January 21, 2013

Red Line FEIS - Part 1

Red Line FEIS "Travel Forecast Results" report: Utter failure

Astonishingly, the MTA's gargantuan Red Line Final Environmental Impact Statement (FEIS) provides virtually no justification for the multi-billion dollar project - despite or perhaps because of being based on almost totally revamped population and ridership data from what was used to reject all other alternatives.

The proposed MTA Red Line is even slower than comparable buses. It cannibalizes the existing bus and rail transit system in an extremely cynical ploy to prop up Red Line ridership projections by exploiting poor captive bus riders. It restructures the MTA's already inefficient route system in an even more chaotic direction. And it is based on some assertions that are just too bizarre to believe.


There is far too much in the Red Line FEIS to cover everything, but the "Travel Forecasts Results Report" demonstrates its utter failure. Here in three parts, addressing the data assumptions and their impacts, are the FEIS comments I will submit to the MTA. Comments are due by January 28th emailed to FEIS@baltimoreredline.com with the subject heading "FEIS COMMENT".

PART ONE - TRAVEL FORECASTS DATA



Here is the problem in microcosm: The #150 bus line currently runs the length of US Route 40 from Howard County to Downtown. East of Edmondson Village, it's route resembles that of the proposed Red Line. It takes 11 minutes for the #150 bus to go from Edmondson Village to downtown at Greene/Saratoga. But the Red Line is projected to take 16 minutes for a similar trip to Howard/Lombard, five minutes more than the bus. The MTA wanted to provide a comparable Red Line station at Greene Street, which would save about a minute, but they had to eliminate it to save money.
Proposed future Edmondson Village Red Line Station hypes up the estimated 16 minute travel time to downtown, but the existing #150 bus line can do it in 11 minutes.

Under the MTA Red Line plan, however, a trip downtown originating on the #150 bus would become far lengthier than 11 minutes, by deviating off US 40 route onto Ingleside Avenue to reach the I-70 Red Line station, forcing bus passengers to transfer to rail, and thus slowing down the trip in order to include it in the Red Line's ridership numbers.

But messing with the efficient but obscure #150 bus route is just the tip of the iceberg. A large proportion of the MTA's indecipherable urban bus network would be even more mangled to feed the Red Line, even passing up far more logical connections to the existing subway and stealing bus service from non-feeder services that have little to do with the Red Line. And much of this is proposed for low income transit-dependent areas.

Phantasmagorical population projections

The foundation for all this is the revised 2035 population projections used for the Red Line, provided by the Baltimore Metropolitan Council. The key stat is highlighted on page 7 of the Travel Forecasts Results Report: "More than half of the region’s population growth is expected to occur in the Red Line Corridor for an increase of approximately 45,000 residents."

The table which follows shows the specific share of the region's population growth as 62%. From 2005 to 2035, almost two-thirds of all the population growth forecast for the entire Baltimore urban, suburban and exurban region, including the city and five surrounding counties, is presumed to occur in just a slender two-mile wide strip of land through the city and western Baltimore County. And much of this strip is water, large cemeteries and Leakin Park (FEIS Travel Forecasts Results Report, Figure 3).

Think about this. The Red Line plan assumes that the vast majority of regional population growth, starting eight years ago in 2005, would not occur in Howard and Anne Arundel Counties being fed by the far faster growing Washington region. It would not occur in Harford County which is booming from military base relocations. It would not occur in Carroll County where new subdivisions seem to sprout up monthly. And it would not even occur in urban/suburban "smart growth" districts like Charles Village, Westport, Owings Mills, Towson, Hunt Valley, and White Marsh.

The narrow Red Line corridor would get far more than all the rest of the region combined - from Pennsylvania to the DC suburbs to the Chesapeake Bay.

The Red Line's population growth projections and tortuous feeder bus plan combine to define its eye-popping daily ridership estimates of 50,000 in 2030 and 54,000 in 2035.

Breaking down the ridership

But the ridership numbers become even stranger when examined in more detail. For example, the Red Line is projected to carry nearly 3000 more daily riders eastbound than westbound (Table 23). Yes, sometimes transit riders take a different way home, but why would eastbound riders be so much more inclined to do this than westbound riders?

And the aggregate totals mask even bigger discrepancies on a station-by-station basis. At the east end Bayview MARC station, 2923 riders would get on throughout the day but only 504 riders would get off, almost a six-fold difference. At the Canton Crossing station, 5945 westbound riders would get on, but only 1906 eastbound riders would return, less than as third as many. These makes even less sense because they are major parking lot stations. What is to become of all those parked cars that the riders leave behind?

Practically every station is like this. Perhaps the most absurd example is that the ridership from the Bayview MARC station to the Bayview medical campus is given as 277, but the return volume is zero, nada, zilch.

These are examples of what is called the "black box" effect of computer models in a BaltimoreBrew.com comment by Jamie Kendrick, who has worked on the Red Line for both the MTA and the city. He was referring to the cost-benefits model, but that is largely based on the ridership model. Reality checks are supposed to be done during the model calibration process, but of course, the whole purpose of computer models is to forecast the future, not reality.

In Table 23, the severe on/off discrepancies are resolved by computing averages of the disparate numbers. Thus at Bayview MARC, the 2923 on-folks and 504 off-folks are each counted as half a person, split to yield what are inaccurately described as 1710 "total boardings".

Other parts of the analysis handle it differently. The next graphic (Figure 11) calls the same data "production / attraction", which is transpo planning lingo for defining trips by how they are generated rather than whether they are coming or going, but the volumes are the same.

Then comes the station-by-station ridership by access mode (Table 24). Only the original boarding data are presented, not the very different numbers for folks getting off the train or those odd split averages. Some of the stranger numbers in Table 24 can only be explained from the equally strange feeder bus operations plan.


The Red Line ridership numbers are propped up by numerous proposed convoluted changes to Baltimore's already confusing bus system, which would exacerbate its already fundamentally flawed structure that is the root cause of the MTA's chronic reliability problems.

One of the strangest Red Line stations for these Table 24 data is Rosemont, on Edmondson Avenue near Poplar Grove and Franklin Streets. Incredibly, throughout the entire day, only 36 people in this very dense mostly black working class community are expected to walk to the Red Line. That's only about two an hour. In contrast, 3368 riders will arrive at the station by bus, 98% of the total. This includes six bus routes, the #10, #15, #16, #23, #38 and #47 Quick Bus. All those 3368 transferring riders would require a lot of buses, but the plan calls only for standard on-street curb bus stops.
Proposed Rosemont Station location in the middle of Edmondson Avenue, with projected ridership of
3368 per day transferring from buses and incredibly only 36 walk-ins

Four of those six bus lines - the #16, #23, #38 and #47 - would stay about the same in the Rosemont area, although the #23 would have major changes at the ends, westward in Catonsville and eastward from Bayview. The #47 (the Quick Bus version of the #15) would remain intact even though the #40 (the Quick Bus version of the #20 and #23) would be scrapped through West Baltimore. So #47 buses would be allowed to continue to whizz past Red Line trains along the "Highway to Nowhere", even though #40 and #150 buses would not.

The biggest messes would occur with the #10 and #15 lines, which would be split into branches off the existing routes which would terminate at the Rosemont station instead of going all the way downtown as they do now. Most bus riders would be confronted with a dilemma as to which branch of the watered down service they should take - the slow but more direct bus route or the more convoluted branch which requires an extra transfer onto the Red Line, and perhaps another transfer off the Red Line, and so is very likely even slower. Riders are screwed either way.

Some of the remaining #10 buses would branch off the existing route after serving Catonsville and Yale Heights in the Frederick Avenue corridor, cutting through some narrow residential streets to feed the Red Line, while other #10s would stay on the existing route to downtown and the Eastern Avenue corridor to Fells Point, Highlandtown, Bayview and Dundalk.

The #15 route would stay essentially the same from Security Mall to Forest Park to Rosemont to Poppleton to Downtown. But some of the buses would terminate at the Rosemont Red Line station, forcing the riders to transfer. And starting downtown, other buses would deviate from the current route, which proceeds out to the Gay Street/Belair Road corridor to Overlea and Perry Hall. The deviating buses would proceed eastward in the Baltimore/Fayette Street corridor to Bayview.

Bus route system structural problems

There are numerous fundamental problems with this proposed route structure. Basically, there is very little hierarchy or organization. Most of the bus routes would remain very long and unwieldy, requiring riders to slog though many miles of slow local Baltimore streets as they do now. Too many bus routes would continue to run parallel to the Red Line. To a distressing extent, bus routes and the Red Line would compete with each other since neither mode would clearly provide better service.

As part of this precarious balancing act, even some bus lines that would seem to have very little to do with the Red Line appear to get service cuts. The Bus Operations Plan Table 5 presents an inaccurate compilation of existing bus service headways (frequencies) which hides many of the changes.

For example, midday service on the crosstown #22 and #51 lines, arcing across the city from Bayview to Mondawmin to Cherry Hill, currently runs every about 20 minutes, but is presented in Table 5 as every 50 minutes (#22) and 40 minutes (#51). These are then shown as the "feeder bus" headways in Table 6, masking an apparent radical cut in service. The MTA seems to think that since these routes serve captive passive low income riders, no one will notice.

The travel demand has been rigged to assign a huge number of these wide-ranging trips to the Red Line, but many riders would not benefit from that. Moreover, riders don't like to make transfers unless there is a very good and clear reason to do so. In the MTA fare structure, riders without a pass are actually charged two full fares for the privilege of transferring, adding injury to insult.

Subway and kiss 'n' ride transfers

"Kiss and ride", planners' cheeky term for spouses, taxi drivers and other people who drop off other people at stations, has even more pronounced ridership differences from station to station. The Security Square Mall station is one of the more amorous, with 518 daily kiss 'n' riders. In contrast, the Edmondson Village station, whose kiss 'n' ridershed would include Ten Hills, Hunting Ridge and most of Catonsville, yields only two kissers per day. Is it a demographic distinction?

But the strangest numbers of all might be the Inner Harbor central downtown station at Lombard and Light Streets, which is supposed to provide access to the existing Metro subway though a proposed two block long pedestrian tunnel. Not surprisingly, it would have the highest number of boardings on the entire Red Line, a huge 9010 per day. But unbelievably, less than 20 percent (1742) would be local downtown walk-in riders.

The lion's share of the Red Line boardings at this station, over two-thirds (6062), would be subway transfers from the Charles Center Metro station to the Red Line. As a comparison, the current total ridership at this Metro station is about 6500 boardings (and 6500 de-boardings) per day. The Metro's ridership numbers have stayed essentially flat since its last expansion back in the mid '90s. The Metro to Red Line transfer through the two-block long catacomb passageway would be nearly as much as all the current Metro riders. Whew!

The MTA has long argued that a direct station connection is unnecessary between the Red Line and the Metro, even though virtually every other modern rail transit system has at least one. But with numbers like this, they're only kidding themselves.

The travel zone matrix

The FEIS trip matrix, the core of the ridership data from one place to another, is more revealing for what it does not say than for what it says. 

The FEIS does not provide a station-to-station matrix of rider origins and destinations, so it is impossible to do much sleuthing of all these eyebrow-raising numerical claims. The only substitute is a matrix of travel zones throughout the city and five-county suburban region, showing the number of Red Line riders that would begin or end at each (Table 21). Thus we know that Howard County would provide 2231 trip "productions" (mostly by people who commute from there) and 1182 "attractions" (mostly jobs). Some of these would be existing #150 bus riders diverted to the I-70 Red Line station, but this station is supposed to have 1811 transfers daily from bus to rail, so a lot more bus riders would be needed. And since the #150 service is the only bus line that would serve Howard County, and it would be much worse than it is now, the vast majority of these riders would have to come from elsewhere.

The report provides only eight travel zones in the Red Line corridor itself. With nineteen stations, that means the zones (Figure 4) are far too large and aggregated to be useful in modelling travel patterns, sometimes extending beyond what is defined as the "study area" (Figure 3). The four most central zones extend all the way up to North Avenue or beyond. Harbor East and North/Gay, two of the most disparate communities imaginable, are lumped together in the same zone. On the other hand, many Red Line stations straddle two zones - Highlandtown, Canton Crossing, Fells Point, Poppleton, West MARC, and I-70 - so it's impossible to tell who will come and go from where.

These ungainly travel zones are also used to compute rider "benefits", in terms of travel time savings accrued to transit riders. A very simple-minded summary of travel time benefits is presented in Table 18. The Red Line takes 36 minutes IVT ("in vehicle time") to go from Bayview to Social Security, while the "low cost" bus alternative takes 66 minutes. But the current #40 Quick Bus takes only 53 minutes to go even farther to CMS (which takes the Red Line 45 minutes). But somehow, this table inflates the bus vs. rail difference for in-plus-out of vehicle travel times to 84 versus 48 minutes. 

This is allegedly explained by greater congestion in 2035, but that is a relatively small factor and is largely caused by the Red Line anyway due to narrowing Edmondson Avenue and the city's refusal of confront traffic demand caused by the rampant proliferation of parking garages. Section 5.1.4 of the "Uncertainty Analysis" summarizes this by saying "there is little impact on ridership based on 2005 and 2035 bus speeds." This flies in the face of the assumed increase in the bus vs. rail travel time advantage from today's 53 vs. 45 minutes "in vehicle time" to a future 84 vs. 48 minutes total time. That makes no sense.

No wonder the model is described as a "black box". This leaves lots of FEIS wiggle room for the ridership forecasts. Data chopped up into much smaller zones is available, but there is a trade-off of confusion from too much to too little detail. They opted for less detail. As a result, we should gratefully be less confused about how much confusion there is. Capice?

Bottom transit line

The MTA is obviously willing to move heaven and earth to make their Red Line numbers look as good as possible, even though it requires making the rest of their transit system even worse than it already is. The MTA never examined other alternatives in the same way. And even with their inflated ridership claims, their plan still does not come even close to meeting the downplayed federal cost-effectiveness standard.

The MTA's ongoing track record of rail transit failure is perhaps best seen in the system's stagnant ridership and its almost complete lack of Transit Oriented Development, despite such hyped-up project areas as Howard Street, State Center, Westport and Owings Mills. The initial 8 mile Metro project promised a first year ridership of over 80,000 per day, but three decades later, it still hasn't consistently risen above 50,000 riders per day despite two extensions to 15.5 miles. All indications are that the Red Line would perform even worse.

Baltimore needs to get off this Red Line train. Continue to Part 2 of this critique.

Thanks to Edward Cohen for important contributions to this analysis.

No comments:

Post a Comment