OpenStreetMap

When preparing to embark on a hike, I want to know if a trail suitable for me (my skill level and the time/energy I have for the hike). I would like to review a map and have trails clearly identified based on their level of difficulty and level of required exertion. There are numerous classification systems which grade trails for hiking and mountain biking including the Swiss Alpine Club Rating scale, German Single-track Scale, International Mountain Mike Association scale, Yosemite Decimal System, Australian Walking Track Grading System, the Sierra Club scale, the Appalachian Mountain Club, and Shenandoah National Park and Pigeon Forge National Park, and the US Forest Service and others. These rating systems variously take into account surface condition and obstacles, steepness, trail visibility, required gear, required wayfarer skills, risk/exposure, length, and elevation change.

Several of the established scales focus on exertion level based on length/elevation with different formulas but don’t take into account technical difficulty (National Park Service and Appalachian Mountain Club, Sierra Club). Several focus terrain and skill required with criteria for peak and/or average incline, surface material, smoothness, visibility, required gear, and risk/exposure (SAC, MTB, YDS, and USFS). While others combine these factors (Australian System, and Adventure Nerds). Others have proposed modifications/alternative scales that have not been widely adopted, but attempt to merge and isolate key factors.

OSM has well established tags for ways based on the Swiss Alpine Club Rating scale for hiking and the German Single-track Scale for mountain biking, but these are not always used. There are other tags that are related to the factors which contribute to the difficulty assessment (surface, smoothness, visibility, incline), again with variable adoption. The stamina and time required for a hike are a combination of elevation change and distance which are also OSM tags recommended for routes, but stamina and time will also be influenced by the technical difficulty of the ways.

Personally, I think it is useful to segregate the technical difficulty of a path from the exertion level required to complete it. I think it would be interesting to render a map with ways based only on technical difficulty, and routes based on a combination of difficulty and exertion required to complete.

I have taken a pass at correlating all of the various scales and harmonizing their criteria and factors, and then identifying the values for the relevant tags in OSM. The scales are not entirely consistent, so this table represents a number of choices I have made. I intend to use this table along with an algorithm to render trail difficulty based on a combination of whatever tags are available. Such an algorithm could also flag paths which are missing critical information about difficulty or have confusing/conflicting metadata potentially in need of validation (e.g. surface=paved & sac_scale=mountain_hiking, or smoothness=horrible & mtb:scale=0). I have also started on a MapCSS style for JOSM based on this table for my own use and exploration, I hope to upload it to github shortly. I also intend to expand the algorithm to take into account access indications as being identified by the trails working group.

Discussion

Comment from ChristianA on 8 August 2022 at 09:01

This is important, and rather difficult. I try to add the mtb:scale tag when adding/editing trails, but I find it hard to be consistent. For instance, a path may be rather easy (technically) for the most part, but has a small section that is difficult. Sometimes I then divide the path into sections with different values for “mtb:scale” but I do not want to divide the path into too many sections. It is also hard to remember this when I edit the map after a bike ride; it is easier to remember the overall feel of a whole path.

Would it be possible to use, say, data from Strava (or similar) to calculate the difficulty? Or directly from uploaded gps tracks? Of course one needs to know if each individual trace is from biking or hiking and so on. And it is quite dependent on who is hiking/riding as well.

Regarding visualization, I know that the OsmAnd mobile app can show mtb:scale on the map. I use it quite a lot and agree that it would be useful for other map applications as well.

Comment from Troy Hartwig on 8 August 2022 at 22:38

I have also wondered if incline might be computable from the path and SRTM data, but thinking the precision may be insufficient. It could also be computed from recorded GPS data (GPX recordings), again not sure about sufficient precision for a peak, but probably good for averages. There is an open source project which proposes to do this GIScience/osmgpxinclinecalculator or maybe the JOSM plugin ElevationProfile could be modified to do this too. In either case, this would be a significant development effort and learning curve for me personally.
This approach could address the incline parameter (and probably also the total distance and elevation change), but it couldn’t address obstacles and use of hands as independent factors.

Comment from ChristianA on 9 August 2022 at 05:22

SRTM data is probably far too coarse in resolution and has too low accuracy. The trail can have lots of small “ups and downs” that are too small to be in any DEM dataset, but still contribute to the trail being hard.

Thanks for the links, I will check them out!

I hope you will be able to report your progress on the OSM diary, it will be an interesting read. Best of luck!

Comment from SK53 on 11 August 2022 at 12:16

I think cycle.travel uses DEM data to compute inclines & therefore timings on roads.

I think for this general purpose even a 20-25 m grid DEM might be adequate, and in some places one can compute a much tighter elevation data on a much finer grid (for instance much of England & Wales is now covered by 1 or 2m Lidar data, DEM models are available for the whole of Switzerland, and at a more detailed level for some cantons, LINZ has elevation data for most of NZ, etc., etc.).

What I think we lack on hiking (& possibly MTB) trails is a good indication of objective hazards, which in my view contribute to this evaluation. Here’s a tentative list:

  • paths across tidal areas
  • quicksands
  • boggy ground. peat hags etc
  • unexploded ordnance
  • slippery when wet (often signed or remarked in guidebooks for much of the limestone pre-alps: “Heikel bei Nässe”)
  • exposure (“nur für schwindelfreie”)
  • loose rock (underfoot)
  • risk of slipping (associated with exposure, but sometimes not when the rock is very grippy, e.g., Skye gabbro) (“Trittsicherheit”)
  • glacier travel (crevasses, bergschrunds, boilerplate ice, serac fall)
  • rockfall (from above)
  • windthrown trees
  • avalanche danger (including cornice collapse)
  • cornices (potentially underfoot for the unwary, a common danger on the Scottish hills in the spring)
  • unstable ground, especially cliff tops beside the coast
  • rivers or stream crossings dangerous in spate or just fast-flowing

Some can be inferred, if the underlying natural feature (glacier, tidal flat etc. is mapped), but probably are best indicated explicitly using tags. Things such as ability to cut a hike/tour short, and overall strenuousness can be calculated.

In the main I think gauging things on a personal level is beyond the capability of a regular scale, and most scales really don’t work very well outside their original intended range (plenty of SAC scale T1 could be differentiated for older infirm walkers, for instance; and tagging every “voie normale” alpine ascent as T6 is equally coarse & uninformative). For instance I have COPD which mean that I avoid a couple of things if I can as they really impair my breathing: runs of steps over about 20 treads, and ploughed fields (200m on a ‘flat’ ploughed field ~ 800 m on an equivalent grassy one).

Comment from ke9tv on 17 August 2022 at 15:54

In your second paragraph you mention “exertion level based on length/elevation”. That’s highly dependent on the specific trip, of course, which is why those scales are usually for an outing rather than for a trail. Of course, there are extremes: if you want to visit the Ouluska shelter it’s a 30-km walk from anywhere, but in many cases you can choose your adventure to be as long/short as you have time and stamina to handle. Length and elevation gain are also, of course, measurable if you have a DEM and can drape a trail segment over it.

So the idea of focusing on specific hazards and specific technical demands is a good one. SK53’s list is a good starting point. Locally to me, I might add a few things (beaver activity), and if asked to come up with an overall rating, I think I’d need to separate summer from winter. A relatively straightforward YDS class-3 scramble in summer is an entirely different proposition when it’s glazed with black ice or covered by a metre-and-a-half of snowpack. A lot of entry-level guidebooks to the country around here will say of a trail, “hikeable mid-May to mid-October, best left to the experts outside those times,” while more advanced ones will be more frank about “do you need ice axe and crampons (and likely a rope and helmet) or will ski poles and Hillsound or Kahtoola spikes do?”

SK53’s point that the scales don’t work outside their intended range is also a good one. I don’t understand SAC’s T5 and T6 at all, really. They seem to stumble into the class four is a myth problem - once you’re skilled enough to grade a route like that, you no longer see the difference among the lower grades. Perhaps a mountain guide with extensive experience in assessing the abilities of clients could do a better job?

Everything also falls into the realm of subjectivity - what’s difficult for me might be easy for you, or vice versa. Personality also comes into it. I’ve had some other mappers criticize my trail grading - of trails they haven’t hiked, yet! - and my personal opinion is that they’re serious contenders for the all-European sandbagging title.

My concern is mostly motivated by hiker safety and impact on the land. A lot of my travel is into an area within a couple hours’ drive of New York City, and there are times when it gets invaded by legions of the clueless, too many of whom find themselves totally out of their depth. Some of that is probably unavoidable - they arrive having no idea what a ‘strenuous’ or ‘difficult’ trail is! (Moreover, you can’t fix stupid!) Still, we should do whatever we can to reduce the likelihood of having a party of novices arrive at a spot like steep scramble or even easy YDS 3 with ice starting to form and too little daylight left to go back the way they came.

Log in to leave a comment