lines will have equivalent capacity per mile of operation.
While the RL runs a peak time 4.5 minute headway on its trunk/mainline, the branches each have 9 minute headways. The headway, 6 car trains and average crush capacity (Red Line coach avg. of 270), gives an hourly load capacity of 10,800 (60 min./ 9 min. = 6.667 trains per hour; 6.667 x 6 (coaches) x 270 = 10,800). Operating at crush capacities is, of course, not desirable, but its use here is only relative (not quantitative) for comparing resource allocation.
Though the demand on the Braintree Branch is more than twice that of the Ashmont Branch, the MBTA provides the same amount of capacity to each in order to equalize the frequency of service. Thus, half of the trains go to Braintree and half to Ashmont, with each branch having a 9 minute service interval (headway). Equity might dictate that comfort, particularly on the much longer trips of the Braintree branch, would have the MBTA send more trains to the branch with the greater demand. It of course is a trade off, of time and comfort which has to be resolved.
The inbound demand is reflected in the weekday entries as noted on the top map, with the Ashmont Branch having 12,276 and the Braintree Branch 27,005. The Braintree Branch with the highest demand of the two and yet the same level of supply as Ashmont reflects the baseline which we must meet, or better, in providing service to a branched Orange Line. Again, for comparative purposes, we'll use the four hour AM rush period for our time factor and the total daily entries as the demand factor. The supply/demand ratio for the Braintree Branch is thus 43,200 (4 x 10,800)/27,005, or 1.60.
Can a similar rate be attained when branching the Orange Line, within current system constraints and making expenditures significantly less than those of the GLX?
A View of the GLX
How does the service capabilities and impacts of the GLR compare with that of the GLX? As already noted it is easier for a line to extend rather than branch. That understanding should be applied to the GLX, just as it has been to the GLR proposal for branching the Orange Line.
Lost in arguments to extend the Green Line as a primary load carrier from Somerville into Boston is that it is already a heavily branched line, albeit on the other end. In fact, the current official headway on the main trunk of the Green Line from Government Center to Copley Square is a mere 94 seconds, on average. Some evidence indicates that service on two of the Green Line branches actually runs at 5 minute headways during peak hours, and that being the case, the 'average' headway on the trunk is then 82 seconds.
Fortunately, extending the Green Line won't entail adding vehicle traffic to its already congested trunk line, but the restriction will, most likely, prevent increases in service via an increase in frequency. This is a significant concern as the demand estimates for the extension are just that, estimates. And, while the questionable demand issue holds true regardless of which method is used to provide the service, there are significant differences between how the two methods, GLX and GLR, could handle higher demand levels than those forecast, or for future growth.
The GLX proposes extending the routes of the 'D' and 'E' lines to College Street and Union Square, respectfully, both of which operate with 'official' 6 minute, peak time, headways on their branches. Thus, with 10 trains per hour, operating the typical 2 car train, and a crush capacity per coach of 234 (using Green Line tram average), the hourly load capacity during peak hours is 4680 (10 x 2 x 234). Over a 4 hour period that provides a supply factor of 18,720.
If we look at the 'D' line which will run out to College Ave., it will serve all but one of the listed GLX stations, thus we'll eliminate the Union Square demand and, since the 'E' line will also serve Lechmere with the same frequency, we can reduce the Lechmere demand by half. These adjustments give an inbound demand of 16,416. The supply demand ratio is then 18,720/16,416, 1.14.
A supply/demand ratio of 1.14 is only 71% of the 1.60 we calculated for the Braintree Branch, which we've used as our minimum guideline in judging the GLR proposal. We can work around the demand estimate of 16,416 by having the Lechmere stop served only by the shorter 'E' line; this reduces the demand on this end of the 'D' to give 11,511. Now, the ratio improves to 1.62. This is roughly equal to the GLR's rate for both branches.
A significant flaw though in the calculation is the crush capacity of 234. The bulk of the Green Line's current fleet are Breda coaches which are also its newest, for which 199 is the Breda 'official' crush capacity. Moreover, their crush capacity uses 1.3sq.ft./person as its basis, compared to a rate of 1.5sq.ft./person for all other coaches, on any line. The comparison crush capacity should be 172 (199 x (1.3/1.5)). The adjusted hourly supply is 3440 (10 x 2 x 172). The adjusted supply/demand for the 4 hour AM would be (4 x 3440)/11511, a 1.20, and recall that is without stopping at Lechmere
As the subject of seat availability was raised in the latter part of the GLR's capacity findings, both quantitatively and on a per mile basis, the hourly offering of seats on the GLX with the 'D' line to College Ave. would be 20 (coaches) x 45 (seats) = 900. The GLR would hourly offer (a little more than) 6 trains with 6 coaches each, 36 x 48 (seats, originally 58) = 1728. The per mile rates are, Green 'D' Line, 900/5.07(from Park St.) = 178/hour/mile,
Orange Line, 1728/6.15(from DTX) = 281/hour/mile.
Under the GLR's proposal, the two branches of the Orange Line would attain rates above 1.60, well higher than the 1.20 of the GLX's light rail extension for its College Ave. branch. Beyond this disparity, there is a larger, more critical gap.
Built-in Obsolescence?
If the numbers and analysis so far have not adequately made the case that the GLX's plan to operate an extension of the Green 'D' (light rail) Line may well struggle to meet current demand estimates, or that this site's, Green Line Revisited, OLB proposal to branch the Orange Line is capable of providing adequate service to meet a similar demand in the same market area, it should be deduced from the data, already provided, that the ability to increase service in the future, or to immediately adjust to greater than expected demand on start up, is starkly limited for the Green Line service.
It is difficult to make exacting comparisons of one route/line to another. The travel habits particular to patrons of each line and the varying ratios of destination points to origination points from line to line make each unique. One common characteristic though is that some branch lines or major portions tend to be wholly collectors in the AM commute time, serving predominantly residential areas; this would be more true of the Green 'D' Line (Riverside) than the other Green Line routes. And it is likely to be true of most all of the GLX's route to College Ave. The Boston Region MPO's report on transit crowding for the MBTA's urban rail lines reveals that the Green 'D' Line has the highest rate of crowding.
 |
| Population Density |
If the demand on the new 'D' line route to the north only equals its south end, it will begin at near max capacity. This raises the question of what ability does the MBTA have to increase capacity on 'D' line service to College Ave., if demand exceeds estimates?
The options to increase capacity are limited for the GLX proposal. With its branches operating at 6 minute (or 5 minute) headways, the trunk/mainline may well be at traffic capacity. It could opt to run more trains on the branch portion only, returning them from either North or Haymarket Stations. Another option would be to operate with 3 coach/car trains, if the mainline stations can handle the longer trains.
The MBTA has recently begun operating 3-car trains on the current 'D' Line to alleviate over crowding during peak hours. In all likelihood it will do the same on the GLX portion of the 'D' Line to College Ave. That will increase capacity by 50% during the times the longer trains are run. That will well satisfy current demand, but if demand is 50% greater than the planners have estimated it will likely occur proportionately, meaning that peak hour demand will be 50% greater as well. Will the MBTA then add a fourth car to that line's trains during peak use time? If so, the train will then be longer than a Blue Line train.
Closing Comment
This last hypothetical, hopefully, spurs a realization of just what is going on. We have a round hole to fill, and with very poor judgment have decided to fill it with a square peg. Having started with a square peg, we have whittled away at it until it almost fits. If we keep at it, we will, through much effort and at an outlandish expense, have converted it to a round peg; we will remain clueless as to the waste while the politicians gloat on the accomplishment; and government contractors profit immensely.
We are an embarrassment to our forefathers, whose value judgments on public expenditures mirrored the same care exhibited in their private transactions; today, no such connection exists. Today's political scene has neither good judgment or accountability.
In the case of the GLX, we have taken 'light' rail, meant to be integrated into the life blood of our city, taken it off the street and placed it farther away, we eliminated stops, we built walls around it to control access, we make it go a little faster, we make the cars longer to carry more people. So, now we have a heavy 'light' rail system that cost more than a heavy rail system would have cost, and, too, we are faced with the task of running heavy rail and light rail on the same tracks.
Let's use round pegs for round holes, and square pegs for square holes, and don't remain quiet while the average intellect which dominates our political system 'spins' and 'spends' us into the poor house.
