Proposal:Surface=crushed rock

surface=crushed_rock
Proposal status:	Abandoned (inactive)
Proposed by:	Julcnx
Tagging:	surface=crushed_rock
Applies to:	way
Definition:	A dedicated value for surfaces covered with loose angular crushed rock at cobble size (dominant stones ≥ 63 mm), the step above surface=gravel on the size scale. The material has distinct routing behaviour from gravel: very difficult for cyclists and very rough on foot.
Draft started:	2026-05-09

Forest track surfaced with large angular crushed rock, the lead image of surface=gravel on the OSM wiki. This material is the subject of this proposal.

SUMMARY

• What: surface=crushed_rock provides a dedicated tag for loose angular crushed rock at cobble size (dominant stones ≥ 63 mm), filling the gap above surface=gravel where no dedicated surface value currently exists.
• Why it matters: stones at cobble size deflect wheels unpredictably for cyclists and motorcyclists, risk hoof injury for horses, and can damage standard passenger cars. surface=gravel alone gives no signal that the surface is at this extreme; routers and data consumers cannot distinguish it from ordinary gravel.
• Already in use: 8 organic uses of surface=crushed_rock, 23 of surface=ballast, and 24 of surface=track_ballast exist without any documentation, showing that mappers reach for these terms when surface=gravel feels insufficient. This proposal formalises the gap rather than the specific name.
• Backward compatible: existing surface=gravel tags require no migration. surface=crushed_rock is opt-in; mappers uncertain of the stone size continue to use surface=gravel.

Problem

surface=gravel is a broad catch-all, analogous to surface=ground: it covers everything from small loose chips to very large angular crushed fragments, and mappers use it for all of these without necessarily knowing the particle size. This is intentional and not a problem in itself.

However, cobble-size material — which sits above the gravel range — has behaviour that is qualitatively different from medium gravel:

• Cycling is very difficult or impossible; stones are large enough to deflect a wheel unpredictably regardless of tyre width.
• Motorcycling is similarly hazardous; wheel deflection on loose angular rock at speed risks a fall.
• Walking is slow and tiring; ankles are at risk.
• Horse riding is uncomfortable and risks hoof and ankle injury on unstable angular surfaces.
• Standard passenger cars may bottom out or sustain undercarriage damage; high-clearance vehicles can pass at very low speed.
• The surface does not consolidate under traffic.

This material is uncommon in road construction globally. Typical engineered gravel surfaces use aggregate of 10–40 mm; stones at cobble size (≥ 63 mm) are found mainly on forest and logging tracks, quarry access roads, and rural mountain routes where large crushed rock has been spread for load-bearing or drainage without further compaction. Known occurrences include forest roads in Central and Eastern Europe (Czech Republic, Slovakia, Poland, Austria), routes forestières in France, farm tracks in the UK, and logging or mining access roads in other regions. The community has noted that "use for ways with railway ballast-sized stones is really rare" worldwide. When encountered, it stands out precisely because its behaviour departs so sharply from ordinary gravel.

The surface=gravel wiki page already acknowledges this gap. Its gallery caption for a small loose gravel photo notes that it "is useful to distinguish it from both well compacted surface and from big chunks of ballast."

The current lead image on the surface=gravel wiki page (Štěrková cesta u Pláštíku.jpg, Czech Republic) itself shows material at cobble scale: large angular crushed fragments clearly above 63 mm. This is atypical: most gravel road construction worldwide uses 10–40 mm aggregate. The lead image inadvertently makes surface=gravel appear synonymous with the large end of the range, reinforcing the need for a dedicated tag that separates the two.

surface=fine_gravel already provides a more precise value for the fine end of the gravel range. Cobble-size material above 63 mm has no equivalent.

Despite having no documentation, surface=crushed_rock (8 uses), surface=ballast (23 uses), and surface=track_ballast (24 uses) have already appeared organically in the database as of May 2026, showing that mappers reach for these terms when surface=gravel feels insufficient. This proposal formalises the gap rather than the specific name.

Proposal

Add surface=crushed_rock as a dedicated value for surfaces covered with large loose angular crushed rock fragments where the dominant stones are 63 mm or above — the cobble class that sits above surface=gravel on all major grain size scales (ISO 14688-1, Wentworth).

surface=crushed_rock is distinct from surface=gravel: objects tagged surface=crushed_rock can be retagged surface=gravel without essential loss of information (since gravel has historically served as a catch-all), but surface=gravel does not imply cobble size. When a mapper knows the dominant stones are ≥ 63 mm, surface=crushed_rock is the appropriate tag.

Definition

surface=crushed_rock describes a surface covered with loose, large, angular crushed rock fragments. The material is:

• Crushed and angular: mechanically broken from larger rock, producing sharp irregular edges. Not rounded by water (that would be surface=pebblestone or surface=gravel).
• Large: dominant fragments are 63 mm or above, the ISO 14688-1:2017 cobble lower boundary (Co, 63–200 mm), consistent with the Wentworth cobble class (64–256 mm). Material below 63 mm is surface=gravel. Standard railway track ballast is typically 31.5–63 mm (EN 13450), placing it below this threshold and within the surface=gravel range. Commercial crushed rock vendors independently corroborate this break: Finnish size ranges run 0–56 mm and 32–60/64 mm, then jump to 90/100–150 mm with no product spanning the gap, consistent with the geotechnical discontinuity at 63 mm.
• Loose: not compacted or bound. If the aggregate has been mechanically compacted, use surface=compacted instead.

The name describes the manufacturing process (mechanically crushed, producing angular fragments) and the size class (rock-scale, ≥ 63 mm), matching the Wikipedia definition of crushed stone as mechanically broken angular aggregate. "Rock" signals the size: it distinguishes this material from surface=gravel (which covers all sizes below 63 mm) and from naturally occurring angular rock surfaces (surface=rock, which is unplaced bedrock or scree).

Relationship to surface=gravel

Value	Size class	Particle size (indicative)	Practical effect
surface=fine_gravel	gravel: fine end	2–6.3 mm (ISO fGr)	passable by most bikes; some rolling resistance
surface=gravel	catch-all	coarse end ~20–63 mm	rough but rideable; difficulty is rolling resistance and sinking
surface=crushed_rock	cobble: above gravel range	≥ 63 mm (ISO 14688-1 Co; Wentworth cobble ≥ 64 mm)	very difficult for cyclists; stones deflect wheel sideways

When to use

Use surface=crushed_rock when:

• The dominant stones are 63 mm or larger (ISO 14688-1 cobble lower boundary).
• The surface is loose angular crushed rock, not water-rounded stones.
• The aggregate has not been compacted.

Typical contexts: forest and logging tracks, quarry access roads, rural mountain routes where large crushed rock has been spread for load-bearing or drainage without compaction.

Field size reference: a standard tennis ball is 65–68 mm in diameter, just above the threshold. If dominant stones are roughly tennis-ball sized or larger, the surface qualifies as surface=crushed_rock. If they are noticeably smaller, use surface=gravel.

Do not use surface=crushed_rock for:

• Stones below 63 mm, including standard railway track ballast (typically 31.5–63 mm per EN 13450): use surface=gravel.
• Rounded (water-worn) stones: use surface=pebblestone or surface=gravel depending on context. Angular character is what defines surface=crushed_rock; a water-worn cobble of the same size should not be tagged surface=crushed_rock.
• Compacted aggregate: use surface=compacted.
• Naturally occurring angular rock (scree, rocky mountain paths, exposed bedrock): use surface=rock or surface=ground as appropriate. surface=crushed_rock applies to human-placed quarried material only.

Routing implications

surface=crushed_rock sits above the gravel range on the size scale. Routers should treat it as:

• Very difficult for cyclists and motorcyclists; loose stones at this size deflect wheels unpredictably and increase fall risk. The difficulty mechanism differs from ordinary surface=gravel: on gravel the obstacle is rolling resistance and sinking (slow but controllable); on cobble-size stones the wheel can be deflected sideways suddenly, causing falls regardless of tyre width or rider technique.
• Very slow and uncomfortable on foot.
• Uncomfortable and potentially injurious for horses.
• Risky for standard passenger cars (undercarriage damage); passable for high-clearance vehicles at low speed.

How the wiki would look if passed

Tag:surface=crushed_rock

Forest track surfaced with large angular crushed rock (surface=crushed_rock).

surface=crushed_rock is a dedicated tag for loose angular crushed rock at cobble size (dominant stones ≥ 63 mm). It is the step above surface=gravel on the size scale: use it when the dominant stones are clearly at or above tennis-ball size (65–68 mm) and the surface is angular, not water-rounded.

Relationship to surface=gravel

Value	Size class	Particle size (indicative)	Practical effect
surface=fine_gravel	gravel: fine end	2–6.3 mm (ISO fGr)	passable by most bikes; some rolling resistance
surface=gravel	catch-all	coarse end ~20–63 mm	rough but rideable; difficulty is rolling resistance and sinking
surface=crushed_rock	cobble: above gravel range	≥ 63 mm (ISO 14688-1 Co)	very difficult for cyclists; stones deflect wheel sideways

Size reference

The lower boundary is 63 mm, the ISO 14688-1:2017 cobble lower boundary (Co). This is consistent with the Wentworth cobble class (64 mm) and represents the same geotechnical break across major classification systems:

Scale	Gravel upper limit	Cobble lower limit
ISO 14688-1:2017	63 mm (top of cGr)	63 mm (Co)
Wentworth (Udden–Wentworth)	64 mm	64 mm

Standard railway track ballast (typically 31.5–63 mm per EN 13450) falls below this threshold and maps to surface=gravel.

Practical effect

Very difficult for cyclists and motorcyclists; stones shift unpredictably under wheel and increase fall risk. Unlike ordinary gravel where difficulty comes from rolling resistance and sinking, cobble-size stones can deflect a wheel sideways suddenly, causing falls regardless of tyre width or rider technique. Walking is slow and tiring. Horses risk hoof injury on unstable angular surfaces. Standard passenger cars risk undercarriage damage; high-clearance vehicles can pass at low speed.

Possible tagging mistakes

• Do not use surface=crushed_rock for rounded (water-worn) stones: use surface=pebblestone or surface=gravel depending on context.
• Do not use surface=crushed_rock for compacted aggregate: use surface=compacted.
• Do not use surface=crushed_rock for smaller or uncertain gravel: use surface=gravel.
• Do not use surface=crushed_rock for natural exposed rock surfaces: use surface=rock.

See also

• surface=gravel
• surface=fine_gravel
• surface=compacted
• surface=pebblestone
• surface=rock

Features/Pages affected

Approval of this proposal triggers the following documentation edits.

Tag:surface=gravel

• Add surface=crushed_rock to the See also list as the dedicated value for cobble-size material above the gravel range.
• Update the definition to note that cobble-size material (≥ 63 mm) is outside the intended scope of surface=gravel when the mapper knows the stone size; surface=crushed_rock should be used instead.
• Consider replacing the current lead image (Štěrková cesta u Pláštíku.jpg, Czech Republic), which shows cobble-scale material, with an image representative of typical global gravel (10–40 mm). The Czech image could be reused on the crushed_rock page once approved.
• Migration implications: existing surface=gravel tags are not wrong and require no blanket migration. The vast majority correctly describe sub-63 mm material and should not be touched. However, mappers who knowingly tagged cobble-size (≥ 63 mm) surfaces as surface=gravel are encouraged to retag to surface=crushed_rock on resurvey. Mechanical retagging of all surface=gravel is explicitly not appropriate.

Tag:surface=fine_gravel

• Add surface=crushed_rock to the See also list as the dedicated value for cobble-size surfaces above the gravel range.

Key:surface

• Add surface=crushed_rock to the list of specific unpaved surface values.

iD editor

• If approved, open a pull request to openstreetmap/id-tagging-schema to add surface=crushed_rock as a selectable preset value, with the label "Crushed rock (≥ 63 mm, angular)".

Migration

surface=track_ballast has 24 organic uses and surface=ballast has 23. Both require resurvey before retagging: standard railway track ballast (EN 13450: 31.5–63 mm) falls within the gravel range, so these uses may describe coarse gravel rather than cobble-size material. Mappers should check the actual stone size on the ground and retag to surface=crushed_rock only if the dominant stones are ≥ 63 mm, or to surface=gravel otherwise.

The 8 existing surface=crushed_rock uses require no migration.

Why surface=crushed_rock and not gravel=*

A subtag approach (for example gravel_size=* or gravel=* with values like fine, coarse, crushed_rock) has been discussed in the community. This proposal uses a dedicated surface value instead, for three reasons.

Routers read the surface key directly. A router or data consumer that wants to penalise large loose rock does not need to parse a subtag; it reads surface=crushed_rock the same way it reads surface=fine_gravel. A subtag approach requires every consumer to additionally check for gravel_size=* or similar, and to handle the many objects where the subtag is absent.

Consistent with existing surface tagging practice. surface=fine_gravel is already a dedicated surface value for the fine end of the loose unpacked surface spectrum. A dedicated value at the cobble end follows the same pattern: named values at both extremes, with surface=gravel as the catch-all in between.

The subtag approach reached no consensus. The distinction between passable and impassable gravel has been discussed since at least 2023 (see external discussions) with no resolution under the subtag model. A dedicated value is a simpler, lower-friction change.

Why "crushed_rock" and not "ballast". Standard railway track ballast (EN 13450: 31.5–63 mm) sits within the gravel range, below the 63 mm threshold this proposal targets. The 23 organic uses of surface=ballast and 24 of surface=track_ballast may therefore describe coarse gravel (20–63 mm) rather than cobble-size material — the name is ambiguous about which size range the mapper intended. "Ballast" also describes the purpose of the material (stabilising a road or railway, analogous to ballast on ships) rather than its physical character, making it less descriptive than a material-based name. "Crushed_rock" names what the mapper can observe in the field.

Why "crushed_rock" and not "coarse_gravel". In ISO 14688-1, coarse gravel (cGr) is 20–63 mm — the same range as EN track ballast (31.5–63 mm per EN 13450). Applying "coarse_gravel" to ≥ 63 mm material would import that ambiguity: a mapper reading "coarse_gravel" would reasonably expect 20–63 mm aggregate, not cobble-size fragments. The geotechnical break this proposal targets (ISO 14688-1 cobble lower boundary) is not a gradation within gravel but a step above it. A separate proposal for surface=coarse_gravel covering the 32–64 mm range is now in draft.

Why "crushed_rock" and not "cobble_rock". Per the Wikipedia cobble article, cobble is a size class only (64–256 mm, Wentworth) with no angularity requirement: it includes both rounded water-worn stones and angular crushed fragments. cobble_rock would therefore describe a broader category than this proposal, which is specifically about angular crushed material. A water-worn cobble of the same size has different surface character and different routing implications.

Why "crushed_rock" and not "riprap". Riprap is specifically angular rock placed for hydraulic erosion protection: shoreline armour, levee facing, streambank stabilisation. Using "riprap" for a road surface would import hydraulic engineering connotations that do not apply to forest tracks.

Router integration

As an undocumented value, surface=crushed_rock is currently unknown to all major routers and will be treated as neutral or equivalent to surface=gravel, with no extra penalty. If this proposal is approved, routing applications that support surface-based cost models should add surface=crushed_rock with a cycling penalty higher than surface=gravel, equivalent to smoothness=horrible or above. Foot routing should treat it as slow and uncomfortable but passable.

The penalty must be qualitatively different from the gravel penalty, not merely higher on the same scale. On ordinary gravel the obstacle is rolling resistance and sinking: slow and effortful, but controllable. On crushed rock at cobble size the fragments are large enough to deflect a wheel sideways suddenly under load, causing falls that are independent of tyre width or rider technique. The routing signal this tag provides is not "slow" but "unpredictably hazardous": a category difference, not a degree difference.

Likely objections and responses

Why not just use smoothness=horrible?

Surface describes the material (stable, verifiable); smoothness describes the riding experience (subjective, weather-dependent). A router can penalise surface=crushed_rock for cyclists without needing a smoothness value at all. Both keys complement each other and can be used together.

In practice, mappers encountering cobble-scale surfaces disagree on whether smoothness=very_bad or smoothness=horrible applies. That disagreement is itself evidence that smoothness cannot carry the signal: both values are documented primarily around vehicle clearance requirements, not around the lateral wheel deflection and fall risk that makes large angular stones specifically hazardous for cyclists. On ordinary gravel the difficulty is rolling resistance and sinking: slow and effortful but controllable. On crushed rock at cobble size, fragments deflect a wheel sideways suddenly under load, causing falls independent of tyre width or rider technique. The two surface types are difficult for cyclists for different reasons and to a different degree. surface=crushed_rock names that material directly, independent of conditions or observer.

Only 23 uses: too rare to standardise.

Low counts reflect how uncommon the material is, not how uncommon the need for precision is. surface=fine_gravel covers an equally narrow case at the fine end and has 547,000 uses, demonstrating that precision tags in this family are adopted at scale once documented. The 55 combined organic uses of surface=crushed_rock, surface=ballast, and surface=track_ballast show mappers reaching for this concept independently, without any documentation.

How does a mapper judge 63 mm without a ruler?

A tennis ball is 65–68 mm, just above the threshold. Dominant stones roughly tennis-ball sized or larger: use surface=crushed_rock. Noticeably smaller: use surface=gravel.

Doesn't surface=gravel already cover this?

Yes, and it still will for mappers who do not know the stone size. surface=crushed_rock is opt-in for when the mapper can confirm the stones are ≥ 63 mm. Existing surface=gravel tags need no migration.

The tagging schema is already complex enough.

A single opt-in value for an uncommon but distinct material adds less complexity than leaving it undocumented. Undocumented edge cases tend to generate more discussion over time, not less.

Why not surface=cobblestone or surface=unhewn_cobblestone?

Cobblestone variants (surface=cobblestone, surface=sett, surface=unhewn_cobblestone) are manually placed paving surfaces: individual stones set one by one in a closed layer to create a stable, durable pavement. surface=crushed_rock is bulk-dumped loose angular material spread without placement intent; the surface does not consolidate and behaves as a loose aggregate, not a pavement. The construction method, the surface character, and the routing behaviour are all fundamentally different.

Can this tag be used for large water-rounded cobble-size stones?

No. surface=crushed_rock is scoped to angular manufactured material: "crushed" encodes the manufacturing process and the angular character that results from it. Large rounded cobble-size stones already have partial coverage under surface=pebblestone, whose documented size range extends to approximately 80 mm (see the pebblestone note in the Definition section above). Routing behaviour does differ between angular and rounded stones of the same size: angular fragments can deflect a wheel sideways suddenly because of their irregular edges; rounded stones of the same size tend to shift or roll. Both are hazardous, but for distinct reasons, and keeping the two cases under separate names preserves the option to differentiate them in routing if the community later identifies sufficient need.

External discussions

• OSM Community: Confusion between surface=pebblestone and surface=gravel; routing distinction for large gravel: includes Mateusz Konieczny's explicit statement that the community needs "to distinguish between reasonably sized gravel and ballast size one", and mapper testimony that 5–6 cm stones are impassable by bike while 2–3 cm are manageable, both currently tagged surface=gravel
• OSM Community: Tracktype and large-stone surfaces: includes independent proposal for a dedicated tag for "surfaces mostly consisting of stones larger than track ballast", and Map_HeRo's riprap concept for sharp-edged material above 60 mm
• OSM Community: Pre-RFC discussion for this proposal

Questions for the community

1. Are there regional names for this material in your area? (Confirmed so far: German Schroppen, a producer/trade term for crushed rock where the smallest particles exceed 63 mm, distinguishing it from Schotter/gravel; Danish skærver, meaning crushed rocks/stones, contributed by Linus_W_Frische.)

1. Can contributors share representative photos on Wikimedia Commons, ideally showing crushed_rock-surfaced tracks from different regions or contexts?

1. Is crushed_rock the right name? "Crushed" signals the angular manufactured character; "rock" signals the size. An alternative is cobble_rock, but cobble is a size class only with no angularity requirement, so it would include rounded water-worn stones of the same size — broader than this proposal intends.

1. Large rounded cobble-size surfaces also exist — for example deteriorated roads where the concrete surface has worn away leaving the rounded stone base exposed. Should these be housed under a shared cobble_rock tag covering all cobble-size surfaces regardless of angularity, or does rounded cobble-size material need its own proposal (parallel to surface=pebblestone at the smaller end)?

Comments

Please comment on the discussion page.