So probably the best place for it would be a separate database that could be combined with OSM data. There is no reason why it couldn't use OSM format and tools such as JOSM though.
I worked in a library for a while and we had a theory that if you asked five classifiers how to classify a book you'd get six different answers. Cheerio John On 14 March 2015 at 13:38, Charlotte Wolter <techl...@techlady.com> wrote: > Sander, > > I agree with most of your points and would like to add that > surface geology is a highly specialized field requiring a great deal of > expertise. I'm a geology buff myself, and there is no way I would attempt > to map that. Also, there often is strong disagreement among geology > professionals about the nature and dating of rock units, disagreements that > make some of our set-tos about how to code sound trivial. > Also, there generally is a lot of local information about > landslides and tsunami risk, courtesy of the USGS, though sometimes it is > ignored. In Los Angeles, tsunami-prone areas are signed along major roads, > as are the areas subject to debris flows. The recent deadly landslide in > Oregon was in an area known to experience landslides, but apparently the > risk was not widely publicized. > Tsunami risk, perhaps, could work as an overlay, and I believe > that data is available from the USGS. > But, generally, I think this whole area may be too technical for > widespread application in OSM, even though I would enjoy seeing it. > > Charlotte > > > > At 05:59 AM 3/13/2015, you wrote: > > I think this suggestion belongs more on the general OSM talk or tagging > list than on the HOT list, but anyway. > > There are already a number of ways to tag surface, like surface=*, > natural=*, landuse=*, landcover=*, ... Just read the wiki about those (f.e. > http://wiki.openstreetmap.org/wiki/Key:natural ) > > There's also a convention in OSM about sub-tagging. F.e. you could tag > > natural=rock + rock=sandstone > > Thus I guess most of what you want is already possible in OSM. You should > only try to add a few more specific conventions (f.e. about the types of > rock). > > I probably don't really get your 3D attempts, but the general concensus is > that it's hard to get in certain places, and thus you can't make a uniform > map of heights or angles. As such, OSM contains no height or slope data > (apart from the elevation of some peaks), but leaves this to professionals > (such as the NASA). It isn't so hard to extract a general slope from good > precision elevation data, so there's no point in including it directly in > OSM data (with the right preprocessor, it can get rendered on the map > anyway). > > So that doesn't belong in OSM, but it isn't the biggest problem IMO. The > biggest problem I see in your attempt is ignoring that OSM is a > crowdsourced effort. For crowdsourcing, you need a crowd, and that crowd is > most easily found in populated places. Your effort seems to focus on areas > with a low population (a city isn't very vulnerable for a landslide). But > sadly, there's no crowd around there, so the most we would be able to do is > some mapping from aerial pictures. This shouldn't hinder you from starting > the project, but you shouldn't have very high expectations from it. > > Regards, > Sander > > > > 2015-03-12 22:03 GMT+01:00 Hazel <hl...@srcf.net>: > Dear All, > > Can we again discuss putting geological data into OSM? Specifically, I'd > like a recommended way to tag fault lines and surface geology polygons. > > This e-mail assumes the reader knows nothing of geology, apologies to > everyone else. > > First, the usecase: geological data saves lives in natural disasters, it > is useful for common activities like agriculture, and it is interesting in > its own right. It can also be usefully collected by amateurs. > > I am not suggesting that OSM should produce disaster risk maps, or > recommendations for farmers. I am saying OSM could collect the data that > would allow experts to quickly and easily make these things. > > Using OSM contours, they can work out areas of flood risk and tsunami > escape routes. Using contours and and basic geological information, they > can work out areas of landslide risk (landslides kill more people than > volcanoes or floods or earthquakes, but they kill a few dozen at a time). > If we map faults, they'll know more about where earthquakes are likely to > happen (you know the photos of roads after earthquakes, offset by a few > centimeters? The fault is the plane where the offset happens, and > earthquakes use the same faults over and over again). If you map areas of > shallow bedrock vs. unconsolidated sediment, you know which areas may > suffer soil liquifaction in an earthquake. > > https://en.wikipedia.org/wiki/Soil_liquefaction soil liquifaction > > Technical infodump: > > To make a geological map, you map areas with similar surface rock or > sediment2. You describe them (anything from field IDs like "greenish rock > #2" to detailed technical descriptions) and give them proper names (e.g. > "the Tunbridge Wells Sand Formation"). > > Having mapped the boundaries between different rock types, you can also > trace faults and the line of folds in the rocks. These all obviously exist > in 3-D, but are usually represented on 2-D maps. Just mapping the 2-D trace > is enough for many purposes. > > OPTIONAL EXTRA 3-D info: > If you want to add more information about the third dimension to a two-D > map, there are conventions for that. You specify a line (along the axis of > the fold, or on the steepest line down the fault plane or boundary plane). > You map the direction of this line. Then you measure the angle between the > line and the horizontal, and write in on the map (next to standard symbols: > for a plane, a T-shape, and for a fold axis, an X with two or three of the > lines turned into arrows pointing in the two or three downhill directions). > > Plane: > http://web.arc.losrios.edu/~borougt/StrikeAndDip.jpg > > Fold: > http://bc.outcrop.org/images/structural/press4e/figure-11-16b.jpg > > Planes on either side of a fold: > http://courses.missouristate.edu/EMantei/creative/GeoStruct/strkdip.jpg > > This is actually fairly easy to explain in 3-D, but not in 2-D, and I > don't know of a good video. We could make one. > END OPTIONAL EXTRA > > > Example: > Let's look at the Weald area of the UK, since it is well-mapped. > > Read: > https://en.wikipedia.org/wiki/Weald#Geology > > Terms: > "Lower Cretaceous" and "Upper Jurassic" describe age (lower means older) > "rocks", "chalk" and "sandstone" describe rock type > "sands" and "clays" describe sediment type > "Purbeck Beds", "Ashdown Sand Formation" and so on are proper names of > groups of rocks/sediments. These names are hierachical, like taxons, and > are in databases (for the Chalk Group that forms the White Cliffs of Dover: > http://www.bgs.ac.uk/lexicon/lexicon.cfm?pub=CK). > > The cross-section may help make the 2-d map make sense. > > To see how faults and folds (synclines/synforms, that sag, and > anticlines/antiforms, that hog) are mapped as lines, see this map: > > https://commons.wikimedia.org/wiki/File:Geologic_map_SE_England_%26_Channel_EN.svg > (just gives rock ages, not type). > > Faults are usually much more obvious on small-scale maps than they are on > this map. > > For sediments, there exist multiple soil classifications, with mappings > between them, and OSM could support them all, but the classes we have > (sand, gravel...) would be enough to start with. > Examples: > https://en.wikipedia.org/wiki/Australian_Soil_Classification > https://en.wikipedia.org/wiki/USDA_soil_taxonomy > etc. > > QGIS is increasingly used for geological mapping, so it works increasingly > well with many other geological tools. QGIS is already well-integrated with > OSM. The barrier for geologists new to OSM to upload their maps is > therefore low. Classes of students could do it. > http://wiki.openstreetmap.org/wiki/QGIS > > End infodump, requests for clarification and corrections welcome. > > Could anyone suggest a set of minimal changes that would make it possible > to enter data like this? As I said, just having a recommended way to enter > a surface geology polygon, a geological contact line (between two > polygons), and a fault line (with optional dip direction and inclination) > would be very useful. > > Pseudo-3-D perfection would also allow keeners to input the contact > between two rock formations (line, with dip direction and inclination) and > input folds (line, with dip inclination), but this can also wait. > > Regards, > Hazel > > > _______________________________________________ > HOT mailing list > HOT@openstreetmap.org > https://lists.openstreetmap.org/listinfo/hot > > > _______________________________________________ HOT mailing list > HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot > > Charlotte Wolter > 927 18th Street Suite A > Santa Monica, California > 90403 > +1-310-597-4040 > techl...@techlady.com > Skype: thetechlady > > > _______________________________________________ > HOT mailing list > HOT@openstreetmap.org > https://lists.openstreetmap.org/listinfo/hot > >
_______________________________________________ HOT mailing list HOT@openstreetmap.org https://lists.openstreetmap.org/listinfo/hot