Why Dan Gilbert is wrong about Curbside Rail

[Detroitnerd]

Michigan avenue would be great. where exactly should it stop downtown. How far into dearborn and beyond should it go? What, if anything, has the current mayor of Dearborn said about the possibility?

Guido was favorable. I imagine O'Reilly can't be all that different.

[GUSHI]

Ya I see what your saying, but can't they increase the speeds to cut down on a hour trip after lets say 8 mile?

[jsmyers]

Just to reiterate, instead of small spur lines, which pose a lot of technical hurdles, we should be thinking about adding another radial line and then a crosstown line or two. Let's not get bogged down in linking little, special areas to M1. Let's think in terms of a comprehensive rapid transit system.

What are the technical hurdles you're worried about? I don't agree they exist, but I'm curious to know what specifically you're concerned about.

The biggest hurdle is money. LRT costs are a function of track miles and stations. A short 1-2 mile extension with 3 stops is a lot cheaper than a 6-10 mile extension that would presumably have a station about every mile. It is about 10 miles from Cass to West Dearborn on Michigan Ave.

While I don't at all disagree that LRT on Michigan as Far as Dearborn is a good idea. In a bang-for-buck consideration of what is going to provide the most trips and most positive impact to the city's economy and livability, I think that a complete system in the central city is the way to go, before [[not instead) of one long linear extension.

[Detroitnerd]

Ya I see what your saying, but can't they increase the speeds to cut down on a hour trip after lets say 8 mile?

Maybe, but it's always at the expense of something. Shorter trip times means fewer stops, for instance. You can only go as fast as possible when you're not slowing down for the next stop.

Maybe the main hurdle is that that kind of diffuse area isn't set up very well to take advantage of a mode that can take many, many passengers through a dense area. For instance, what are the current bus ridership statistics on Gratiot north of Eight Mile right now? That would be a good test for upgrading service to light rail out there.

[esp1986]

Interesting. Another factor would be the ease with which you can drive to the city center. In Detroit, it's pretty easy. Maybe not so much for the Windy City.

By the way, does the CTA have express lines? Or is it just all stops on all lines?

The CTA has s few lines that, for a period of time, follow the same route. Red, Purple and Brown. The Red Line is faster to a certain point, because it has fewer stops, then the Purple Line bypasses 12 Red Line stops, I guess this would qualify as the "Express" part of the line, but since it has a bunch of stops in the city center before it skips a bunch, I would hardly call it express. The Purple Line begins stopping again after the Red Line Terminus.

[Detroitnerd]

What are the technical hurdles you're worried about? I don't agree they exist, but I'm curious to know what specifically you're concerned about.

The system we're proposing is pretty straightforward. You get on a downtown LRT and it takes you downtown, to RPTS. You get on an uptown LRT and it takes you uptown, to State Fairgrounds. It is the M1 train.

But then you're running these spurs. Now you have to wait for the M1-A to pass, heading downtown, and get on a less-crowded M1-B train to get to Eastern Market. You need to reserve rolling stock for that trip, and you don't get the optimal ridership on that LRV. Then you're also running an M1-C to get to the casino, and M1-D to get to Jefferson East, and M1-E to get to Woodbridge. Before long, you're basically trying to run a light rail system as if it were a bus system. Low ridership per vehicle, long wait times for the proper vehicle, expensive track to maintain without the throughput to make it more cost-effective.

Now, I'm not a huge fan of buses, but this is what buses do well: The feed the main LRV line to increase ridership, and then they offer transfers off the LRV line to let the line do what it does best.

Better to let buses pick up the slack and keep our eyes on the prize: An effective, citywide rapid transit system built line by line by line.

[Detroitnerd]

The CTA has s few lines that, for a period of time, follow the same route. Red, Purple and Brown. The Red Line is faster to a certain point, because it has fewer stops, then the Purple Line bypasses 12 Red Line stops, I guess this would qualify as the "Express" part of the line, but since it has a bunch of stops in the city center before it skips a bunch, I would hardly call it express. The Purple Line begins stopping again after the Red Line Terminus.

Interesting. Thanks for sharing. At a certain point, I have to bring up an apples vs. oranges argument, because we've spent decades building downtown Detroit's parking capacity, whereas Chicago is so filled-in I've had to street-park a lot when there. Definitely made me think twice about driving downtown vs. taking the CTA.

I think the main problem may have more to do with how well diffuse, suburban areas of metro Detroit are set up to utilize light rail.

[esp1986]

Interesting. Thanks for sharing. At a certain point, I have to bring up an apples vs. oranges argument, because we've spent decades building downtown Detroit's parking capacity, whereas Chicago is so filled-in I've had to street-park a lot when there. Definitely made me think twice about driving downtown vs. taking the CTA.

I think the main problem may have more to do with how well diffuse, suburban areas of metro Detroit are set up to utilize light rail.

Agreed. I do not at all think it's too far fetched to mention 15 Mile / Maple in in Birmingham as a logical endpoint. Anything past that should be out of the question. All forms of density pretty much cease to exist at that point, and with Birmingham and Troy right together, there is certainly the potential for ridership there.

Woodward is, however, the only corridor that really warrants going this far. Into Macomb County, there is virtually no density past I-696, so not much reason to go very far there. Michigan Ave. hits a complete deadspot between Telegraph and Merriman. Grand River an argument could be made for going to Downtown Farmington, but not much further and East Jefferson would best be ended in the Gross Pointes. This is all provided the Woodward line even gets built, and there becomes consideration for a second line, third, etc. I would love to see a Michigan line become a continuation of the Woodward line though, much like many of the lines in Chicago do, where they come from one direction, into Downtown then out another direction.

Oh, to dream!

[jsmyers]

In terms of what to expect for speeds:

From GCP to McNichols, you should expect the vehicles to achieve about 30-35 mph when not stopped at a station or waiting for a signal at a major intersection. With stations, the average speed will probably be about 15-20 mph, faster where there are fewer stations and slower where there are more.

From McNichols north, speeds are more limited by the ability to protect the ROW and provide clear signals. Many LRT systems have top speeds of about 55 mph, with a few a bit faster. But every station stop adds about 3 minutes to the trip, between the deceleration, the dwell time, and the acceleration.

It is about 18 miles on Woodward from McNichols to downtown Pontiac. If a LRT vehicle is going 50 MPH the whole way, that is about a 22 minute trip. But if the train stops at every mile road, and each stop is a 3 minute penalty, the total trip time is 1:07, and a commuter is still at least 3.5 [[New Center) or as much as 7 miles [[financial district) from their office. The commute from Downtown Pontiac to Downtown Detroit on usual LRT would be approximately 1:30.

22 minutes + 3 minutes * 15 stops = 67 minutes to McNichols

+ 20 minutes to downtwown = 1:27

Maybe I'm being pessimistic about the stop penalty and expecting too many stops. Lets figure only 2/3 of the stops and a 2 minute time penalty:

22 minutes + 2 minutes * 10 stops = 42 minutes to McNichols

+20 minutes to downtown = 62 minutes.

To use another illustrative example: The MAX Light Right Blue line in Portland, OR is 32.7 miles end to end, with downtown Portland roughly in the middle. It average speed end to end is about 20 mph, for a total trip length end to end of ~1:40. this line is largely built on old railroad ROW outside urban centers. While I'm sure some people ride from end to end, the system excels and trips of half that length or less. Most people take the trains 15 miles or less to downtown Portland. For comparison, it is 17 miles on Woodward from Downtown to Birmingham and about 12 miles from Downtown Royal Oak to Downtown Detroit.

It is unreasonable to expect a lot of transit trip on across a 20-25 mile LRT system. That is a distance at which express, non-stop commuter buses and commuter trains excel. Now that doesn't mean the LRT shouldn't be installed over that whole distance, it would work well for a commute from Ferndale to Pontiac, or Birmingham to New Center. The vast numbers of possible overlapping destinations could make for an effective transit system at a more local scale. But the reality is that the largest attractors are in Downtown, New Center, and Midtown. There isn't much of that magnitude [[yet) close to Woodward north of New Center. If Southfield or Troy's office towers were walking distance from Woodward, things would be very different.

An alternative system, where a commuter train operates from Pontiac to Detroit and then on to Ann Arbor and keeps the same time as the current Amtrak schedule, a Pontiac to New Center commute is 40 minutes, and would still be under and hour to downtown including a transfer to LRT and a walk to the office.

[professorscott]

It does NOT take 3 minutes per stop. The deceleration and acceleration time are only several seconds each, and unless there are a huge number of people getting on, everyone can usually board in 10-15 seconds. Remember, unlike the buses, people won't be paying fare on the train so you won't have those ridiculous delays while people try to get their crumpled-up singles into the farebox.

I suspect the total delay caused by each stop will be much less than a minute. And to think the trip will take more than an hour is ridiculous; the DDOT Woodward bus makes essentially the same exact trip in 37 minutes at weekday peak hours.

[jsmyers]

But then you're running these spurs. Now you have to wait for the M1-A to pass, heading downtown, and get on a less-crowded M1-B train to get to Eastern Market. You need to reserve rolling stock for that trip, and you don't get the optimal ridership on that LRV. Then you're also running an M1-C to get to the casino, and M1-D to get to Jefferson East, and M1-E to get to Woodbridge. Before long, you're basically trying to run a light rail system as if it were a bus system. Low ridership per vehicle, long wait times for the proper vehicle, expensive track to maintain without the throughput to make it more cost-effective.

...

Better to let buses pick up the slack and keep our eyes on the prize: An effective, citywide rapid transit system built line by line by line.

I see you're point...and it is not at all trivial. I'll first point out that the things you bring up are still a concern when you have all of these lines fully built out.

But I don't think I'm quite explaining how it would work clearly enough. The attached schematic has lines that represent a train every specific period.



Lets say this period is every 15 minutes. What it would mean is that from Grand Blvd to Congress, there would be a train every 5 minutes. North of Grand Blvd, there would be a train on average every 7.5 minutes, but some waits would be 10 and some 5. The spurs have trains every 15 minutes. All of the trains go downtown. And all of the trains have the same rolling stock.

The only downside to this type of system is that if you wanted to run super low headways on the spurs, you run out of capacity at the core of the system. This issue is why Portland updated their transit mall a few years ago so that now there are two routes in downtown for LRT. The original route reached capacity.

But the limits are somewhere around a train every 2 minutes. The schematic I drew should allow a train every 6 minutes to Eastern Market, Corktown, and the Riverfront, with a train every 2 minutes in Midtown. It might even allow for tighter train spacing. A 2-car LRT train every 6 minutes has a capacity of 1,300 seated or 3,300 standing passengers per hour.

You could add other trains that didn't go on Woodward in downtown as well, if there was a need for capacity that didn't go to the exact core. For instance Royal Oak to Grand Blvd trains, or Corktown to Riverfront trains.

These aren't technical hurdles. If the system is widely successful, they lead to capacity constraints, but the counterpart of that is they make the system more useful so that it serves more people and more types of trips. It makes sense to plan for an expansion in the future to relieve the capacity constraint [[my idea is to also run trains on Gratiot//Grand River perpendicular to Woodward), but that is not a reason to avoid building a system that serves the entire central city.

[Detroitnerd]

I see you're point...and it is not at all trivial. I'll first point out that the things you bring up are still a concern when you have all of these lines fully built out.

But I don't think I'm quite explaining how it would work clearly enough. The attached schematic has lines that represent a train every specific period.



Lets say this period is every 15 minutes. What it would mean is that from Grand Blvd to Congress, there would be a train every 5 minutes. North of Grand Blvd, there would be a train on average every 7.5 minutes, but some waits would be 10 and some 5. The spurs have trains every 15 minutes. All of the trains go downtown. And all of the trains have the same rolling stock.

The only downside to this type of system is that if you wanted to run super low headways on the spurs, you run out of capacity at the core of the system. This issue is why Portland updated their transit mall a few years ago so that now there are two routes in downtown for LRT. The original route reached capacity.

But the limits are somewhere around a train every 2 minutes. The schematic I drew should allow a train every 6 minutes to Eastern Market, Corktown, and the Riverfront, with a train every 2 minutes in Midtown. It might even allow for tighter train spacing. A 2-car LRT train every 6 minutes has a capacity of 1,300 seated or 3,300 standing passengers per hour.

You could add other trains that didn't go on Woodward in downtown as well, if there was a need for capacity that didn't go to the exact core. For instance Royal Oak to Grand Blvd trains, or Corktown to Riverfront trains.

These aren't technical hurdles. If the system is widely successful, they lead to capacity constraints, but the counterpart of that is they make the system more useful so that it serves more people and more types of trips. It makes sense to plan for an expansion in the future to relieve the capacity constraint [[my idea is to also run trains on Gratiot//Grand River perpendicular to Woodward), but that is not a reason to avoid building a system that serves the entire central city.

You don't make a bad case for it. Thanks for clearing up what your design is. but I think that this plan is sort of pie-in-the-sky at this point. And you could accomplish much the same thing in the short term with feeder/spur buses and in the long term with more comprehensive lines and transfers. Plus, I must maintain it would be in all our best interests to develop systems that move people rapidly citywide, not just to certain neighborhoods/attractions.

[jsmyers]

It does NOT take 3 minutes per stop. The deceleration and acceleration time are only several seconds each, and unless there are a huge number of people getting on, everyone can usually board in 10-15 seconds. Remember, unlike the buses, people won't be paying fare on the train so you won't have those ridiculous delays while people try to get their crumpled-up singles into the farebox.

I suspect the total delay caused by each stop will be much less than a minute. And to think the trip will take more than an hour is ridiculous; the DDOT Woodward bus makes essentially the same exact trip in 37 minutes at weekday peak hours.

Does the DDOT Woodward bus go all the way to Pontiac? I'm not sure we are comparing apples to apples.

You are right that the time penalty is not 3 minutes when the train is going ~30 mph [[like it will on Woodward south of McNichols). I'm talking about the total time penalty for decelerating from 50 mph stopping, loading and unloading, and then accelerating again to 50 mph. Perhaps that is too long, and in retrospect, perhaps 2 minutes is a better, but as I wrote, Pontiac to Downtown would be about an hour if that is the assumption. In comparison, the Max Blue line is similar to LRT in a Suburban Woodward's median. That line has 50 stops in 30 miles and takes 1:40. LRT to Pointiac is about 25 miles and would have at fewest 25 stops, probably more like 30, and I'm pretty sure it would take around an hour to go end to end. That is about the time it takes the 445 and 475 Woodward Smart buses to go a slightly shorter distance at a higher operational cost with less comfort, reliability, and capacity.

[jsmyers]

I think that this plan is sort of pie-in-the-sky at this point.

Touché

And you could accomplish much the same thing in the short term with feeder/spur buses and in the long term with more comprehensive lines and transfers. Plus, I must maintain it would be in all our best interests to develop systems that move people rapidly citywide, not just to certain neighborhoods/attractions.

The way I'm looking at it is a matter of triage. It goes hand in hand with a plan for Detroit that focuses on making a livable and walkable city at the core, rather than trying to serve an entire city that was largely built on a suburban, auto-oriented model.

You are correct that a good cirulator bus system do a lot of the same stuff. And I don't disagree with you that the long term goal should be complete lines. I just think that getting the minimally effective operating start on each [[or most) of the lines is more important than extending them through what is largely suburban [[even if it is older and gritty than the suburbs). One of the beautiful things about how the compromise proposal is set up is that it lends itself to pretty simple piecemeal expansions that don't require the existing system to shut down or a huge city-wide construction project to happen.

People always talk and write about how Detroit is such a great city but it is not really connected together. The first phase of "M1" [[I hate that name.) makes a huge leap forward to correcting this. but I think that my ideas take it a good step further.

[Detroitnerd]

Well, let's not bog down the elegance of M1 by trying to make it into a mixture of crosstown and uptown-downtown, or by trying to make it do the buses' job too.

As for Detroit's density, the main thoroughfares are set up for it [[lot sizes, fire hydrant spacing, existing apartment buildings) because that's where the streetcar used to run.

I'm probably warier than most of the plans you propose because linking destinations that are built is nice but [[a) one of the best examples of that is the People Mover and [[b) one of light rail's most interesting abilities is that of attracting transit-oriented development where it goes in.

Anyway, thanks for the fresh ideas. I may not agree, but it's nice to have some people dreaming. [[And some news suggesting we may not be dreaming much longer.)

[jsmyers]

I didn't see this posted:

http://www.crainsdetroit.com/article...ard-light-rail

Bing lays out the facts pretty well. As a transportation planner, I'd be curious to know how much he understands everything he said, because it is obvious that the city's professional staff put together most of that statement [[as it should be).

[j to the jeremy]

I didn't see this posted:

http://www.crainsdetroit.com/article...ard-light-rail

Bing lays out the facts pretty well. As a transportation planner, I'd be curious to know how much he understands everything he said, because it is obvious that the city's professional staff put together most of that statement [[as it should be).

The points about bridge footings, viaduct height, and blocked driveways should be enough to convince M1 to settle for this compromise. The fact that it's completely economically impractical to put side running rail in near New Center is interesting.

[Gistok]

OK I have a question for those of you who think that the next line [[if even the 1st one is built)... should be out Michigan Ave.... Michigan Ave. is NOT as wide as Woodward or Gratiot [[which is very wide). If putting the rail lines down the middle of Woodward is the plan, then there will only be 2 lanes of traffic in each direction [[plus parking lanes).

If you use this same scenario on Michigan Ave., then much of that street will only be 1 lane of traffic in either direction. That does NOT sound feasable. Eliminate the parking lane, and it could help kill whatever retail is still there along the thoroughfare.

???

[jsmyers]

Good question. Looking at google earth:

Michigan Ave is 120' in Corktown transitioning from about 100' closer in downtown and 100' @ Livernois.

Gratoit is 115' in Eastern Market.

For comparison: Woodward is 120' in Midtown and about 100' north of Grand Blvd to McNichols.

A 100' ROW in Detroit is usually currently set up with 7 lanes for cars and sidewalks about 12' wide. This is what I measure a couple of blocks north of Grand Blvd on Woodward.

Parking is allowed in the outer 2 lanes, and the middle lane is for left turns.

In a 120' ROW, you get an additional travel lane in each direction. Travel lanes are 10-12' parking lanes are 9-11' and the left turn lane is about 10'

So currently they look like this when 120'

|12'|10'|11'|11'|11'|10'|11'|11'|11'|10'|12'|

And this when 100'

|12'|11'|11'|11'|10'|11'|11'|11'|12'|

Bold is center turn, italics is parking, and underlined is sidewalk. Everything is approximate, I don't work with the city of Detroit to measure everything or know standards.

Next post: how does LRT fit?

[Detroitnerd]

Michigan is, at many points, a pretty wide thoroughfare. Also, with light rail, there should be fewer cars and less call for parking. A rise in foot traffic and less burdensome parking requirements could actually help sustain and foster business growth, not hamper it.

Plus, I-94 pretty much duplicates Michigan's route through Detroit anyway.

Part of becoming serious about mass transit means also de-emphasizing automobiles. Don't panic!

[cramerro]

Gistok,

It may be that the portions where there are would only be 1 lane left will be forced to do shared lanes for the rail. Or, maybe the rail will hop over to run parallel with Amtrak through Dearborn, only a block+ away to the south. Otherwise you're pretty much stuck.

Sharing lanes isn't ideal because it really hampers speed, but if you include signal priority [[the light will turn green and stay that way longer than usual timing when a train is approaching) it can be mitigated. Much in Detroit and then again west of DB has plenty of room to take away a lane each way.

Another option might be eliminating parking on one side, or narrowing the sidewalk [[ I think it's already pretty narrow ). This is something they struggled with during the MI-Grand River plan in Lansing that I worked on... in order to keep two lanes each way they were going to have to give up potentially 100's of on-street parking spaces, so they were considering sidewalk, track alignment, etc. They also have a 1/4 mile stretch that is 66' wide ROW, with two lanes each way and sidewalks at the curb - that stretch had to be shown as shared lanes in all scenarios b/c you can't have a major state trunkline [[which MI Ave is) with only 1 lane each way. The Mi/GR Study [[http://www.migrtrans.org/) might be a good framework to see what type of factors would be in a Michigan Ave project - similar downtown on one end, smaller DT in the middle, suburban and urban cross-sections...

Would be interesting to go through that exercise through downtown Dearborn - let's hope that day comes soon!

[kraig]

Maybe, but it's always at the expense of something. Shorter trip times means fewer stops, for instance. You can only go as fast as possible when you're not slowing down for the next stop.

Maybe the main hurdle is that that kind of diffuse area isn't set up very well to take advantage of a mode that can take many, many passengers through a dense area. For instance, what are the current bus ridership statistics on Gratiot north of Eight Mile right now? That would be a good test for upgrading service to light rail out there.

That's a very good point. The best example of this I can think of is the People Mover stop at Bricktown, only to stop again right across the street at Greektown.

Hopefully, this thing will be able to expand to routes at Gratiot, Grand River, Michigan Avenue and Fort Street going from downtown. East-West along Eight Mile-Vernier, 16 mile and M-59 will be cool as well.

[Detroitnerd]

Re: Michigan Avenue:

I know I'm getting painfully close to pie in the sky here, but: There is one way to put in rail without losing lanes of traffic, but it's very expensive: A subway. I'm not saying put a subway in all the way, but maybe where there is no clearance you could go under and come up later. Just a [[fanciful, I know) thought.

[detmsp]

It does NOT take 3 minutes per stop. The deceleration and acceleration time are only several seconds each, and unless there are a huge number of people getting on, everyone can usually board in 10-15 seconds. Remember, unlike the buses, people won't be paying fare on the train so you won't have those ridiculous delays while people try to get their crumpled-up singles into the farebox.

I suspect the total delay caused by each stop will be much less than a minute. And to think the trip will take more than an hour is ridiculous; the DDOT Woodward bus makes essentially the same exact trip in 37 minutes at weekday peak hours.

Exactly. The lightrail in Minneapolis travels at 40 MPH outside downtown and arrives at a stop every 2-3 minutes. If stopping took 2-3 minutes, it wouldn't travel anywhere!

[jsmyers]

With center running LRT in a 120' ROW:

|12'|11'|11'|11'|30'|11'|11'|11'|12'|

and 100' ROW:

|12'|11'|11'|30'|11'|11'|12'|

-or-

|12'|11'|11'|30'|11'|11'|12'|

Except that now the bold is the LRT. Alternatively, you can cut out one of the 11' lanes and put a bike lane on both sides. The side with the extra vehicle lane can alternate as the LRT moves back and forth across the ROW [[or the ROW moves under the LRT).

So yes, where there is a 100' ROW, there will either be no parking or there will only be 1 travel lane. In practice, usually there is 1 through travel lane with parking, and then at intersections, dedicated turn lanes are provided [[but no parking).

Note that putting a station platform on one side takes out a travel lane for the length of the station. So at each station, one side looses a lane or parking. This is true in both a 120' and 100' ROW, but is more noticeable on the narrower street.

I'm familiar with the MAX Yellow Line [[Interstate), which I frequently used the summer after it opened. I think it is the perfect example of what LRT on Detroit's arterial streets would be like. Note that this neighborhood is some of the poorest parts of town, and the street used to hold the main highway between Portland and Seattle.

Please take a tour in aerial:

http://maps.google.com/maps?q=Portla...69922&t=k&z=20

and streetsview:

http://maps.google.com/maps?q=Portla...,345.73,,0,7.5

I can imagine it now "oh noes! the cars will be choked out!" In reality, the reduction in left turns in and out of the street help smooth the traffic flow a lot, keeping it safer and more consistent in velocity...but it is not a highway anymore...it is an urban street. Gratiot, Michigan, and Woodward were widened to bring all the cars into town before the interstates were built. They have way too many lanes now.