Flooding on land in Alaska

A couple of years ago I started making maps for Garmin with OSM data from geofabrik mostly, occasionally protomaps, and compiling with mkgmap. I distribute these maps on another site. Alaska from the beginning has been a challenge. Parts of Southwest of Alaska have had excessive water, or flooding.

At first I used the most recent, at the time extract, but was unhappy in the visuals. At one time I had downloaded and compiled everything available and only one didn’t have the flooded look, that one from January of 2019. I was able to merge in missing features converting shape files with josm and gpsmapedit to pbf’s and mp’s so all was good. I’m looking to update and recompile with recent data but at present the flooding is still occurring.

The first image, a screenshot from opentopmap shows the area that is flooded, the second is a section (I have tried with and without compiled sea, with and without typ and the results are the same, the third is the map I have been distributing, all these shots were in basecamp on the highest detail.

I’ve looked all through josm (with small sections), associated tags, comparing old and new, etc and have been unable to find what I need to solve this.

At present I would just like to do the edit(s) for my purposes on my computer, any uploads can wait.

Is there something I’m missing or I need to look at to correct this. I personally would be embarrassed to release my maps with this much, or any flooding, but apparently opentopomap and some other international map makers and distributors have released maps with the same issue. I will note this area may be boggy and have some swampy areas, but not this much!




When I saw you were talking about flooding, I immediately thought that it would be an issue related to the sea/coastline, which can often be an issue. However, once I saw your screenshots, that isn’t what’s happening.

If you look at the map here, you can see that this is all data that has been added to the database. It looks like a user imported a bunch of data from the USGS National Hydrography Dataset in July 2020. I was a bit skeptical of the accuracy of this data at first, but aerial imagery does appear to confirm that there are a lot of small waterbodies in that area.

What you might need to do is modify your mkgmap style to omit some of the smaller waterbodies at certain zoom levels. If you look at the documentation, you could use a comparison (section 4.2) to check for the size or length of an object (4.2.4), and only display the larger ones at lower zoom levels. You could leave the smallest ones to only display at the highest zoom levels. You might have to play around with the comparison values to find a good level at which the most notable objects still display, but without so much of the small stuff.

Alaska is full of areas like this, almost more water than land, especially in the region you’re referring to. Those small lakes and ponds (“potholes”) are leftovers from glaciers that once covered Alaska. I think they are probably accurate.

Adjusting your rendering rules is the way to resolve this.

Thanks alester for responding. I had considered messing with the style sheets, but could end up omitting small but important/significant water bodies elsewhere in the state. The information about the USGS datasets has given me some ideas though about deleting specific objects that were tagged by “the user” in josm, time for some patience and zen, I’ll dig in when I have some time.


Well, I used a bit of logic and a lot of trial and error. What I did is open each affected split in the Bethel/Kusko region in JOSM (there were 9 or 10) and searched “nhd” (National Hydrography Dataset). I edited the key “natural” and changed the value from <different, to water, edited the key “type” from <different to multipolygon, saved and moved on to the next tile. On the positive side, it looks better, to me at least, at the mid zoom levels and flows with the rest of the state. I have messed around with the style and get about 95% of the water features at level 24. The “downside” the 5% the is not displaying small connecting streams and tiny lakes are not showing, though I may experiment more with the style. The bigger issue is I lost the name when hovering about above major rivers like the Kusko, but should be able to edit the name back in.

When user **** uploaded the data set it helped to nearly double the size of the extract. I’m not sure if they should of changed the tagging or if it’s at the end user level to deal with. In the end, I’m happy how it worked for me and if this information can assist others, great.


I only took a quick look, but the data appeared to be appropriately-tagged and accurate. It just happens that this part of Alaska is very wet, so mapping that “wetness” takes a bunch of data to do so. At its heart, OpenStreetMap is just a big database, and it’s up to the end-user to use that data how they wish. I’m glad you found a method that works for you to sort through the data you don’t need.

Potentially, but a lot of systems are in place for flying stuff up. Places like Tuktoyaktukin Canada fly in lots, Anchorage already has a massive airport.