The diameter_crown tag is fairly well used for trees. It’d be nice when showing trees to show larger ones as larger than smaller ones. One challenge is that the values can be a bit hit and miss. There are a couple of sorts of issues there:

One is “obviously silly values from an import gone wrong” like the 289 diameter_crown=width: 2.0; color: green; that have found their way into the database, perhaps by a failed JSON conversion.

Another is “human but not machine readable values” like “4 - 7 m”

Another again is implausible values for certain tree types. For example, this is apparently a London Plane that is 5 times wider than it is high. That’s technically possible with a bit of pruning, but unlikely. Much more likely is that the data was not sanity checked before import, and the “diameter” figure here is actually a “circumference”.

In order to work around these issues, when processing the data prior to display I:

• Make sure that diameter_crown is actually a number > 0. If it isn’t, I default it to 5.
• Make sure that diameter_crown is not more than 35. 35 was chosen as a reasonable cut-off for “large, but not unfeasibly so” (there are genuinely larger trees of course, but many in OSM are just rubbish data).

It would have been nice to have used the actual “diameter_crown” value when rendering. On raster (which I’ve not looked at yet) that would have meant one icon per zoom level, which isn’t ideal. On vector, it would be nice to have the MapLibre map style do calculations based on diameter_crown directly (the documentation suggests that something might be possible, but is spectacularly unclear, and attempts to do something with that haven’t yet worked), so instead for now I went with three “large” sizes and one “normal” one.

Here is a map, centred on a large rhododendron.

Edit 12/06/2025: The raster version of that map (that updates every 5 minutes) is now live too. Here is that big rhododendron again, but on a raster map.