too much water

When viewing the OSM map for colorado there is way too much water shown. There are many small creeks that just do not exist. It looks like the water features were autogenerated somehow and are just not right. Is there some way reduce the water detail and increasing the road detail? I’ve downloaded the tiles and opened them in qlandkarte and also viewed the map in the browser.
Is it possible to add contour lines?

I’ve noticed lots of water where it isn’t in the mountains near me too. Looking at it with JOSM I see that they are typically ways marked as intermittent streams. The old USGS topo maps would render that such that it did not look “too wet” but the standard OSM render seems to think that it ought to rendered as a regular stream which is pretty misleading.

I’ve left the ones I am familiar with as is as “tagging for the renderer” is frowned upon and they really are intermittent water courses. Maybe filing a bug against the standard OSM style sheet would be the way to go so the default presentation would look less wet.

I’ve come across lots of references in the OSM wiki, the help forum, mail lists, etc. that give reasons for not adding contour lines to the OSM database.

For me the main one is that they are only one of several ways to attempt to represent elevation information. So you’d need a consensus as to the method to do that. If contours would be accepted then we’d also need to standarize on the intervals (10 meter, 40 foot?), how many intermediate contours to put between index contours, etc. Topping it off, elevation data is freely available from other sources along with tools like those in the gdal library that allow you to easily make contour maps with index and intermediate contours at the separation you desire with the units you desire. There are some cookbook examples around the web on how to create geotiff files for contours, hill shading, slope shading, hypsometric tinting, etc and using them with Mapnik to create finished maps.

It looks like most of the streams in Colorado have been imported from the National Hydrography Dataset.
I don’t know how accurate any of that data is, or whether those streams actually exist.

There are a variety of different styles of maps based on OSM. Some of them highlight particular things, so some show roads, and not water etc. eg look at Mapquest Open or Cloumade have a range of styles
Or even more options here:

Or you can produce your own maps, adjusting the style to look how you like. This is known as ‘rendering’, you can start with this as a guide:

Yes, as n76 says, contour lines/elevation data is not usually part of OSM. But you can get the data from other sources, and combine it with OSM to render maps. There are a few maps doing this, eg OpenCycleMap

What is being shown is not even intermittent. I see that one of the codes in the national hydrography data is ephemeral stream which shouldn’t be shown on any normal map. I wonder if that is being shown. I’ll see if I can find the suggestion box or bug report & file something.

Thanks for the links, I think the OSM cycle map is pretty close to what I want. They have done much better with the water.

I imported most of the Colorado river basins from NHD, starting with Summit County as I was interested in seeing ski lines and in the absence of contour lines the hydrography is a good substitute. AFAIK I kept significant NHD tags on the ways I imported, but I do not recall any stream or river features corresponding to the ephemeral class. A quick check suggests that 46003 feature classes should be tagged intermittent=yes, and many are not.

From a conversation a couple of years ago, the NHD dataset is clearly rather old. IIRC there were some reservoirs around Grand Junction which were filled with scrub (probably removed now), one at least was a significant site for migrating birds as a non-reservoir)

The major issue to my mind at the time of import were irrigation channels (often covered) which have been tagged waterway=canal. These give a very misleading view of the actual situation. To date I don’t think OSM has resolved a decent tag to separate off irrigation channels from major irrigation canals.

The cycle map actually shows the same water features (at least to z16) but with much less emphasis. For some reason they seem to disappear at z17 and z18 at the same time as some forests appear, see A-basin for instance.

I think I retained the data used to do the import so could check on the NHD attributes. You could also cross-check with what Lars did with TopOSM ( as he used the same NHD data, but rendered intermittent waterways with a dashed blue line. When you look at TopOSM with contours and hillshading it is obvious that even intermittent water courses have a powerful effect on the landscape. Noting their presence is probably highly important for hydrographic purposes, although no-one is yet using OSM data for hydrographic simulations.

Regarding adding contour lines and hillshades: Please look at this project: We do it :slight_smile: