Talk:Tag:landuse=farm

From OpenStreetMap Wiki
Jump to: navigation, search

Discuss Tag:landuse=farm


animals or crops

I'm surprised nobody has proposed such a thing yet. I can only thing of the following:

  • animal=yes|cattle(non-specific) |mixed|sheep|cow|horse|* implies crop=no
  • crop=yes|wheat|strawberries|hops|* implies animal=no

- LastGrape/Gregory 21:52, 18 July 2010 (UTC)

Perhaps the reasoning for not proposing it is because crop rotations mean that the data is invalid within months Beelsebob 20:39, 4 March 2011 (UTC)

This is true in areas where there is crop rotion! Welcome to Haiti, here we have _regions_ that have grown specific crop for _decades_. So, +1 for defining crop=[name of crop] in the article. ... Rotation could be one option? --JaakkoH 19:38, 2 April 2012 (BST)

landuse=farmland

is it not Tag:landuse=farmland now? --katpatuka 11:55, 4 November 2008 (UTC)

We have landuse=farm and landuse=farmyard on Map Features at the current time. But there is Proposed features/farmland -- Harry Wood 23:12, 28 January 2009 (UTC)
I really don't understand the way land cover is handled in osm ... as if there were no cartographic conventions set up in the real world already. I can understand that osm as a "street-centric" project in the beginning just neglected that.
farm: Correct me if i'm wrong, but this sounds very "british". Reflecting on "farm", in my imagination, i see a neat little farmyard, surrounded by fields of crops waving in the breeze. So we call them farm. (?!). In "my world", we have villages with farmyards, and the villages are surrounded 100% by a patchwork of agricultural areas (of whatever use and ownership). So the alternative tag "farmland" makes much more sense then "farm", cause there is hardly never a direct connection between the farmyard and its agricultural areas. "Farm" only makes sense when the agricultural areas are directly connected to the farmyard.
Agricultural use of the land is a dominant factor in the landscape, and should be able to be presented accordingly (provided that a renderer decides to take care of this). The patterns of agricultural landuse can tell "the knowing" a lot about soils, climate and cultural heritage of a region. Would be a pitty to have only "primitive" ways of tagging them.
hedges/fences: again, this sounds very british. In my region, we have no fences at all and only some hedges along the tracks. So how could they be used for indicating field boundaries when they simply do not exist?
The "evolution" of osm tags for natural and agricultural items is a pretty mess. Have a look e.g. at http://etc-lusi.eionet.europa.eu/CLC2000/classes to see how easily things could be handled.
So the main tag for arable land should be landuse=agriculture, on the same level as the tags for urban landuse tags like residential, commercial etc, containing areas with tags like pasture=yes or permanent=yes, irrigated=yes and so on.
Following no carthographic conventions for non-urban, non-street/navigation connected items degrades the landscape to a "nice decoration" of the streets.(a humble geographers point of view)--Bernd v. 20:40, 13 August 2009 (UTC)
Hear, hear. landuse=farm is probably now embedded, but clearly should have been something like landuse=agriculture, but it might be too late and we'll have to accept farm as an unpleasant and misleading synonym. It's not just non-British places which don't have farms with nice fields delineated by hedges or fences. The village of Laxton[1] in Nottinghamshire, retains the open-field landuse pattern which was widespread in lowland Britain until a few hundred years ago. Field boundaries are distinct from hedges and fences, and may be distinct from the cadastral land parcels (which we might also want to map in the end). Similarly, I agree that there are plenty of well structured schemas (e.g., CORINE) for representing natural and landuse categories which would ideally form the basis for a coherent set of OSM tags. SK53 21:17, 13 August 2009 (UTC)
I've moved your comments up here since they relate to the landuse=farmland tag idea. For some reason you'd added them to discussion below about landuse=field tag idea. The other place we could have this discussion is over at Talk:Proposed features/farmland
As I said in my blog. We're using the tag landuse=farm for mapping farms fields AKA farmland AKA farms! A field is part of a farm is it not? I really don't understand what you are talking about or why you find this confusing. Obviously something to do with German meanings of the words, but I can't quite put my finger on it.
What difference does it make whether the farm is next to the farmyard? Step 1: Draw in two different polygons (next to eachother or not next-to eachother) Step2 : Give them two different tags.
We could have used other words 'farmland', 'fields', 'crops', 'agriculture', but the words we have used, and documented as tags are landuse=farm and landuse=farmyard Even if was feasible to migrate to different words (which it isn't) I'm yet to understand why you object to these.
-- Harry Wood 13:45, 4 February 2010 (UTC)

Very large swathes of land

"Obviously in many parts of the world we can expect very large swathes of land (almost everywhere outside of our towns/cities) to be used for farming, and therefore tagged landuse=farm. This may raise some interesting questions about how to handle landuse in general. But for time being perhaps we should follow the Wakefield example, and just get on with it!"

I'm mapping one of those parts of the world. Using this tag is basically..just inpossible..I can't acutally beleive it's now expected. Maybe its plausible if I just select the county admin boundary and have that as a outer relation tagged as landuse=farm and then make every other land-use or natural= as an inner?!..that's the only way I can think of this not being a sickening impossibly huge task. The Wakefield example is great for people mapping 'Wakefields', but as noted, many parts of the world are just huge areas of farmland. I don't really know what to ask?! My "Interesting questions" are:...who, what, where, when, why, how?..with emphasis on the how. Ben 00:42, 27 January 2009 (UTC)

It's not expected but plausible, even if it takes quite a lot more effort than entering just the streets. "Interesting answers" would then be: anyone with interest, anything visible from the roads, anywhere, any sunny day, because it's a dominant feature just like forests. And as to how: interpolation between known locations goes a long way, as does extrapolation (even with landsat images to assist) if the other side of the field is unreachable. With little practice and photomapping it's sufficiently easy to locate the corners near the roads (with adeaquate precision) even without stopping - it just takes more time when post processing the photos. Adding just one field at a time gets them all added eventually. Alv 08:38, 27 January 2009 (UTC)
I'm not exactly put off by the concept of mapping the landscape to a high level of detail, it's just making almost the entire area a landuse= area that is concerning, and rases many questions. Should the landuse be split by farm, or by parish, or by field or by some other method? I think it will result in stupidly complex areas of data, just to state that the land-use is what could be assumed anyway. (Assumptions made on a country level anyway) In an areas such as the Wakefield example I see it's reasons as it may be quite rare, but when it's everywhere, there is little benefit, and thousands of square miles of land will be glowing green. When I seek to find the benefits of this, I'm left a little stumped. The important data is the elements that divide up the landscape, so hedgerows and walls for example. Upping the detail level would be to split livestock from crop, which would be another mammoth task. I don't oppose effort; that would be stupid on so many levels. There is no reason why others can't do what I can't be bothered to do. I'm questioning the value of data like this, and how to realistically use it in areas which are 99% farmland.Ben 17:41, 27 January 2009 (UTC)
That's true. On the other hand: at least when the more distant areas - that is those where the ratio of farmland to forest gradually swings to the other end of that scale - get thoroughly mapped the tagging for farmland is likely to creep to really cover all those thousands of square miles... Adjusting the colors used might then make the map more visually pleasing. Naturally where forests or other natural features are less common, mappers will be happy for a long time adding just those. As to how to divide the area, I personally couldn't yet say anything other than "what looks convenient" - from convenient size ways (node count wise along the perimeter roads, all the way down to splitting by farm and by any impenetrable bushes if the mapper happens to live there. Maybe someone else has more experience on the matter. Alv 20:19, 27 January 2009 (UTC)
I wrote the little paragraph you're quoting there. I stuck it on this page mainly to prompt exactly this kind of discussion. You've raised many of the questions which are in my head Ben. Obviously I've phrased it to encourage people to follow Wakefield's example, but then part of me wants people to look at Wakefield and say "what the f***? Are we really going to do that for all the farm land?", because there surely must be some better technical solutions. If more people do following Wakesfield's example around the edges of more towns, then we'll surely see the question crop up more.
Better technical solutions? Part the problem is that drawing in landuse areas is pretty fiddly in the editors, and generally makes the data quite messy and difficult to work with. As discussed down the pub, editors could support some kind of landuse "layer", but the technical details of how that would work are yet to fleshed out.
The landuse problem crops up in all sorts of places of course. For example in Las Vegas I noticed that people started adding patches of desert, where it is deemed useful to show this (between patches of houses in outer suburbs [2]), but how to manage the fact that on all side of the city, it gives way to... the whole Nevada desert??
-- Harry Wood 23:38, 28 January 2009 (UTC)
For colour, I take that back, it's basically irrelevant. The renderers can sort that out, and hopefully it will be a very subtle difference between the default colour. I imagine if I tagged the entire county as farmland, the glowing green splodge in the middle of the UK would see the colour get changed soon enough.
The idea of a landuse "layer" was what I was thinking about today. Should the layers just be in the editors, rather than OSM's data? Ideally it would be possible to draw the closed-way in the normal layer, but when tagged with landuse= it would drop that data onto a separate layer. This would then keep the data easier to handle (as along hedgerows there would usually be at least 3 ways), but not make this an even greater task by making us have to draw anything again. Drawing again, on a seperate layer, also makes me wonder how it would be possible to keep data aligned at all.
Because of the enormaty of this task, if ever possible, I think a faster method for creating the way is needed. I'm wondering if it's possible when you have the field borders already mapped, then a tool similar to the 'fill' tool in an image editing program should fill as far as it can, and apon that 1 click it grabs the nodes touching the edge of that area and creates a way. Although..because a high percentage of fields (of which I've mapped anyway) aren't completely enclosed, individually, drawing single additional ways across gaps would be required additionally. Basically work with it as though it were pixelated, but take vectors from it.
In short, I can make an image of the landuse (crop/livestock) in photoshop, by editing a rendered image very quickly, but to put this into vector form field to field would be incredibly time consuming, so ideally a fill option with some method of blocking gaps would make this whole thing plausible.
Field names could be applied to areas rather than a node also; that's one advantage I can see in this. The tag should be for each field, rather than each farm, parish or county, so that it can take the name of the field with it. The fields can be in a relation for that farm. This would allow some interesting data to be used, although maybe not really map data, excluding if and when the field names are used as a destination/start point. This would take a huge amount of additional research to add, and upkeep though.
Minor point really in relation to katpatuka's point above;=, as I dislike debating word choice usually; Farmland is defiantly more appropriate than farm, since the 'farm' is the area of buildings from which the farm operates. As long as people know the tag choice though it doesn't matter much. Ben 00:34, 30 January 2009 (UTC)

Here's another example of a somebody tagging a very large swathe of land: near Stockton California. Big brown stain. -- Harry Wood 15:03, 31 March 2009 (UTC)

I'm another victim of the lack of policy on farmland. I live in an agricultural area where villages are effectively little islands in a sea of pinkish-brown. It does seem a waste of effort to draw farmland boundaries carefully following the outline of each village, and presumably it adds to the renderer's workload too. But I'm unable to just leave it alone, because I think we should know the difference between an area that has been positively identified as farmland and an area where nobody has looked. I would be happy with either of these two solutions:
  1. Draw a polygon enclosing an entire island, like Great Britain, or even a subcontinent, like Europe; tag it as farmland; and then let everything else sit on top of that. Pro: least effort for mappers. Con: no way to mark unvisited land; risks telling lies.
  2. Draw every field or farm separately, as it gets surveyed, and admit that we don't know about the rest. Pro: the most honest, accountable solution. Con: time-consuming, lots of data.
It's obvious to me, having set out the options like this, that (2) is the way to go for people who care about the accuracy of the data. The quantity of data, IMO, is a good thing, not a drawback. Who agrees or disagrees? --Jeff 16:54, 29 August 2009 (UTC)

A very similar problem arises for otherlanduse and natural tags. Some example are wood (http://bit.ly/PFFuX), water (http://bit.ly/Gk6dv), glacier (http://bit.ly/qD0gN), fell (http://bit.ly/49JUH) and I assume a few others. This is a largely unsolved problem so far, even if multipolygon might help somewhat. -- Gustavf 10:34, 30 September 2009 (UTC)

Yep and Las Vegas Nevada desert example I mentioned above. I've created a page Land use and areas of natural land, which is a tagging topic page (a bit like Railways) Eventually it should be a page giving answers, but currently it mostly poses questions -- Harry Wood 09:42, 27 April 2010 (UTC)

I think it would be a good idea to use default tags for boundary areas (renderers/tools could look at the top level down, so you can choose the level you wish to tag it). Such an idea with defaults has been suggested with other tags such as maxspeed in the talk SotM 2010 session: Tag Central: a Schema for OSM. - LastGrape/Gregory 21:32, 18 July 2010 (UTC)

I had a specific idea how the problem could be handled, introducing a new layer into osm. Travelling salesman 18:36, 9 July 2011 (BST)

I'd like to show you an example of how this problem is solved in Russia: Krasnodar Krasnodar. Mapped (and is still being mapped) by Kotelnikov. --Zverik 10:45, 28 December 2011 (UTC)

Marking areas used for gene-modified plants?

The European Court of Justice just anounced that public authorities must hand out information were gene-modified plants are being grown. (n-tv.de: "Behörden müssen informieren", German)

I think such areas should be marked in OSM, too. How about adding gene-modified=yes/no to the landuse=farm area? --Spartanischer Esel 09:49, 18 February 2009 (UTC)

+1 from me! --Lulu-Ann 13:09, 25 May 2009 (UTC)

Already used: gene-modified=yesedit

Why not turn it around too? Think they call it biodynamic agriculture where no artificial products are used (i.e. only natural fertilizers, no insecticides, etc.) biodynamic=yes/no --Skippern 23:20, 25 May 2009 (UTC)

Sounds good to me, maybe we can have it in one tag like: cultivation_method=bio/ordinary/gene --Lulu-Ann 08:51, 26 May 2009 (UTC)

Good idea! I'd propose changes in detail, though: cultivation_method=organic/conventional/gene-modified
Biodynamic agriculture seems to be only one kind of organic agriculture. --Spartanischer Esel 19:33, 26 May 2009 (UTC)
OK with me.

Already used: cultivation_method=gene-modifiededit --Lulu-Ann 06:53, 27 May 2009 (UTC)

-1 (sorry Lulu-Ann, it seems we never agree on tags). For me, this landuse shouldn't go into details about what plants or vegetables are growing in each parcel. Are you going to update them one or two times every year when it's changing ? It is simply impossible to maintain. You can do it for some permanent crops like vineyards or fruit trees but not more. -- Pieren 09:26, 27 May 2009 (UTC)
Could say the same about shop/restaurant mapping though :-) I think people should go for it and map gene-modified crops if that's what they're keen on doing (assuming they're not copying this information from restricted/derived sources "public authorities must hand out information" based on what maps? )
But I guess it goes without saying that people should go ahead and map. The tricky question is whether/how to document the tag. I think the tag could be listed on this page as a "possible tag to use in combination". It should not be listed on map features page, just on this Tag:landuse=farm page as a combo tag (of somewhat minor specialist interest)
-- Harry Wood 10:00, 27 May 2009 (UTC)
Well, continuing the comparison with restaurants, it's like writing the menu and price list into OSM. Going too far in details makes the maintenance and suitability more difficult. --Pieren 11:37, 27 May 2009 (UTC)
I somehow agree with Harry Wood on this, if people really want and have the resources to do so, than go ahead. It have to be clear though that most mappers will not go that much into detail. BTW: OSM can be used by official resources to map their needs too, I see no reason why the agricultural administration cannot use OSM to see where people grow organic or gene-modified crops, but than I rather see them do the maintenance of that data. Restaurants have a tendency to open and close, move around and so, same with clothing shops and much more. I have seen restaurants appear and vanish in less then a months time. That means we can have more than 12 changes to the same node in a year, so if maintenance is the argument, than stop mapping restaurants, shops, and other amenities that may change frequently. --Skippern 15:21, 27 May 2009 (UTC)
Mappers don't have to go that much into detail. It's a can-Tag, not a must. --Spartanischer Esel 08:29, 31 May 2009 (UTC)
Pieren, you are asking me "Are you going to update them one or two times every year when it's changing ?" - Yes, I do. I would like gene food farmland to be mapped on one day, burned on the next and removed from the map on the day after....--Lulu-Ann 15:06, 28 May 2009 (UTC)
It should also be noted, that a cultivation method is almost permanent if it's bio or gene-modified. It doesn't change every year - bio status, contracts, politics, the public ...--Landwirt 15:50, 28 May 2009 (UTC)
I agree with Landwirt, organic farms cover about 16% of the agricultural land in use in my area, Austria. But in other countries it is usually even much lower (like France about 4-6% I think) - so the maintainance seems more feasible.
Concering Europe, there is an ongoing dispute about gene-modiefied plants and with the exception of some corn plants most are illegal. There is also changing legislation about distance to normal farmland etc. So, having interesting future appliances in mind, I definitely would welcome a possibility to tag organic/conventional/gene-modified -HQQ

External links to places where gene food is grown

What about greenhouses?

That's it. How do we tag a massive area of greenhouses like this?:

Mar de plástico.JPG

This is near my city, it's called the "Plastic Sea", the only man made structure visible from the ISS (The Chinese Great Wall is not), so you can imagine its extension. Is it farm or farmland? --Schumi4ever 16:40, 26 June 2009 (UTC)

I've made the maths... it is 45km long by 15 km wide... in Spain, by the Mediterranean. --Schumi4ever 16:43, 26 June 2009 (UTC)

How about building=greenhouse. So that'll be the biggest building polygon in the database then. Spectacular! -- Harry Wood 16:48, 26 June 2009 (UTC)

That's a new tag, isn't it? That will require a new proposal... --Schumi4ever 17:05, 26 June 2009 (UTC)
Already proposed on Proposed features/building. Actually I didn't know that until I looked just now. building=greenhouse is a tag I naturally thought of in answer to your question.
I suggest you voice your support for that proposal, or for the greenhouse value in particular, but then go ahead and use the tag in your data anyway. People who find tags useful and use them in the data, are the main driving force behind wiki proposals (using the tags comes first)
It will actually get rendered too since Mapnik currently treats all building tags the same.
-- Harry Wood 17:25, 26 June 2009 (UTC)
+1 --Skippern 15:27, 27 June 2009 (UTC)
I suppose this is not one huge building, is it? I guess these are a great number of smaller greenhouses, right? So lots of buildings to add... --Spartanischer Esel 21:06, 26 June 2009 (UTC)
Of course, there are thousands of greenhouses, but I think that they can be tagged just as one big area, and the roads between them will separate them... --Schumi4ever 00:37, 27 June 2009 (UTC)
Having around 100 square km of greenhouses in my neighbourhood [see Google aerial] I am also quite interested in a solution. Drawing and tagging each and every greenhouse as a building is not feasable(there are thousands of them) and it wouldn't add much value IMO. A new value for landuse seems much more practical. I would simply suggest landuse=greenhouses (plural!) as I don't know any better terms for such an area. Polderrunner 20:01, 27 June 2009 (UTC)
Aha. If you don't want to trace around the actual outline of buildings then building=greenhouse isn't the right tag after all. It should be a landuse tag as you say.
I don't know the "correct" landuse tag to go for really. It probably depends on the outcome of Proposed features/Crop. There's a mention of greenhouses amongst that chaotic discussion, but it seems we're not even close to consensus on how to handle different crop types. landuse=greenhouses wouldn't fit well with the other landuse tags really (it's too specific about the type of building in the area rather than the use the land is being put to) I'd recommend landuse=farm along with a note=greenhouses as an interim solution until crop types is decided on (other mappers can see your note, and set the correct tag later)
-- Harry Wood 21:59, 27 June 2009 (UTC)
Sounds good to me... --Spartanischer Esel 19:42, 1 July 2009 (UTC)
I agree, landuse=greenhouses is not appropriate. So far, I have used landuse=farm, farm=greenhouses as an interim tagging scheme. However, yesterday I googled a bit around and found out that the correct term for cultivating plants under cover seems to be "greenhouse horticulture". Will make a proposal for this kind of landuse. Polderrunner 20:47, 18 July 2009 (UTC)

Farm vs field

[3] shows someone (not me) using landuse=field. What's interesting are what happens when two fields share a segment - you get a line. Someone's done something interesting in Mapnik for an unapproved tag...

Personally, I reckon fields should be white, and should be the default; what we should be mapping is field boundaries.

--RichardMann 00:51, 8 August 2009 (UTC)

As far as I was aware, landuse=field was a rejected proposal (Proposed features/agricultural Field) Rejected because it pointlessly duplicates landuse=farm. Not sure why mapnik is rendering it.
I guess maybe the whole point of the landuse=field idea was to capture field boundaries, in which case maybe it did have purpose. That was never made very clear on the proposal though.
We also have barrier=hedge, barrier=fence. Maybe this is a better way to denote field boundaries.
-- Harry Wood 09:50, 10 August 2009 (UTC)

cranberry 'bogs'

I live in Massachusetts, and I'm trying to figure out how to map cranberry 'bogs'. These are fields sunk lower than the surrounding land that are flooding in the fall and kept flooding over the winter, but then drained again before the plants are growing in the spring/summer. (see the cultivation section of wikipedia's cranberry page)

landuse=farm certainly makes sense, more so than natural=wetland, since they aren't natural. On the other hand, their effect on actually getting-around is that of a seasonal marsh (and in Massachusetts they typically freeze over in winter, adding another wrinkle). seasonal=yes probably makes sense.

i'm guessing rice paddies would have the same issues, but there aren't any nearby...

Morganwahl 04:49, 15 October 2010 (BST)

I think you maybe can use natural=wetland together with managed=* and seasonal=*, and as far as I know, natural=* can be combined with landuse=farm --Skippern 10:58, 15 October 2010 (BST)

tillage vs. grass land

"As a consequence the tag farmland was adopted to denote tillage and pasture" There ist a better tag, 'meadow' for pasture and hay land already -- which both are in some cases difficult to discern,.

Why not use farmland just for acre/ tillage? (which is easy to differentiate from grassland) Official maps here showing grassland different than acre. They distinguish between the two since more than 200 years (Tranchot-Müffling), for a reason! At least in summer they can be distinguished in nature. This makes more sense than mix it alltogether just because farmland looks like farmland to some urban mappers.

Please use farmland exclusively to tillage, or you will need, sooner or later, something like 'field" and 'farm/ farmland' will become obsolete because of it's indecisiveness.

Please also remove the picture. Probably it shows growing corn but so young that it can't be distinguished from young grass. Taken two weeks later it would have been clear what it is. A picture should explain what is written in the article. This one confuses.

Unlike, 'farmyard' is a very usefull tag. It shows the area of a very specific farm. If you tag 'farmland', please also tag the 'farmyard' it belongs too. And the service road belonging to it.--Taunide 11:02, 8 March 2011 (UTC)

natural=* and landuse=* aren't mutually exclusive. All land that is used for farming, and outside of the farm yard, that is tillage and pasture, is such. Forget for a second the rendering, and consider the fields (be it for crop or for pasture) around a farm yard: they're all part of the farm, hence landuse=farm. Trying to find out the exact difference between the German words Wiese and Weide, I believe that's where this stems from: there are three kinds of areas: fields for tillage (Acker), fields growing meadow-like and used for pasture (as wiktionary reads "Wiese, auf der man Vieh hält"), and areas that naturally grow, or are made to grow mostly flowers and grass and might be (but are not necessarily) cut annually for later feeding to cattle. The first two kinds are definitively used for farming, and hence worthy of a landuse=farm. You can describe the natural element, or the use with additional tags. Alv 15:28, 8 September 2011 (BST)
not so sure about that. landuse=* is intended for managed land, natural=* for unmanaged. --Taunide 16:47, 11 September 2011 (BST)
You are mistaken, then. The first and most simple example that comes to mind: you can have an area of natural=scrub inside a landuse=railway (or just about any landuse). Natural describes what elements of the nature exist at that area, and (most of the time) totally independent of that the key landuse describes to what use are the people putting it. Or think of a recent clear cut forest: landuse=forest + natural=scrub. Alv 21:23, 11 September 2011 (BST)
Some mapper may see (means: tag) very small pieces of scrub, whereas some mappers will see the railway (or forest). Both may be right at some point, but this don't mean that both tags are equally important for a certain map.... My idea was about larger pieces of land here, and that a *map* should classify pieces of land how they looks like, primarily, not what it logically belongs too, or not belongs too, in a enhanced and theoretical denotation. After all, I understand OSM work as working on a map. A mapper very seldom is able to judge if a piece of grass land belongs to a farm, grazed by livestock, cutted by machine once or twice a year, or if the grass grows naturally, grazed by wild animals. So it simply don't make sence to distinguish between these uses in *maps*. Whereas ploughed land looks quite different from it and is not "natural" in any case. you may know that tillage and grassland are shown different in topographical maps since ~200 years for a reason, right? --Taunide 21:06, 12 September 2011 (BST)
Yes and no. Working on osm is working on a database that can be used, amongst other things, for making maps, but not directly drawing maps. If we describe land use as "used for farming" and use some other tags to make a difference between tillage and other contents, one can make a map showing the difference. At least here most pastures and hay meadows look more similar to a tillage field, than to a naturally occurring (or transitional) meadows. If it weren't autumn already (everything has been harvested already, and it's raining), I could snap a photo... There has been some discussion around the issue at Talk:Tag:landuse=meadow, and even those come to the conclusion that the tagging has been messed up. Things used to be consistent, but nobody documented stuff thoroughly and explained to nonnative English speakers what the keys should be. Then someone added rendering rules for landuse=meadow, and others used it to the goal you're repeating here, when the proper way would have been to discuss and file a ticket at Trac. Alv 17:59, 15 September 2011 (BST)