Crap facilities in LTN 2/08

Some folk think that things would be better — or less bad, at least — for cycling in this country if only LTN 2/08 “Cycle Infrastructure Design” were strictly followed. They see some good recommendations in the guidance, and perhaps a solution to the more bizarre makeshift crap facilities. A few go further, thinking that the document could actually be the basis for the better, safer, more attractive streets that would support a mass cycling culture.

I’ve explained why I think LTN 2/08 is not fit for the purpose of guiding cyling infrastructure design, but I don’t blame people for seeing the contents of LTN 2/08 as an improvement on the cars-only street designs and crap facilities that we have now. There really are good things in the guidance. Instructions for filtered permeability are given — though sadly at the level of individual streets rather than whole neighbourhoods. There are strong words about the use of “cyclists dismount” signs — though perhaps they could be even stronger. Replacing centre lines with wide advisory cycle lanes, tightening junction geometry, and other cycle-friendly traffic-calming solutions are suggested. The streets and cycle facilities in LTN 2/08 generally look more attractive than those that most of us are used to.

But at the same time, the document clearly encourages certain kinds of crap. There are two in particular that I feel like discussing: bus lanes and shared pavements.

On bus lanes, LTN 2/08 says:

6.1.1 Bus lanes are generally popular with cyclists (Reid and Guthrie, 2004).

A chapter is then devoted to bus lanes, giving, for example, guidance on designing out close overtakes by specifying lane widths, and specifying that cycle lanes can not continue through bus stops.

The cited source for the claim that bus lanes are popular with cyclists, Reid and Guthrie, is behind a paywall. But the abstract says:

Surveys and interviews carried out in Edinburgh, Hull, Derby and London found that riding in bus lanes (including contra-flows) was generally very popular with cyclists because it appeared safer and more direct than cycling in general traffic.

That is, lanes from which all motor vehicles except buses are banned are more popular than lanes which are full of trucks and fast cars. Bus lanes on busy roads are something that cyclists use to get by, they are not an aspiration and they do little to reduce the barriers to would-be cyclists switching their travel mode. In fact there are many problems with bus lanes: they’re shared with some of the biggest, most intimidating and most polluting vehicles on the road, constantly stopping and starting; they’re shared with impatient and frequently hostile taxi drivers (though there is no good reason why they should be, and this could easily be remedied with a change to the rules if only politicians were willing to make it); they do nothing to solve the junctions problem; they don’t solve any problems outside of their hours of operation or the hours of parking restrictions; and they suffer from much the same left-hook problem as cycle tracks — and one that is far more difficult to solve with engineering than that of cycle tracks.

More important are the things that LTN 2/08 has to say on “off-road cycle routes”. Off-road cycle routes should mean cycle tracks and paths. But little in LTN 2/08 comes anywhere close to resembling proper cycle tracks. It’s all shared paths — the basic unit of the crap facility. Indeed, the first line of the chapter on off-road cycle routes prescribes these shared pavements:

8.1.1  Off-road cycle routes almost invariably accommodate pedestrians too.

(The switch from prescriptive to descriptive language is a bit bizarre. Perhaps the authors knew of the problems with shared pavements and couldn’t quite bring themselves to explicitly endorse it, but were prevented from recommending anything better? The document could equally state that cycle routes are almost invariably blocked by ridiculous obstacles and “cyclists dismount” signs. But it doesn’t, it strongly discourages such things. This is a prescriptive document therefore this is a de facto prescription for shared pavements: that is how engineers are going to use it.)

I’ve already discussed the damaging endorsement of “dual networks”, and the idea that standards can be compromised on cycle routes because Real Cyclists will naturally always prefer to ride on the roads. It shows up right from the start, in the introduction:

1.3.8: inexperienced and/or leisure cyclist – may be willing to sacrifice directness, in terms of both distance and time, for a route with less traffic and more places to stop and rest;

And it can be found again, in the section on off-road routes. After a good start on design speed, it explicitly recommends compromising on quality, all because it can’t imagine a cycle route having a separate footway:

8.2.1 On commuter routes, cyclists usually want to be able to travel at speeds of between 12 mph and 20 mph, preferably without having to lose momentum…

8.2.2 A design speed of 20 mph is preferred for off­road routes intended predominantly for utility cycling…

8.2.3 Where cyclists share a route with pedestrians, a lower design speed may be required. Routes with design speeds significantly below 20 mph are unlikely to be attractive to regular commuter cyclists, and it may be necessary to ensure there is an alternative on­carriageway route for this user category.

There are certain situations where a shared path may be acceptable. Outside of urban areas, where usage is low, for example. And shared use can be appropriate if applied not as a route but at destinations, to help get the final few yards to the parking. It is rarely the right way to build a through route in urban areas where usage both on foot and on bicycles will be high, leading to conflict. A manual should be explaining such things. This one isn’t, it’s just endorsing low quality shared paths — for that’s how it will be, and has been, interpreted — whether it intends to or not.

The formula for crap facilities continues where width is discussed:

8.5.2 A minimum width of 1.5 metres is recommended for a one-way cycle track. The minimum recommended width for a two-way cycle track is 3 metres

8.5.3 Where there is no segregation between pedestrians and cyclists, a route width of 3 metres should generally be regarded as the minimum acceptable, although in areas with few cyclists or pedestrians a narrower route might suffice.

These are, of course, minimum widths, and they are indeed acceptable minimum widths where, say, there is a short section where a pre-existing, immovable and unworkaroundable building or geographical feature makes the desirable width impossible. But they’re rarely appropriate over sustained distances, except perhaps, depending on the exact circumstances, on the lowest trafficked rural routes — and even then, routes that are predicted to be low usage do not always turn out to be so. That these are merely the minimum widths for low usage routes is mentioned in the document, and the authors can not be blamed for their misuse — though I would like more to have been said about what the actual desirable widths are.

But misused the widths are. Every new relief road and shopping centre distributer and every big new road submitted to the DfT for funding last year — even those in so-called “cycling cities” — has a 3.0 metre bidirectional shared pavement on one side.

Obviously the problem here goes far wider than just this document alone. The way that at least some local authority engineers and consultants approach this stuff is revealed in this delightful discussion on those other crap facilities — Advance Stop Lines:

My colleagues and I have been looking through LTN2 /98 and its more of a compendium of How Not To Do Traffic Engineering than anything else. I would hope that Figure 9.4 was swiftly removed from street – in fact I have to wonder why DfT even published the picture in the first place! Another one is Figure 7.2 which invalidates the double yellow lines – and thats given as a good example? Come on!

These are figs 9.4 and 7.2:

There are other marvelous comments in that thread…

I have NEVER seen the point of ASL across full width when a R/T is NOT permitted (and some even show this across three lane approaches.

I agree re the suggestion that 5 metres max depth is excessive. This measurement is applied as a standard in Edinburgh and I have queried the use of such a distance in a city where under 1% of daily commuters are cyclists.

Obviously the content of LTN 2/08 itself is not even half of the problem when highways departments are populated almost exclusively by non-cyclists who think that the worst thing about the cycling infrastructure guidance is a non-standard bicycle-shaped red traffic light and that advance stop boxes don’t need to be deeper than a truck’s blind spot, and when politicians are reinforcing that cars-first culture by pursuing fanciful programmes of “smoothing traffic flow”. But fixing the guidance looks to me like the easiest step in the change that is needed. If things are going to continue to be built by a formula with no understanding of the theory, we should at least make sure that the formula is right.

(Thanks to Mark and Paul, who helped to annotate the good and bad in LTN 2/08 a few months ago — though I don’t claim to speak for anybody other than myself in this post.)

The DfT’s crap cycling manual

No sooner had I posted a list of things for Norman Baker and his colleagues to be doing — to prove that they were doing things that will actually make a difference, rather than just passing the buck to under-resourced and poorly supported local authorities — than they acted. Baker and road safety minister Mike “Petrolhead” Penning have written a letter to local authorities, encouraging them to, er, look at their junctions and invite businesses to sponsor cycle lanes.

I don’t think many people have fallen for this charade. There is little point in simply suggesting that local authorities rebuild junctions. If they did — and they’re not going to on any significant scale unless and until they find the money from somewhere (and that’s unlikely to be from sponsorship), but suppose they did… — they would go through the normal design process and, er, the instructions would tell them to build the same cycling hostile crap as before.

There are reasons why we have atrocious junctions and crap cycle facilities. Our engineers and lowest-bidder contractors have been trained to build these things. They are also told explicitly not to build the sort of high quality infrastructure that we need. If we are ever going to make any significant advance, the government — central government — needs to do something to help our engineers into the 21st century. The first and most obvious step is to revise the guidance — the manual — that makes local authorities build crap.

The Department for Transport have, over the years, produced a number of guidance documents that tell council officers and consultants how to build roads and streets. Things like the “Design Manual for Roads and Bridges”, which tells you how to build a motorway… or city streets, if you like your streets to look and be used like a motorway. Better streets are built according to the principles of the more modern and civilised “Manual for Streets 2”.

“Cycle Infrastructure Design” (PDF) — commonly known by its serial number, “Local Transport Note (LTN) 2/08” — is the document which sets out the principles for building for cycling, and all the technical details of the government’s recommended facilities. The devolved administrations in Scotland and London have produced their own manuals which vary slightly from LTN 2/08.

Some cycle campaigners are fans of LTN 2/08 and think that if only it were strictly followed things would be better. In their briefing to The Times last week (PDF), for example, cyclenation say:

DfT publication LTN 2/08 (Local transport note no. 2, 2008) is generally good at setting out guidance for cycling provisions, but frequently goes unheeded.

I understand where cyclenation are coming from, and I think no ill of them for writing this. Because most British main roads and cycle facilities are even worse for cycling than LTN 2/08 recommends. Following the manual would be an improvement. But not much of an improvement. Saying that LTN 2/08 is good shows just how abysmally low our expectations have sunk.

The manual largely consists of guidelines rather than strict rules, and the guidelines are frequently broken. Certainly there are, as the cyclenation briefing says, cases where the guidance has gone unheeded and we have ended up with crap cycle facilities. But there are also cases where the guidance has gone unheeded and we have ended up with something far better than would have be provided had it been followed: some of the best (and yes, in this country “best” is hardly “great”) examples of on-street infrastructure — the tracks on Camden’s Royal College Street, for example — break all of the rules of LTN 2/08. And all too often — through a combination of poor training in how to use the guidance, competing political demands like “smoothing traffic flow”, and the many fundamental failings of the guidelines themselves — the guidance is heeded, and the result is still a crap cycle facility.

Because LTN 2/08 isn’t good. It frequently endorses the wrong things. It recommends against international best practice infrastructure and omits almost every detail of it. And it fails right from its first fundamental principles, which is why anybody can “heed” the guidance and still build whatever crap they like. I think that LTN 2/08 is a greater hindrance than help for cycling and that replacing it is a necessary step.

The introductory section of LTN 2/08 is the most widely endorsed. It contains a series of underlying principles for designing for cycling. Some of it is very good — the need for “convenient, accessible, safe, comfortable and attractive” space for cycling, for example, and the need to think at the level of the network, not just streets and routes. There is something of a disconnect between these principles and the rest of the guidance, and the good principles rarely shine through in the built designs. But it is also far from the case that the underlying principles are all good.

The first problem that leaps out while reading the introduction is the Hierarchy of Provision. I’ve written before about why the Hierarchy is the wrong approach to the problem, so I won’t here, except to reiterate that the Hierarchy is not fit for the role that it has been given — that of central formula for deciding which solution is appropriate — which is one reason why so many inappropriate solutions have been implemented.

The second fundamental problem is that LTN 2/08 endorses “dual networks”. It correctly identifies that different cyclists have different needs and abilities, but from this fact it draws some very wrong and damaging conclusions. “Some cyclists are more able and willing to mix with motor traffic than others. In order to accommodate the sometimes conflicting needs of various user types and functions, it may be necessary to create dual networks offering different levels of provision, with one network offering greater segregation from motor traffic at the expense of directness and/or priority.” That is, new, nervous and child cyclists will be grateful for a crap facility that gives way to every side road, or a winding backstreet route, while confident cyclists will want to be in their natural place — on the road, with the traffic, riding in the vehicular style. Indeed, the former category are expected to eventually cast off their training wheels and graduate into the latter  category.

I would have hoped that “dual networks” could have been the one thing that might be able to unite cyclists in opposition. As cyclenation say in their briefing to The Times, crap cycle facilities can do more harm than good when other road users get indignant at your refusing to use them. But I know there is one cycling campaigner and consultant who is proud of his dual network, and just in the past few weeks LCC’s Go Dutch campaign has also taken a turn down the dual network path. I think this is the wrong path: when you stop designing infrastructure that’s good enough for everybody, you tend to end up with stuff that’s good for nobody.

The effect of the “dual networks” principle in LTN 2/08 is that neither “network” is satisfactorily designed. The low-traffic “network” can be designed down: it can concede priority, take circuitous routes, share busy pedestrian spaces, and even advise dismounting — yes, LTN 2/08 says elsewhere that those solutions are undesirable, but, hey, this is just the training network, they’ll soon graduate onto the road so what does it matter? And when it then comes to fixing the main roads and busy junctions, engineers will “take into account the type(s) of cyclist expected to use it”, conclude that the inexperienced and nervous cyclists will be usingthe other “network”, and design the roads and junctions accordingly. You can see the wretched result of the dual networks principle all over our cities — famously on the Euston Road, where the cycle route leads you along “a sort of fiddly thing”, while Real Men like Boris Johnson prefer to “scoot down the underpass“.

Theoretically the dual networks don’t have to be substandard, of course. But if you design infrastructure that isn’t substandard, there’s just no need to think in dual networks. The Dutch also recognise the variety of cyclists. Their engineering manual recommends designs of sufficient quality to accommodate that variety. Their designs work. The idea that cyclists will want to graduate on to vehicular cycling — that it is aspiration rather than a survival strategy — is perhaps one of the reasons why LTN 2/08 entirely omits quality separated infrastructure… except where it gives spurious reasons not to consider it.

The authors of LTN 2/08 have obviously not looked at Dutch solutions or the Dutch manual. There are a total of three references to the Netherlands and three further references to the continent in the document. Three of those references are about cycle parking. One is in an aside about roundabout geometry. A Dutch study measuring overtaking distances — probably irrelevant to current British conditions — is mentioned. Finally, the authors have this to say about modern European cycle track design:

“As a result of concerns over the safety of parallel cycle tracks crossing side roads, it is becoming common European pratice to reintroduce cyclists to the main road in advance of a junction. Cyclists pass the junction on the carriageway and then rejoin the cycle track.”

It’s just bonkers.

The final fundamental conceptual problem with LTN 2/08 is not explicitly stated, but is written right through the guidance. Despite being the cycling-for-transport infrastructure guidance, despite being introduced with a reminder of why cycling should be supported, the document just doesn’t treat cycling as a serious form of transport. That’s not a problem specific to LTN 2/08, obviously, and it will take more than just revisions to a document to change the entrenched culture of the nation’s highways departments. But it’s especially dissapointing to find the document so riddled with it. It is clear that the authors are stuck in the car-centric paradigm and lack imagination for how things could be.

“Advisory cycle lanes,” for example, “are not recommended where they are likely to be blocked by parked vehicles.” Not, “car parking should be restricted in cycle lanes.” We’re told that we like cycling in bus lanes: “They are preferred over off-road facilities as a result of the advantage of remaining in the carriageway and therefore having priority at side roads” [my emphasis]. This is the guidance for providing for bicycles and it can not even imagine a world in which bicycles might have priority over turning vehicles. This is especially bizarre given that, technically, pedestrians have priority over turning vehicles — though pedestrians bold and brave enough to take it are ever rarer. To me it seems so blindingly obvious that the natural arrangement would be that anybody continuing straight would have priority over those turning, regardless of the means of travel of either party. The authors of LTN 2/08 can’t imagine that world — can’t imagine that there could be any alternative to our might makes right of way world.

What of that top-of-the-hierarchy solution, “reducing traffic volume”, if highways authorities can’t even imagine a cyclist having priority over car parking or motorists leaving their driveways? This is a problem that obviously goes far wider and deeper than this one document — Karl’s experience of the LTN 2/08 in practice illustrates the cultural problem we face. But replacing this document has to be one of the first steps to changing that culture. This is the document that Norman Baker says “provides comprehensive good practice advice on a range of practical infrastructure measures to help cyclists,” when he tries to shrug off the Cities Fit For Cycling campaign. It doesn’t. It’s part of the problem, and it’s his problem.

These are just the problems with the fundamental underlying principles. Just wait ’till I get around to listing the ridiculous details — the crap facilities it recommends and the almost complete absence of of best practice solutions from this “comprehensive good practice” guide…