Render landuse=quarry fill at the same zoom level as surrounding land cover

[ZeLonewolf]

Expected behavior

landuse=quarry is rendered at the same zoom level as other land cover features such as natural=wood.

These features can be quite large, so rendering them along with other large-scale landuse features is appropriate. For example, the world's largest limestone quarry in Michigan is 6.4 km by 2.4 km in size.

Actual behavior

landuse=quarry is rendered only at zoom 10 and higher. At zoom 9 and lower, quarry areas show up as holes in the map in areas where landuse and land cover are comprehensively mapped.

Screenshots with links illustrating the problem

This area of West Virginia, USA has a large number of quarries surrounded by woods (shown at zoom 10):

Location: https://www.openstreetmap.org/#map=10/37.9464/-81.6270

The same area, zoomed out to zoom 9, shows blank areas where the quarries are:

Location: https://www.openstreetmap.org/#map=9/37.9464/-81.6270

[kaneap]

I'd worry that on low zooms, this would start to appear very similar to the rendering for populated areas landuse=residential etc. (Luckily this is currently avoided on high zooms with the hammer pictograms)

[imagico]

This requires providing a bit of a wider context.

We so far have no consensus on the low zoom strategy for landcover rendering. The original strategy was to start the different landcover types at different zoom levels as you zoom in to successively add more information to the map. This turned unsustainable as the level of detail in mapping increased and the focus of OSM and of OSM-Carto widened to different climate zones where different landcover types dominate. As a result the attempt was made (in #2654) to fade landcover into a purely decorative depiction at the lower zoom levels and subsequently in #3458 to extend the shown landcover types somewhat and selectively to lower zoom levels. This failed (see #3513 #3647 #3935 #3936) and we partially reverted the fading in #3952 - though many of the problems of that strategy still remain, including the one this issue is about.

A suggestion how to address this in a comprehensive and sustainable fashion was made in #3670 (together with #3896) - but that has been lacking consensus support for several years now. No other strategy for low zoom landcover rendering has been suggested so far.

The suggestion of this issue would - as @kaneap hinted - aggravate the existing problem of similarity between the low zoom builtup color and the quarry color (see also #2905) and it would not be a sustainable strategy for low zoom landcover depiction in general because keeping rendering all the different landcover colors from the high zoom levels as is at the low zoom levels in combination with the inevitable AGG rendering artefacts just results in a mushy mess that gets increasingly more messy as mappers add detail to the landcover mapping.

Long story short - i would suggest to close this and instead finally get either consensus on #3670/#3896 or a practically workable alternative strategy if such exists and that is generally preferred.

Side note on the specific area used as example here - general consensus among mapper is (and this is also documented on the wiki in that way at the moment) that landuse=quarry is to be used for the actual area of current material extraction only and not for the wider industrial area around it with transportation, processing, storage etc. or for areas reserved and scheduled for mining in the future or for areas that have been mined and that are in the process of ecological restoration. In the sample area landuse=quarry seems to be used quite a lot beyond the core consensus use of the tag:

https://mc.bbbike.org/mc/?lon=-81.704764&lat=38.063251&zoom=14&num=2&mt0=bing-satellite&mt1=mapnik