Apologies for posting in English.
I have just been doing edits from my journeys through Patagonia after State of the Map, and am somewhat dismayed by the presence of large polygons which bear little or no resemblance to what can be seen on the ground. I am aware that some of these are discussed in other threads, but for convenience I have noted my major issues with what is currently mapped here.
There seem to be three major classes of polygons which at best have incorrect tagging on them:
-
**Woods. **Major river valleys tagged as natural=wood. A single polygon encompasses some 100s of kilometers of the basin of the Río Coig and another the Rio Belgrano. I imagine these were created from satellite imagery in the mistaken belief that greener areas are woods. There is also another amorphous polygon of natural=wood around Sarmiento, Chubut, and probably others which I have not come across. I am in the process of adding Mapillary sequences between El Calafate and Rio Gallegos which show clearly that there are no woods in this area. These polygons are anyway far too large for sensible mapping and create problems with on-line editors.
-
Estancias in Chubut. This has already been discussed at length. My view is that they should not be tagged as landuse=farmland. The use of the tag is now largely restricted to arable crops, with other tags for intensively managed pasture (such as landuse=grassland) and for grass crops (hay /silage - landuse=meadow). Extensive areas of natural landscape used for pasturing animals should use natural tags (steppe or scrub), which is exactly what we do in the UK for areas where sheep are farmed in the uplands. It is interesting to note that the one area in Chubut noted for intensive agriculture, the irrigated area around Gaiman, is, ironically not mapped as farmland. Secondly these polygons are not very accurate, have overlaps and underlaps, so their quality is also low. This can be seen on Peninsula Valdes where I have mapped cattle grids on the Ruta Provincial which ususally correspond with boundaries of Estancias. Thirdly, I think it would be more useful to map the central habitation of Estancias using either place=locality or place=farm. Fourthly, some kind of tagging scheme needs to be agreed for mapping the full extents of Estancias. I noted with interest a map in the museum on Peninsula Valdes dating from the 1940s showing the Estancias in the area, and also saw one showing all the Estancias in Argentine Tierra del Fuego. Lastly, there is, of course, the licence issue.
-
Industrial polygons for Oil exploration areas. Large polygons delineating oil exploration areas. These make use of the main OSM site very hard for cities such as Comodora Rividavia which sit inside these areas. Once again, I feel that these have been mis-tagged, and if they are to be mapped then a discussion needs to take place as how they should be tagged. They clearly should not be tagged as landuse=industrial, which should be reserved for observable oil & gas extraction and transportation facilities (many of which can be readily found on aerial images). (One side note I see a lot of oil pumps have been incorrectly mapped with power:generator tags, they should be man_made=pumping_rig or similar - http://wiki.openstreetmap.org/wiki/WikiProject_Oil_and_Gas_Infrastructure#Well_tags)). I would actually question whether this data belongs on OSM at all: is it verifiable on the ground?
Taken together these degrade the quality of mapping in Patagonia. This is a real shame, when one looks at places like Caleta Olivia (http://osm.org/go/Jfv~gUjg-?layers=N), which have been mapped extremely well by local people (or perhaps one local enthusiast). In fact, every populated place I visited in Patagonia is mapped to a high degree of usefulness on OSM.
For me, one of the most important things I learnt at SotM14, was how useful OpenStreetMap can be for all sorts of organisations in the “Campo” throughout Latin America. The ability to build up really useful information in sparsely populated areas will be a significant factor in extending the existing community. People need a “Wow” factor, which comes from things like " even the emergency telephones and road distance markers are mapped in the middle of nowhere", not “it’s useless, you cant trust the data because someone thinks there are woods on the Patagonian steppes”.
There is still lots and lots of stuff which can be used to add information from aerial photos, outside the cities, towns and villages. Notably many major rivers are still missing, but there’s a lot of oil/gas infrastructure, power lines, and even fences, all of which help build up detail in a useful way.
So in conclusion: many of these polygons should be removed for multiple reasons, those that are retained need to re-tagged to reflect the more common usage for these tags.
Let’s make sure we keep the same standards for mapping in the countryside as in the cities.
Jerry (SK53)