Sparked in part by a suggestion from a reader, I spent two days last month in Karlsruhe, a pleasant but not touristed small city in the southwest of Germany. In rail transit circles, Karlsruhe is famous for inventing “tram-trains,” a vehicle and service type that can operate in the street as a streetcar/tram, but can also go onto standard railway lines, often shared with intercity passenger rail and freight, to go longer distances into the surrounding suburbs. This means that the service you board in your outer suburb can flow right into the city’s core streetcar network, and thus get you closer to your inner-city destination than the train station would be, thus saving you from having to make a connection.
This may seem like not much of an invention, but rail engineers will tell you it’s harder than it looks, especially because the crash-worthiness standards that you need to meet to share tracks with freight tend to conflict with the relative lightness than you want in a streetcar.
(Language note: "Streetcar" and "tram" mean the same thing, but since we're in Europe here I'm going to tend to use the European word "tram." Don't get confused.)
But like many great transit ideas, the innovation is hard to photograph. All I can show you is a fairly ordinary-looking light rail vehicle running as a tram through a pedestrian precinct…
… and then running in a commuter rail mode at a suburban station.
Because Karlsruhe's model is being exported to other cities, however, the best service I can offer here is notice how few cities the idea would really apply to, and notice what's been sacrificed to achieve the tram-trains' benefit.
Karlsruhe is a small city, about 300,000, with a very strong single downtown (including a major university) and not much to compete with it elsewhere in the city. The geographical problem that drove the tram-train invention is that this urban core is about 2 km from the main rail station. This separation is a feature you’ll find only in pre-rail cities (developed before 1850 or so) or some poorly planned late 20th century ones. In the main era of rail development, roughly 1850-1950, towns naturally grew up adjacent to railroad lines, and while the rail line itself tended to be an urban barrier, you typically ended up with a downtown that was adjacent to the rail station on at least one side of the tracks.
If the city was there before the rail line, and the rail line couldn’t get closer without disrupting other travel markets, then compromises often had to be made between a station site ideal for the town and one that was efficient for the through operations of the rail service. (The same tradeoffs are made today, at many scales, whenever land use and rapid transit interact in a new town or transit-oriented development setting.)
So with that background, here’s a bit of
It’s very, very complicated, in part because there’s no effort to sort the system into a hierarchy that would show shich services are frequent. But the thing to notice is that most lines converge on the rail station (“Karlsruhe Hbf” in the lower left) and again on the downtown (whose focal point is “Karlsruhe Marktplatz” in the top-center of the image). Many of these lines run in a frequent-stop streetcar mode through the core and then continue right out into the exurbia in the regional rail network. (Lines that do this have numbers starting with S.)
When exploring a system that’s new to me, I often take an intentionally random trip. I just get on whatever vehicle’s coming, and see where it goes and how it seems to work. Here’s the trip I ended up taking, and the travel time (in minutes) for each segment. Again, the rail station is in the lower left. The "tram" segments are from the rail station up to "Downtown/Marktplatz" and then east to a point just before Durlach station -- evident from the longer travel times.
At the rail station (lower left) I happened to get on an S4, the dark brown line on the network diagram. First,
the S4 goes north as a tram, covering the 2km into the center in 8 minutes. Then, it
turns southeast and goes, still as a tram, about 3.5 km to Durlach station in 11 minutes. The tram segments are in exclusive right-of-way but still very slow because it operates through so much shared pedestrian space.
But once you get to Durlach
station (“KA-Durlach Bf” on the network diagram) the vehicle slides onto the rail network and speeds up, so it took only 7 min to cover the 6 km to Berghausen-Hummelberg, where I
got off. It's a minor stop with not much more than a pad and a shelter (second photo above) serving low-density housing and open space.
On alighting, I noticed that I was on a single-track segment. Bad news: that usually means low frequency. Sure enough, the regular pattern of service back into the city was at :35 and :55 after the hour. That’s uneven spacing, which is bad news too. Two trains an hour should mean a maximum wait of 30 minutes, but here the maximum wait is 40. I can’t think of many really good excuses for this. Even though this line must merge with many others on a downtown segment, and must thus find a space in a more crowded hourly cycle, it’s almost always possible to do this while maintaining regular spacing. The likely explanation is that the schedule has to work around some kind of less-frequent intercity train movement sharing the same single track. Such compromises are sometimes necessary, but it’s worth noticing what they cost for the customer.
Returning on the same line, I wanted to get back to the railway station, but that’s not what the S4 does. It turns and goes through the city first. What if I lived on the outer reaches of the S4 and just wanted to get to the
Enough about my troubles. What can we conclude about tram-trains and their value?
The crucial necessity that drove the tram-trains’ invention is that Karlsruhe is (a) a small city with (b) a fairly compact downtown that is (c) the primary destination for a large share of trips but is (d) a little too far from the main rail station where normal commuter rail trains would logically terminate. Generalizing a little, I'd suggest that tram-trains make sense if:
- You're in a small market where the scale of demand doesn't justify, say, running suburban commuter rail lines into a subway under the city, as large cities (e.g. Paris, Sydney) sometimes do.
- The destination area you want to serve (the core downtown in Karlsruhe's case) isn't right around the rail station (or candidate rail station site) where a normal commuter rail service would cover it.
- This destination area is so dominant that you don't need to be too concerned about anyone wanting to go somewhere beyond that area.
Think about it this way. Imagine you’re on the “train” portion of the service, cruising along stopping just once every km or so, like a typical rapid transit service. Because of the relatively high speed, this service can be useful to a lot of people going to and from many different places, as long as they’re willing to change at one of these stations. But now the train suddenly slows way down and begins making local stops and crawling slowly through pedestrian zones. Suddenly, the service is no longer useful to anyone going further than this area, because you can only run really slowly for a few km before your travel time becomes unacceptable to someone going further. So if you’re going to have a good load of passengers they all need to be going to the area you’re slowing down for. In other words, that area needs to be the only major activity center for the city, or the last you're going to encounter in geographical sequence.
There's a bit of "tram-train" in the design of many North American light rail systems, and the same principle applies. For example, the Los Angeles Blue Line runs fast from Los Angeles to Long Beach but then very slowly on-street through the Long Beach downtown. That's OK in that case; nobody's going beyond Long Beach, and thus riding through this point, because beyond Long Beach is the ocean. On the other hand, light rail systems that have slow on-street operation in the middle of a line (downtown Portland and San Jose, for example) do tend to be less attractive for people trying to ride beyond the slow segment.
The key issue with the tram-train, then, is that by running trains into slow, pedestrian-oriented tram lines, you make them useless for anyone who wants to go much further. Once a train slows down to become a tram, it’s going to lose all its passengers within the next 3 km or so, because that’s the rough upper limit of how far people are willing to travel on a very slow tram (especially if they’re so time-sensitive that they really value the fast “train” part of the trip). So everywhere that people may want to commute to needs to be along one line, and in that distance. In effect that means one fairly compact downtown where almost everyone is going. So you need a strong downtown that also contains your big university and medical complex, if you have those things, because those will both be big transit destinations. I can’t think of many North American cities where this is still true. The only Australasian city that begins to fit this description is Adelaide, which is planning a tram-train project.
That downtown itself must be compact or linear enough that all the lines can serve it in a way that gets everyone within walking distance of their destination. (If they have to connect to another tram, you’ve lost the connection-avoidance benefit that is the tram-train’s whole purpose, and you might as well have just asked them to transfer from train to tram, as most cities do.)
If you’ve got all those features, tram-trains may well be for you!
*S3 is not a tram-train despite the number. It's the outer end of a Mannheim-based commuter rail line, and the "S" makes sense in Mannheim's nomenclature, though it's misleading in Karlsruhe's.
Portland's MAX more or less complies with most of the train-tram conditions and characteristics, and immediately came to mind as soon as you started describing what they are. It has, however, had major growing pains, partly due to capacity (thus the new bus mall alignment) and partly due to slow surface running. Currently, Metro (the regional government that does conceptual planning and funding for major infrastructure) is considering a new "express" alignment through downtown as a long term solution.
As cities such as Portland grow, there is a lot of reluctance to speed up downtown transit times, however. Most planners acknowledge that part of the problem is too many stops, but nobody wants to actually remove one as that might reduce boardings. Indeed, there was a proposal for a downtown tunnel for crosstown trains but the idea was dismissed in the current RTP process partly because it would reduce the number of stations and therefore the amount of people who are within a quarter mile walk of MAX in downtown.
A lot of this reluctance no doubt stems from the original political support for many light rail projects laying with supporters of progressive urban planning, whose desire was to reshape the street environment more than it was to create rapid transit. Much of the principles of these early light rail lines are being inherited by domestic streetcar projects, whose focus is largely on economic development and land use rather than on transportation.
Posted by: Alexander Craghead | 10/18/2009 at 11:01
Actually, the slow downtown part has been identified as a barrier and it has been decided to put underground the section between the stations Mühlburger Tor, Durlacher Tor and, I believe, Ettlinger Tor and Mathystr. Additionally to the slow speed, the pedestrian zone section barely has enough capacity to put trains through and one accident can lead to a blockage of the whole section between Kronenplatz and Europaplatz - what people tend to call the "yellow wall" through the midst of the pedestrina zone.
Posted by: Hermann | 10/18/2009 at 12:26
Jarrett, I'm surprised that you invoke the crash safety standards here. Germany isn't the US; its rail safety standards demand positive train control to ensure crashes don't happen, instead of high train weights to ensure people think the train will survive crashes.
Besides, is there any difference between the European tram-train and the American interurban?
Posted by: Alon Levy | 10/18/2009 at 14:43
I agree that tram trains, while very useful in some cases, have somewhat limited applicability. Among other things, you can't have too many regular trains already using the tracks, and the tram-trains are almost certainly slower than regular trains and thus take up more room in the schedule. Plus, tram-trains have lower capacity than regular trains, so it's really not a very efficient use of track capacity. Thus, it can only work in a relatively small city like Karlsruhe, where the commuter rail network is not very busy.
Posted by: anonymouse | 10/18/2009 at 15:13
Yes. There was always some distinction between the interurbans and the mainlines, and the interurbans pretty much all died out or transformed into mainline operations. The idea of the tram-train is running what is basically an interurban on the mainline, mixed in with the big trains. Also, Europe does have crashworthiness standards, they're just based on actual facts rather than the rather arbitrary figure of 800,000 lbs of buffing strength that was obtained by direct rectal extraction back in the 30s or so.
Posted by: anonymouse | 10/18/2009 at 15:16
Tri-Met seems to have learned a bit--it appears that the Milwaukie MAX line, which should see groundbreaking in 2010 IIRC, will be a new line (tentatively the "Orange Line"), rather than an extension of the Green or Yellow, both of which terminate at PSU, near where the Milwaukie line will enter downtown.
This, of course, is one argument against through-running--if the line needs to serve a lot of stops downtown, it makes more sense to have two lines with an overlapping segment, rather than one long line with a slow section.
Posted by: EngineerScotty | 10/18/2009 at 20:07
Well, except that someone going from Milwaukie to North Portland, say, will still have to ride all the way through the slow section, but in addition they'll have to make a connection. So I'm not sure that's an improvement.
If the proposal is as you describe, I hope the real reasons are (a) operational reliability due to shorter lines and (b) increased intra-downtown frequency on the mall, such that you really can always see the next one coming.
Posted by: Jarrett at HumanTransit.org | 10/19/2009 at 01:47
I find it interesting that in Europe, towns and cities with populations of around 100,000 have streetcars, trams, or light rail of some kind.
Posted by: W. K. Lis | 10/19/2009 at 06:27
It may well be the latter, keeping in mind that I don't have a reliable source (and the matter may well not be settled anyway). An Oregonian article reported that Tri-Met runs circulator trains (running every half hour) up and down the mall on weekdays, to increase service frequency to an average six-minute headway (this, plus a green and yellow every 15 minutes); were this replaced with two Orange trains every fifteen, that would result in five-minute headways.
I doubt that many commuters are making Milwaukie to North Portland trips (or Oregon City to Vancouver, if we imagine the yellow and "orange" lines extended to their likely eventual termini).
It could also be that Tri-Met instead intends to extend the Green Line (or the Yellow) to Tigard--a Barbur Boulevard alignment is one of the possible extensions after Milwaukie gets done.
Posted by: EngineerScotty | 10/20/2009 at 11:30
If I were cynical, I would expect that the FRA crashworthiness standards were promoted by freight railroading interests who a) operate rolling stock that either already meets the standards (by virtue of being a heavy freight locomotive capable of hauling dozens of railcars), or is exempt (by virtue of being a freight car which does not carrying humans, hobos nonwithstanding); b) really would rather not spend the money to upgrade signalling and control systems, especially on the long stretches of rural track found in the US, and c) would rather not have passenger services on their lines in any case.
But I'm not cynical.
Posted by: EngineerScotty | 10/20/2009 at 11:36
Lest us not forget that Karlsruhe's tramtrains or zweisystem (two system) LRT were conceived to do away with transfers, providing a truly seamless journey.
What many, on this side of the pond don't realize, is that tramtrain has also been awarded "Product of the year', by the influential German business magazine DM, some years ago by the phenomenal increase in ridership that occurred with tramtrain.
When the first tramtrain line to Bretton was opened in 1993, weekdays (5 days) ridership increased 423% from 488,400 passengers a day in 1992, to 2,064,000 in March 1993! Overall ridership increased 479%, from 533,600 a week to 2,555,000 a week! ( Albtal-Verkengesellschaft Karlsruhe and ABB Henchel)This from eliminating one transfer from commuter train to tram.
The maximum speed of tramtrain is 100 kph on the mainline and in Germany, the signaling system is so designed to proven collision and to date I do not think a tramtrain has collided with a mainline train.
Here we have LRT that can operate as a commuter train, light rail and a streetcar and is now in operation in over 15 cities in Europe, including Paris.
There is a lot to learn about Karlsruhe's tramtrain as the concept is wonderfully suited for North America.
http://railforthevalley.wordpress.com/2009/10/07/french-delights-part-5-pariss-tramtrain-line-t-4-could-this-be-a-template-for-tramtrain-in-the-tri-cities-at-11-million-per-kilometre/
Posted by: zweisystem | 10/26/2009 at 22:09
Did you actually read my post, zwei? ;)
Posted by: Jarrett at HumanTransit.org | 10/26/2009 at 22:12
He has Jarrett, but disagrees with your analysis.
Posted by: bulleid35028 | 11/01/2009 at 09:16
You've spent a total of two days to "analyze" a system that operates far and wide in the area, obviously looking for ways to critcize and belittle the system. I've spent weeks analyzing and photographing the system, and can safely say that it is, in my opinion, the world's finest, well-coordinated tram system. While, like any system, there can be improvements, it is a model that some cities in the world could emulate (and thankfully some are doing so), especially New York City, where billions of dollars have been spent over the past few decades to create a third-rate system, while using the Karslruhe mode, especially in conjunction with the Long Island Rail Road, could have saved billions and created a very useful system. My suggestion is that you stay in Karlsruhe for a few weeks and really see how the entire tram and tram-train network actually works and is appreciated by the population, instead of just popping in for a quick two-day visit and pretending to know everything about the system in that little time period.
Joseph P. Saitta, Editor, Traction Yearbook
Posted by: Joseph P. Saitta, Editor, Traction Yearbook | 11/02/2009 at 05:15
Note the distinction between description and judgment. I was not pretending to understand the whole system, merely observing the limitations of the model and what this says for its range of application.
Posted by: Jarrett at HumanTransit.org | 11/02/2009 at 12:24
Detroit.
The train stations have avoided the urban core due to considerations related to the Lake. It would still be very hard to cram a train station into an appropriate location.
The remnant urban core is the only place which is a destination for significant enough numbers of people to need a train (there are a lot of very diffuse traffic flows suburb-to-suburb, but downtown remains the only *dense* one).
See also: Cincinnati, where the urban core has relocated away from all the train stations.
Of course, these are cities which can barely manage to build a tram, period. :-P
Posted by: Nathanael | 11/03/2009 at 12:13
Hermann. When did they decide to tunnel and put the S-trains under Kaiserstrasse?
Personal opinion of course, but the Karlsruhe system is incredibly easy to use and one of the best in the world.
I too like the OP would just hop on the next train and ride it although in my case it was to the end. It was my way of exploring my new home at the time.
The other point I feel the OP left out is that many of the S-trains go through the city and do in fact slow down in the middle. The thing is they unload and reload.
Posted by: Chris G | 11/04/2009 at 08:33
I broadly agree with your analysis, but as you mentioned it is somewhat generalized.
You do not need to have all trip destinations in the central area as long as enough of them are there. The benefits to the majority just have to outweigh the inconvenience caused to the minority.
Also many of the trips with transfers may be better in this scenario depending on the structure of the city. Even if you lose time coming into the central area, your final destination may be significantly closer than from the railway station. I believe this is the case for many destinations in Karlsruhe.
Another possible factor is overall network design and timetable coordination. In Karlsruhe careful timetable planning at Durlach and other transfer stations may significantly alleviate the problem for passengers going past the central station, although I do not know whether this is the case. They will have to transfer of course, but again more people will benefit. It should also be noted that the light tram train vehicles have significantly more stops than the old heavy rail services, so there are also less transfers at that end of trips.
In any case you are naturally correct in that tram trains are only optimal for specific scenarios. I generally agree with our description of these scenarios, but I would also like to add that tram-trains can have useful applications in larger urban areas as additions to larger systems. You might for instance extend an existing low volume local train line into a satellite town, which is conceptually close to the Long Beach case. Or you might use tram trains to alleviate crowding at a central station by moving some low volume services into city streets when space can't be found for more tracks.
Posted by: www.facebook.com/profile.php?id=609058067 | 11/06/2009 at 10:34
Your revelation regarding the elimination of multimodal transit into a downtown got me thinking. You state that the situation that Karlsruhe is in is only applicable to cities built pre-rail, and a few very poorly planned 20th century cities. I respectfully disagree.
In international cities built before rail, counting most major cities in Europe, and most of the major cities on the east coast of the US, rail took prerogative in regard to use of space and often right of way, as soon as it was practical. A similar phenomenon occurred 100 years later when freeways were given right of way in many American cities. Henceforth, the tram-train would only be useful in cities that had very poor RAILROAD planning, or otherwise later on decentralized or recentralized their downtowns.
A major category you missed out on is the streetcar city. I live in Berkeley currently, and for many reasons, the Bay Area is a logistical nightmare when it comes to transportation planning (primarily because of the large body of water in the middle of the population bases). Although this isn't 100% true, Berkeley, Oakland, and Alameda were built as commuter cities for San Francisco, with streetcars being taken over the bay via ferry until the Bay Bridge was completed in 1939. At the same time, the entire East Bay is also served with heavy rail (now Amtrak) stations, that today sit almost completely unutilized. The streetcars in the East Bay are now completely defunct, and Amtrak is near-completely defunct, and we now have a terribly poor bus system that never runs on time and provides very poor service, and BART, which is only marginally served by bus service, sometimes in a very circumlocutious manner. If the East Bay could revamp the old streetcar lines, and connect it with Amtrak (which also serves a much needed corridor between Oakland and San Jose not served by any other mode of transportation save automobile), we could dismantle the whole AC Transit system, and even figure out a way to get those cars onto the BART tracks and into the city.
While I'm on a roll, I may as well bring up Caltrain too. Caltrain is a hideously underutilized commuter heavy rail from San Francisco to San Jose. From San Jose north to about San Bruno, Caltrain was actually well designed to go through the downtowns of the numerous suburbs lining the El Camino / US 101 corridor. Except, it runs very slowly, and between San Bruno and San Francisco, it runs along a corridor that basically contains no population at all. The SF Metropolitan Transit Agency is spending some ridiculous sum of money to extend the Muni Cars (which already go underground and on the street) from the Caltrain terminus through the Powell St. intermodal station and up to North Beach or possibly Fisherman's Wharf. This would be vastly more useful, if they got rid of the giant industrial trains that run up and down the peninsula, and replace them with smaller Muni cars. Seamless integration would be fantastic!
Posted by: Alex Lemberg | 02/05/2010 at 13:16
@Alex Lemberg
Conversion of Caltrain to light rail was seriously discussed in around 1990-2, when Harry Britt was the city supervisor (i.e. city councilman) responsible for transportation. It made good sense from a San Francisco standpoint. But Caltrain's limitation has always been that while the geography of the line is great for all-day transit, the politics of San Mateo County (and Caltrain's own "commuter rail" operations culture) tend to give much more weight to the needs of the peak commuter than to the more diffuse but potentially transformative market that could be unlocked by all-day frequency. So the energy has gone into "baby bullets" and other ways to increase train speed rather than the all-day frequency that a LRT model would make possible. You're right, though, that the tram-train concept would have been useful here, and I wouldn't be surprised if it comes up again.
I disagree with the notion that the bay is in any way an obstacle to good transit. The Bay Bridge is one of America's great chokepoints, and this is precisely why transit is so competitive. As for the streetcars, the mid-20c Key System did much of what you describe, using a track on the Bay Bridge. The Key System streetcars didn't mix with heavy-rail trains, but they were local streetcars that flowed directly into the long nonstop bridge segment to deliver people into San Francisco.
For more on chokepoints, see here:
http://www.humantransit.org/2010/01/a-carbonneutral-seattle.html
Posted by: Jarrett at HumanTransit.org | 02/05/2010 at 16:46
Commuter rail can run all day at high frequency, too. Caltrain is planning to convert its line to higher frequency, with double the peak frequency, higher off-peak frequency, electrification, and lightweight trainsets.
Posted by: Alon Levy | 02/05/2010 at 18:16
Alon. Yes, that's exactly what we were working on 20 years ago!
Posted by: Jarrett at HumanTransit.org | 02/05/2010 at 18:33
The tunnel is totally officially in construction in Karlsruhe since the 20 January 2010
http://www.diekombiloesung.de/
and especally the flash :
http://www.diekombiloesung.de/kombiloesung/wLayout/design/flash/bauphasen
http://www.diekombiloesung.de/kombiloesung/wLayout/design/flash/MAP/map.php
Then, this will no longer be a tram, but a semimetro (in german: Stadtbahn).
Karlsruhe follows the model of the Stadtbahn of Stuttgart :
Stadtbahn = a tram outside the centers, a subway in the centers.
Stuttgart is the neighbour-city, and the capital-city of Baden-Wuerttemberg, then of Karlsruhe also.
Stuttgart (the city of Mercedes, Porsche, Bosch) is also for me The Model for the S-Bahn
(a very great sub-urban train-network for a city of this size).
Another model is.. Munich : 100km of subway, 100 subway-stations, 500km of sub-urban train...
and all these, in only 48 years : in 1972, the city had only trams...
In a recent german publication, Munich is declared as the best city for public transportation in ... Europa : highest comfort, highest speed
(that's true : between two stations in the center, the subway reaches a lot of time the 85km/h... well, this is the city of BMW),
a very good coverage of the city, etc, etc... )
http://www.faz.net/s/RubCD175863466D41BB9A6A93D460B81174/Doc~ECD9DE54217DC46B78316FD152F567913~ATpl~Ecommon~Scontent.html?rss_googlenews
highest ticket-prices also : well, to have the Mercedes of the public transportation has got its price, in Stuttgart and Munich.
but german like to spend money in transportations (cars or public).
well, i spend 400 euro the year for public transportation.
is that really much more than in other cities ?
one big difference between Germany and France :
the National Government of Germany pays 60% of the Karlsruhe upgrade in Stadtbahn, with the tunnel.
in France, the National Government gives only... 20% of the costs for subway sections.
thats explains why, in France they buy so many "tram", and in Germany, more tunnels.
in France, Paris gets the money... not the other cities.
a big difference also : the German State expects that an official study gives a good "efficacity quota".
it gives no money, if this is not cost-effective....
in france, they only consider the price of km at construction...
not the efficacity of the system, the result in term of service, etc, etc...
alternatives solution, somehow cheaper at construction, but making no sense economically have less chances in germany, than in france.
typically in Munich : they prefer a very powerfull second big and very tunnel thru the center of city.
the solution of train aerian, in the south, sounds cheaper, but makes no sense economically, because it doesn't go thru the center.
by the way : the idea to have connection between the center of a city to other part of the region,
doesn't come from Karlsruhe, but from Berlin with S-Bahn or Paris with the RER hundred years ago :
a sub-urban train in tunnel trhu the center.
Karlsruhe will do the same. It is not a model, because itself, this city changes the system from tram to Stadtbahn.
it is not a model, because it is not applicable to other european cities :
the center of karlsruhe is built with very large streets.
very rare in the old Europe...
my blog : http://metrotramrer.blogspot.com/
my site : http://metrotramrer.201w.com/
Posted by: frederic.mtr@web.de | 02/20/2010 at 08:06