Mountain bike trails in WA (and applicable everywhere)

Hi all,

The mapping of mountain bike trails in WA has been a mess for a long time, and so as part of a larger project I started cleaning things up last week. The work of updating the map is the relatively easy part, the MTB culture/community around unsanctioned trails that has developed over the last few years (post covid boost in rider numbers, ebikes [the ones that don’t count as bicycles is another topic], and the info available online) is much more difficult.

This is a rather broad topic, but I am dumping everything together here. The issue categories are:

  1. Any changes are reverted by a malicious user.
  2. Tagging issues.
  3. Missing trails. [This is an easy one - no discussion needed]

Malicious User

A single user (“DBCA” - impersonating the actual Department of Biodiversity, Conservation and Attractions, who is the land manager for most areas) has been the cause of potentially all of the tagging problems. Deliberately vandalising the map but also actively reverting all other edits to bully out users making legitimate changes, all without using any valid changeset comments.

I can’t make any updates because of this user.

5 years of this has resulted in 3 user blocks already (keeping @Fizzie-DWG busy). Some quotes:

  • concerns have been raised that a large amount of your work concerning the names of bike tracks in Western Australia is, in fact, fictitious. Concern has also been raised over your naming of some of these tracks, and also other places such as here: osm.org/changeset/130619183.
  • when other Mappers have left comments on your Changesets, pointing out issues, or requesting further information from you, you have not responded to them.
  • concerns about your work regarding the Changeset Comments which you have been leaving when you have edited OSM.
  • when other Mappers have left comments on your Changesets, pointing out issues, or requesting further information from you, you have not responded to them.
  • Good to see that you apparently think our messages to you are a joke!
  • if you don’t want to do the right thing by other mappers, your account will be blocked for a long period, and your previous edits reverted.

I have reported them a few days ago (also noting the attempt at government department impersonation). Assuming that they will be blocked, what options will I have for when they create a new account? I know they will have rate limits on the new account, but history has shown that I should not expect this user to stop their vandalism.

Way Tagging

Problems

  1. No differentiation between paved bike paths and mountain bike trails.
  2. No differentiation between (signposted) official trails and (not signposted) unsanctioned/pirate/informal trails.
  3. Fictitious naming (only existing on OSM) of unsanctioned trails.
  4. No differentiation between unsanctioned trails that the land manager has not physically reacted to, and the ones that they have (closure signage, gates, covering, demolishing).
  5. Unsanctioned trails can be closed by the land manager but physically reopened by individuals. ie the land manager puts up signage explaining the closure and/or taping across the trail, adds logs across the trail or completely digs up the area. Then someone comes along and removes the taping, moves the logs, clears a new trail.

Tasks

I’ve broken down tagging updates to these tasks for each riding area:

  1. Change `highway=cycleway` to `highway=path` for all single-track MTB trails and set `path=mtb`.
  2. Add the IMBA rating `mtb:scale:imba` to all offical trails. Remove from unsanctioned trails (which are not rated). `mtb:scale` can apply to either.
  3. Remove any `route=mtb` on a way (only valid to relations).
  4. Add `informal=yes` and change name to `mtb:name` to unsanctioned trails. Remove any fictitious `name` values created by the user “DBCA”, possibly adding legitimate names if known.
  5. Generally `access=no` and only `bicycle=yes` (official) or `bicycle=permissive` (unsanctioned). Never `bicycle=designated`. `electric_bicycle` the same. Other access types (foot/running) only where the trail allows access, which is rare.
  6. Ensure a `surface=[dirt|gravel|…]` is set for all.
  7. Set a `mtb:scale` for all.
  8. Split ways where multiple trails overlap. Combine all of the ways for each named trail into a relation.
  9. Combine all of the trails for a named riding area into a relation.

Munda Biddi

I want to avoid changes to many sections of the Munda Biddi long-distance touring route even though most would be considered a “mountain bike trail” by some people. The route is a mix of roads, fire roads, and singletrack. It also allows hiking the entire length. I’m only concerned with the singletrack sections that have been built for the Munda Biddi as part of this cleanup (especially where they are within a MTB riding network).

Closed Unsanctioned

There’s still the tasks involved with dealing with closed/decommissioned unsanctioned trails. This is probably an ongoing discussion.

National Parks

Some trails may be best being totally deleted from the map due to the environmental and social impacts of their existence, but I won’t get into delete vs not. However, especially in national parks, `bicycle=no` should be set on all unsanctioned trails (even without signage, it is the general rule that using informal trails in national parks is not allowed), and so this will be considered.

Actively Closed, Closure Actively Ignored

Finally, closed unsanctioned trails but without any signage or indication of the closure. This is where the trail closure has been ignored and the closure signage/blockage removed by someone, or a new path cleared. Riders actively use this trail, and many many have never seen any indication of a closure. I would add a note tag about the fact that this was actively closed/decommissioned, and so I feel it would be an `access=no`/`bicycle=no` and the note enough to stop future mappers reverting this. That’s the general legal status, rather than the general user behaviours - what’s the preferred “OSM approach”?

Example from Tas

For reference, here’s an example sanctioned vs unsanctioned trail from Hobart, Tasmania: https://www.openstreetmap.org/edit#map=16/-42.89773/147.26858

“Upper Luge” is official, “Lower Luge” is unsanctioned. This shows the practice of using an invalid `route=mtb` tag, has inconsistent use of `access=no`, but does follow tagging guidelines (and common sense…) for showing as an MTB trail.

I take a keen interest in the Munda Biddi Trail route that I try to keep updated to match the latest route files issued by DBCA (the genuine DBCA :slight_smile: ).

I think its not just WA that has an issue with some of those tags. It would also appear not just that user has been incorrectly tagging paths/tracks.

route=mtb is on over 5,000 ways across the country so I think work would also be needed on the wiki pages to spell out the requirement that its should only be on relations. If we can get some consensus then it will be much easier. The tag has been in use for many years (the Tassie example was tagged 7 years ago) so it looks to be a long standing misinterpretation of those tags.

Task 4 - should be informal=yes (not ‘true’).

In regard to the Tassie example - those two tracks have different land mangers - one side actively checks and updates the tracks/tags - the other side much less so but some consistency (across the country) would be great.

Great start in highlighting the issues.

How did you report them, as I haven’t seen anything come in for them?

If you’re absolutely certain it’s them, report them as a sock-puppet to data@osmfoundation.org

Something that I have seen before, is different MTB Clubs / groups having their own, different, names for the same track! :zany_face:

Good luck with your project!

2 Likes

Yep, the active maintenance of the Munda Biddi info is one of the reasons to not try to change much there. The same on Trailforks. I am of the belief that the Munda Biddi shouldn’t have mtb:scale:imba tags where it’s not actually rated (none of it?), which breaks my “offical trails” have an IMBA scale tag rule and is another reason to treat the MB sections as a special case.

MTB trails tags have definitely been a struggle to make consistent worldwide; there’s been many discussions…especially once you consider unsanctioned. The fact that there is no such thing as a “trail only for mountain bikes” doesn’t help as you could ride any bicycle on any trail if you were willing.

Hopefully a simple case of “any route=mtb on a way is intended to be path/track=mtb”. Good point that it’s worth a specific mention on the wiki pages.

:ok_hand: Fixed in the original post to keep as a reference.

Clicking the ‘Report this User’ link on DBCA | OpenStreetMap with “This user is a vandal”. Actually done twice (the second time to note the reoccurrence of the reasons for the previous blocks).

Should I resubmit via email?

For an unsanctioned trail with a name set by the builder and a fictitious name set on OSM for multiple years (plus made up Strava Segment names), there could be multiple “community names” for the trail - and none of them “official”. “mtb:alt_name” I assume for those?

Yes thanks, as those reports can take some time to reach us, as they go to Admins first.

1 Like

I am not familiar with those “high level” tags on the relation, and leave editing those to people that understand them better. I just focus on the ways within the relation.

If it’s clear it’s the same person, then the new account can straight out be blocked as a sock puppet account. Otherwise if they go straight in and start ignoring prior discussions, point out the issue briefly and any relevant links, and if they persist against that, then it seems like a DWG block would be in order.

(full disclosure I put together the wiki Proposal for that tag in 2020, but at the time I felt there was too much pushback and not enough clear support to push the tag further)

I support this. I usually go with,

  • highway=footway if it was built for walkers (this could include the footpath, or a purpose built bushwalking track),
  • highway=cycleway if it was built for non-mtb cyclists,
  • highway=track if it was built for vehicles, but isn’t a “road”,
  • highway=path + path=mtb if it was built for mountain biking

This has always been a bit of an open question, should mtb:scale:imba be reserved for only those officially assigned an “IMBA” rating, or can it be used for those OSM mappers self rate against the IMBA scale? The same can be said for the AWTGS grade for bushwalking.

I wouldn’t remove mtb:scale:imba from unsanctioned trails unless you gain a clear view from the community on this.

We’ll do things like lcn=yes on the way to indicate it’s part of a route, which is useful if we haven’t yet mapped out the full route as a relation.

Unless you have reason to feel the route is bogous, perhaps it could be left until the way can be added to a route relation?

+1 to adding informal=yes to unsanctioned trails. The names can be tricky for informal trails, if there is a wide ridership community who do use specific names for trails that become more widespread names we should include them. So it can be hard to tell what might be fictitious without a strong connection to the place and people using it. The originally mapper should be able to justify the names and where they came from, if not then they can go.

Personally I don’t like the top left access=no with specific modes =yes, since I interpret that as no access at all by anyone. ie. completely closed off. Instead I like to omit access=* and instead list the modes that are known, so if the sign says no walkers, then foot=no, if the sign indicates mountain bikes can use it then bicycle=designated or we could consider mtb=designated with mtb being an access mode in that context.

Why do you say never bicycle=designated? I see designated as being used when a way is signposted to indicate that mode of transport is permitted, and yes when it’s legally allowed, but not spelled out by any signage on the ground.

What tags would you use for the site relation?

I’m a big proponent of using the Lifecycle prefix - OpenStreetMap Wiki tags like disused:*=* or abandoned:*=*

If you have the local knowledge, I would say that’s okay, but I would steer clear of mass edits without a local check.

Is a sanctioned/unsanctioned trail determined based on presence or absence of signage alone? I’m not sure about this one, usually I set bicycle=no based on explicit signage indicating no bicycles, which could include signage at the main park entrance indicating no bicycles for the whole park, or just at the start of the trail.

The “OSM approach” is usually to map what’s actively happening on the ground.

How do we know it’s been closed without any signage? As a mountain biker myself, I always thought if there was no signage then the national parks people didn’t mind mountain bikers using it, but they just didn’t want to officially sanction it. Is that a thing? And I’m worried your suggested tagging leaves no room for that situation.

1 Like

@sweaty_techie have you seen Australian Tagging Guidelines/Cycling and Foot Paths - OpenStreetMap Wiki?

Thanks @aharvey great inputs and questions. For some better understanding:

The tagging strategy is a mix of the goal of making everything correct and the reality of managing such a large-scale change. Instead of expecting to jump straight to what is perfectly correct, the idea is to have a process of small changes that can be completed, starting with making it easy to filter to only offical and/or unsanctioned single-track mountain bike trails (or close enough to not make review easy). Currently there is quite good consistency (even if poor correctness) with how trails are tagged across the state, and I want to keep that.

I have local knowledge of the areas and almost all of the trails I am looking at, the clubs/organisations responsible for the official riding areas, and many of the unsanctioned trail builders and regular riders. There will be a few trails that I do not know, but I intend to keep them consistent. Some other local mappers of MTB trails have been bullied out by user DBCA, so there is not currently an active community to engage with and I am mostly thinking of data consumers.

The Aussie Cycling Tagging Guidelines was my starting point, plus looking at the other MTB discussions on the forums world-wide.

Having an IMBA rating on the unsanctioned trails may make sense for an idea of difficulty, but there are vary few existing trails with values set and many of those aren’t reasonable values. It is also not verifiable.

The collection of official trails in each area is well known. Having an early step of ensuring mtb:scale:imba is set on each official trail and not on any unsanctioned achieves 2 bonus purposes:

  • Make the trail rating info available (and correct, instead of values like blue).
  • Allows filtering for only official MTB trails without needing all unsanctioned to be correct (ie. filtering on informal requires that informal=true be set on all unsanctioned. Setting informal=false on official trails is extra work that adds nothing).

It’s a different case where the trail is part of a route (that’s a good tip for if there is). Removing is where a standalone trail that is not part of a route has the route=mtb tag.

Where the original mapper was the “DBCA” user, the names are 100% fictitious. As mentioned above, I do have the local knowledge of these trails. There is 1 trail that I know of where the OSM fictitious name is used by multiple people event though the original trail builders have signposted it with the “actual” name, which is a special case.

That’s not the definition of what access=no means though, as it is setting the default for all transport types to no and is overridden by the bicycle transport type tag. Tag:access=no - OpenStreetMap Wiki

For example:

access=no foot=yes – access is only allow by foot and via no other means.

This matches the fact that signage at the trail networks explicitly restrict access to all but those allowed (bicycles, and foot/running in some cases), and says it in a way easily processed by data consumers. I don’t want to worry about kick_scooter, carriage, cycle_rickshaw, …

  • Definitely not using bicycle=designated for the unsanctioned trails. I also prefer not using it for any MTB trails it has the potential for data consumers to prefer it for routing road cyclists / commuters if they don’t process the path=mtb tag or if there is no surface/smoothness/mtb:scale/width tags (tags like surface=gravel, width=1.0 don’t tell me if a road bike could comfortably use the trail). The “correctness” of whether designated should be used depends on if you read it as it must be signposted as a section of path specifically for bikes or just expected because it is for bikes.

I threw the last 2 relations tasks in later, and they should be in a seperate cleanup task as it’s individual trail/area fixes rather than a state-wide tags cleanup. I won’t follow up now as I don’t want to go out of scope of the tagging fixes until that is complete.

I’ve used demolished:highway=path for physically decommissioned (dug up) trails. Where the trail exists but the landowner has added signage that the trail has been closed it would be bicycle=no. For the case of the land manager not wanting trails mapped in sensitive areas in national parks where the vegetation grows back so slowly in our environment (the act of mapping is what makes the trail be used enough to make it exist on the ground) it is not:highway=path with a note.

This is the case where there was signage to mark the trail closed, but someone has removed it (real-world vandalism): “closed unsanctioned trails but currently without any signage or indication of the closure due to it being stolen / tampered with“. This is common.

Especially in the national parks, the land manager does not want riders using any of the unsanctioned trails and regulations mean it is not authorised regardless of the lack of signage.

I was hoping to keep my tasks list up to date with what steps I actually take / plan to take, but my forum account is too new for that.

Good to see this in depth discussion happening, I support your efforts in improving the quality of MTB tracks mapped in OSM.

Okay, and I see that Key:mtb:scale:imba - OpenStreetMap Wiki indicates this too.

Sure, personally I have hesitations against that style of tagging and would rather list out the expected modes of transport explicitly (eg. foot=no + bicycle=designated or foot=designated + bicycle=no, or motor_vehicle=private + foot=yes + bicycle=yes), and reserve top level access only when it applies to all modes (eg. access=private for a private track), but if the wider community prefers it that way, I’ll oblige.

I agree that bicycle=designated should not be used for unsactioned trails, but it should be used for an officially signposted MTB track, as the way is specifically designated/signposted for use by bicycles.

I think you need bicycle=designated on official mountain bike tracks, as that’s the only way to say that bicycles (mountain bikes) are both legally allowed to be ridden there, but also encouraged/explicitly allowed by the land manager.

It’s a good question about how to determine routing. but I think you do need to rely on path=mtb, smoothness and mtb:scale.

:+1:

If think if there has been signage, but it was vandalised, we shouldn’t straight away update the tags as if the signage didn’t exist. We can use other signage, like one at the park entrance that lists out all the official tracks, or potentially first party online sources. As time passes it becomes harder to keep things mapped based on a sign being present here 10 years ago, so there’s a balance.

Fair enough. Either way it would be nice if OSM could still record the subtle distinction between this path is clearly signposted as no bicycle allowed, vs. it’s illegal since it’s not a sanctioned trail but there’s no signage here and you’ll find mountain bikers here every week.