Proposal talk:Dry lake

From OpenStreetMap Wiki
Jump to navigation Jump to search

Difference from natural=water + intermittent=yes?

The proposal suggests that this tag is needed for features different than natural=water + intermittent=yes "because they are almost never wet, filling with water for only a few days a year or less."

However, many dry lakes with that description are currently mapped as natural=water + intermittent=yes in the western USA and western Australia, as well as in other arid regions.

How many few days a year would a lake need to be filled with water to be considered a natural=dry_lake instead of natural=water + intermittent=yes? How would mappers verify this information? Often arid regions have very different rainfall from one year to the next, so in one year a dry lake might be full of water for 2 months, while the next year it might never fill. --Jeisenbe (talk) 23:21, 2 August 2020 (UTC)

"Lake Mungo" in Australia has not had water in over 10,000 years... ok it gets damp .. but not water covered. For seldom water there is, note this defines how often the water is present on an average. Mappers verify information in various was, one of them is personal knowledge. Warin61 (talk) 07:51, 22 February 2022 (UTC)

Perhaps we need a tag for salt flats or playa instead?

The two dry lake examples on the page show a flat surface of white salt (I believe). Perhaps we need a tag for this (a playa or salt flat or salt pan), rather than changing the definition of natural=water + intermittent=yes?

E.g. natural=salt_flat instead? --Jeisenbe (talk) 23:23, 2 August 2020 (UTC)

I have used natural=water + intermittent=yes + ephemeral=yes + surface=salt Warin61 (talk) 07:54, 22 February 2022 (UTC)


Looking around at different salt flats, such as bonneville salt flats in Utah, the natural=desert tag is already under heavy use for large dry lake beds and other dry flat barren areas around the world. --Sytys (talk) 14:42, 6 June 2022 (UTC)

Necessary distinction from lakes that contain water

If you look at Bristol Lake and Cadiz Lake near the Mojave Trail National Monument, these features have never had standing liquid water at the surface. The features are formed by the evaporation of brine from the groundwater 150 feet below the surface. The surfaces are crusted with minerals from the brine, including salt, calcium chloride, and gypsum.

Areas like this have been variously tagged as:

Each of these is incorrect in its own way.

Using the natural=water tag on a feature that never contains standing water is trolling map users. Adding intermittent=yes or ephemeral=yes to negate the meaning of the main tag for the feature is exactly the definition of trolltagging. While it may be the case that small areas of standing water appear in some of these locations after periods of heavy rainfall, these are temporary conditions that aren't permanent enough to be mapped. For a dry lake, it is typically the case that the area never completely fills with standing water and the locations where water pools vary from season to season.

The verifiability of the natural=desert tag is disputed, for good reason in my opinion. It is not clear whether this tag refers to a surface feature, geology, or ecology. And the criteria for applying this tag are unclear. However, the tag is typically rendered as sand or waves of sand, which makes it attractive as a form of tagging for the renderer. The Wiki description suggests its use for dry lakes and playas, which is somewhat misleading. Although dry lakes and playas occur in arid regions, the are typically very small portions of a larger desert area. And a dry lake or playa is distinct from the surrounding desert because of the nature of the surface (typically minerals, clay, or silt) and the lack of vegetation. (Contrary to some popular imagery, deserts do indeed have bountiful vegetation.)

The natural=mud tag may be appropriate in some cases. For example, Death Valley Playa (aka Badwater Basin) is an area of salt crust over mud. The ground in this location remains wet throughout the year with water supplied by nearby springs. Whether the natural=mud tag is properly applied to the nearby Racetrack Playa is perhaps a matter of debate. Is this area truly "mud" if it is normally completely dry and only partially wet immediately after periods of heavy precipitation?

However, in many cases, the use of natural=mud appears to be a choice that mappers make given the alternatives that are typically rendered. In the examples of Bristol Lake, Cadiz Lake, and Danby Lake above, the soil at the surface is never wet enough to be considered mud. However two of those three locations are tagged as "mud." It's possible that this is a choice that mappers make as the best of three poor alternatives.

In most cases, dry lakes and playas are similar enough to be covered by the proposed natural=dry_lake tag. This tag may also be appropriate for some salt flats and salt pans. However, there are also salt flats and salt pans adjacent to perennial or intermittent bodies of water, e.g., the salt pan adjacent to Laguna del Negro Francisco. These salt pans likely deserve a different form of tagging that is more specific to the type of feature.

In summary, I am in favor of using the natural=dry_lake tag in places where it is appropriate for features that normally consist of dry beds of salt, clay, or silt, commonly in endorheic basins, even if they may be temporarily or infrequently wet. I approve of the suggested rendering for the tag, but might suggest an outline of light blue to distinguish the rendering from areas of sand. Where the natural=dry_lake tag is used in combination with surface=salt the rendering might use a white pattern for the area of the feature.

--B1tw153 (talk) 16:03, 7 July 2023 (UTC)

Include sabkha as an option?

Several examples shown and described in this section would better be termed a sabkha. These areas are common and extensive in Gulf Coast Countries (GCC) such as UAE, Kuwait, Qatar, Saudi Arabia, Oman etc. I have been using a tagging scheme in Qatar specifically to overcome a few challenges:

  • Qatar has no natural continuous flowing fresh water or lakes
  • The interior has numerous dry lake beds which do fill with water for between hours and weeks, depending on rainfall and weather conditions
  • These dry lakes form smooth hard ground and can be sites for plant growth (scrub) and grassland
  • They can be dangerous for vehicles as they become almost incompetent mud which will trap even the most robust offroad vehicles
  • These areas can develop into sinkholes as they concentrate fresh water which dissolves the underlying limestone rock and forming karst

Closer to the coast, the broad flat areas inundated by seawater form a proper sabkha which could be described using wetland=tidal flat although the wiki describes these as being areas for animal life such as crabs. These areas are frequently affected by tides with some having direct connections to the sea via tidal channels and others being filled from below via shallow groundwater breaking through. The Sealine Dune System is bordering on an erg and has numerous pockets of sabkha. Many of these are traps for dune bashers in offroad vehicles navigating the dunes. Currently, the option is for mapping the pockets as dry_lake (which misses that these are water filled or muddy often enough to be a hazard) or tidal flats:tidal=yes which causes rendering problems as tidal flats are meant to be offshore and beyond the coastal boundary.


  1. I use dry_lake for features not connected to tidal influence but dislike the current situation where these do not render on my Garmin device
  2. I alternate between using wetland:tidal flat for areas that inundate regularly with tidal influx although this is inside the coastal boundary
  3. Mud for areas that may only have water infrequently but remain a hazard to vehicles for extended periods

Lake Mead high water mark

I have edited the high water mark of Lake Mead in Arizona / Nevada, USA to be a dry_lake, which in my opinion is more correct than the previous type of water=lake. This area has not had water in 30 years, but renderers continue to draw it in blue.


Kevinp2 (talk) 03:05, 13 February 2024 (UTC)