The leaf_type tag has no definition

Hello All,

I’ve sat down once again to check the list of trees supported by my UrbanEye3D plugin, and I discovered an absolutely astonishing thing.

It turns out that the leaf_type values needleleaved and broadleaved have NO definition whatsoever. None at all. The wiki does not refer to any generally accepted standard, nor does it provide its own.

Usually, needle-leaved plants are considered to be members of the Pinophyta division, specifically the seven extant families (‘Araucariaceae’, ‘Cephalotaxaceae’, ‘Cupressaceae’, ‘Pinaceae’, ‘Podocarpaceae’, ‘Sciadopityaceae’ and ‘Taxaceae’), but the OSM wiki says that this is by no means mandatory.

The wiki article Key:leaf_type says:

leaf_type=broadleaved – Broadleaved woody vegetation. Mainly angiosperms, but also some conifers.

leaf_type=needleleaved – Needleleaved woody vegetation. This category includes all plants with needle-like leaves, even if they are not conifers.

Roughly the same in Tag:leaf_type=needleleaved:

Needleleaved woody vegetation. This category includes all plants with needle-like leaves, even those that are not conifers, e.g. some acacias. This category is also often used for scale-like and awl-like leaves.

And all would be fine – anybody can tell a spruce from an oak – but these needle-leaved acacias have already appeared in OSM. Here is a list of what is already out there in the database:

species genus family leaf_type
Acacia aegyptiaca Acacia Fabaceae needleleaved
Acacia nilotica Acacia Fabaceae needleleaved
Acacia peuce Acacia Fabaceae needleleaved
Archontophoenix cunninghamiana Archontophoenix Arecaceae needleleaved
Casuarina cunninghamiana Casuarina Casuarinaceae needleleaved
Casuarina equisetifolia Casuarina Casuarinaceae needleleaved
Cordyline banksii Cordyline Asparagaceae needleleaved
Hyophorbe lagenicaulis Hyophorbe Arecaceae needleleaved
Mauritia flexuosa Mauritia Arecaceae needleleaved
Pandanus utilis Pandanus Pandanaceae needleleaved
Vachellia nilotica Vachellia Fabaceae needleleaved
Afrocarpus gracilior Afrocarpus Podocarpaceae broadleaved
Agathis australis Agathis Araucariaceae broadleaved

And here too we find Agathis australis, the kauri – a most ancient conifer that outlived the dinosaurs – is classified as broadleaved.

Casuarina cunninghamiana – even in the curated list it’s marked needleleaved, despite obviously being an angiosperm.

The question is, does any of this make sense? Has anyone seen a definition in any book that classifies broadleaved vs needleleaved based on leaf shape, e.g., if the ratio of length to width is greater than 10, then it’s broadleaved, otherwise needleleaved?

Or maybe it’s all nonsense, and for any practical purpose (e.g. for rendering in UrbanEye3D) it is enough to consider a plant needleleaved if it belongs to the Pinophyta division, and all the rest as broadleaved.

I would appreciate any thoughts on this matter.

Best regards,
Zkir

4 Likes

Do you have images how leaves of this dubious plants look like?

At least in my local area it is obvious™ whether given tree is broad/needleleaved, with no real occurrence of dubious cases.

5 Likes

Sir, you must be new here. This is not astonishing at all.

11 Likes

Is that not the intention of this reference?

“These values relate to the Land Cover Classification System (LCCS) by FAO”

Yes, there is that reference. But definition in FAO LCCS is different, and more clear:

  • Broadleaved: referring to trees and shrubs of the botanical group Angiospermae, with the exception of ginkgo (Ginkgo biloba), which belongs to the Gymnospermae taxonomically.
  • Needleleaved: referring to trees and shrubs of the botanical group Gymnospermae (Ford-Robertson, 1971) carrying typical needle-shaped leaves.

I can interpret that the classification is based on taxonomy with only one specified exception (Ginkgo biloba)

Personally, I would have said it’s how the leaf looks to a totally untrained observer, who knows absolutely nothing about genus, family or anything else botanical, except (maybe) the common name!

16 Likes

This is my approach as well. If I know the genus or species, I tag that. If I don’t, but I see a broadleaved tree, I tag leaf_type.

I’m not consulting a wiki definition to determine if I should use broadleaved or needleleaved.

In any edge cases I would tag the species if I knew it, and omit the tagging altogether if I didn’t.

11 Likes

And if both are tagged then data consumer may derive own classification and ignore broadleaved/needleleaved classification provided by OSM.

I added now link to this thread in Key:leaf type: Difference between revisions - OpenStreetMap Wiki that quotes @Fizzie-DWG

it’s how the leaf looks to a totally untrained observer, who knows absolutely nothing about genus, family or anything else botanical, except (maybe) the common name

1 Like

Indeed, when the species is known, the precise scientific leaf type can be inferred from that. The leaf_type=* tag serves as a very coarse classification that even untrained people can determine.

This question is asked by Street Complete, which is where the majority of the values are probably from (though that quest is off by default IIRC).

I know Verpucci also has a field for species, which I can only being filled for notable trees like large ones or popular cherry trees

1 Like

Far less than half. Per StreetComplete Statistics AddForestLeafType and AddTreeLeafType each were solved less than 500k times and leaf_type | Keys | OpenStreetMap Taginfo has over 11 000k uses

So at most 8.2% was added with SC - and in reality less if any of these objects were later deleted.

Still more than I would have guessed, but if weighted by area it would be far far less.

3 Likes

Thanks you guys for your comments.

It makes sense to me, but it means that for corner cases opinions can vary (see the table above).

You can easily find them on wikipedia/wikimedia commons :slight_smile:

Some examples of needleaved from totally untrained observers:

Acacia peuce:

Cordyline banksii:

Quite definitely, but I (who I was describing earlier! :rofl:) aren’t going to be arguing if “you” want to call it one or the other!

Oh, & the 2 photos are needle & narrow!

2 Likes

first for me is rather needleaved if anything (and AFAIK no healthy tree would have leaves like that in my area)

second I would not tag either way, unless it grows to be tree-sized?

probably, though I do not care enough about it to look for photos - but if you want to get answers how and maybe also why people would tag, feel free to show some examples

I know I’m coming a bit late, but I read this discussion a few weeks ago and it strikes me as odd that Araucaria would be classified as needle-leaved. Monkey-puzzle leaves look pretty broad to me, even if they’re arranged quite differently from how they would be on angiosperms…

1 Like

@PluMGMK you cannot be late to this discussion – it is of almost existential nature!

Monkey-puzzle leaves look pretty broad to me, even if they’re arranged quite differently from how they would be on angiosperms…

I guess some new values of the leaf_type tag for specific cases like this would be useful, especially if someone volunteer to create JOSM icons/UrbanEye3D textures for such trees.

Also it seems that leaf_type=palm should be rehabilitated.

If there is no solid scientific basis for the leaf_type tag in general, then it is absurd to blame leaf_type=palm for lacking one.

4 Likes

To whom it may concern: the map of leaf_type mixtures:

Notes:

  • leaf_type=palm is adjusted by the species tag: species of the Arecaceae family are considered to be palms.
  • What are leafless trees, nobody knows. There are however areas where they dominate. The wiki page suggests that those should be cacti, but cacti are not usually considered to be trees.
2 Likes

I think that wiki page you linked is quite clear?

Plants without any leaves and plants that apparently do not have leaves in the common sense.

Maybe. But “cactus forest” is a term used in wild, see say Getting Around - Saguaro National Park (U.S. National Park Service)

Sort of yes, but sort of no. The page does not even suggest natural=tree as useful combination.

So the question is, are there any tree species, worth adding to here or here, with leafless in the leaf_type column. Currently there are none :slight_smile: