I’ve been working on & off on this project for a couple of years, but now I really want to get it done with that, so here I am writing in this community about my project to import Wetland ARCGIS Data for the Carteret County in North Carolina.
Why
The reason why I decided this area of the USA is because it’s my favorite place to travel and I wanted to get more accuracy on the map. It started by simply improving some of the existing shapes, but I found out that the county has a great GIS service with a lot of data for the Wetland. Wetlands | Carteret County Open GIS Data Site (arcgis.com)
Steps taken
Back in 2021, I wrote to them about those datas and they were ok for me to use them in OpenStreetMap since the data are Open to the public. I didn’t get a proper approval at this time, but it’s something I will do before starting the importation.
The accuracy difference from the current data is massive.
Right now I’m making sure I’m using the appropriate tags, deleting some of the source’s tags that aren’t relevant to OSM and I need to do a little bit of cleanup and delete useless shapes and shapes I’m not sure what it is. Some tags are irrelevant, but sometime it can be used by boats to know where it’s more shallow.
Currently I have a mix of tidal flat, shoal, bogs, swamp and almost everything related to wetland.
I’m using JOSM and I will import one little area at a time to not make mistakes. I will do that by merging my new shapes with the current one (if there is an existing one).
What I need to do
Getting a proper approval to use the data (even if they didn’t rejected me when I wrote I wanted to use their data for OSM). Is this “term and condition” page enough? GIS Terms and Conditions (carteretcountync.gov)
This looks like an noble and valient effort to improve what can be a challenging and daunting task to map manually. An entire scientific area of study exists around the classification and delineation of wetlands. Why reinvent the wheel is how I’d approach an import like this one. An expert has already delineated the data, so I say have at it.
Some years ago I began a similar import and although it was not finished, I think you would find some utility in the tagging research I had come up with. If you are successful here, I may be inspired to resume my work!
Most useful, I think, would be the conversion of the Cowardin wetland classification system to the OSM wetland tagging structure. I propose the following based on the ATTRIBUTE attribute in your dataset:
Wetland Type
MapCode
cowardin:description
wetland:description
OSM Natural
OSM Detail 1
Estuarine Deepwater
E1
Estuarine subtidal water and wetland
Open water estuary, bay, sound, open ocean
natural=water
tidal=yes
Marine Deepwater
M1
Marine subtidal water and wetland
Open water estuary, bay, sound, open ocean
natural=water
tidal=yes
Estuarine wetland
E2
Estuarine intertidal wetland
Vegetated and non-vegetated brackish and saltwater marsh, shrubs, beach, bar, shoal or flat
natural=wetland
wetland=saltmarsh
Marine wetland
M2
Marine intertidal wetland
Vegetated and non-vegetated brackish and saltwater marsh, shrubs, beach, bar, shoal or flat
natural=wetland
wetland=saltmarsh
Lakes
L
Lacustrine wetland and deepwater
Lake or reservoir basin
natural=water
water=lake
Freshwater Emergent wetland
PEM
Palustrine emergent
Herbaceous march, fen, swale and wet meadow
natural=wetland
wetland=marsh
Freshwater Shrub wetland
PSS
Palustrine shrub
Forested swamp or wetland shrub bog or wetland
natural=wetland
wetland=bog
Freshwater Forested wetland
PFO
Palustrine forested
Forested swamp or wetland shrub bog or wetland
natural=wetland
wetland=swamp
Freshwater pond
PUB
Palustrine unconsolidated bottom
Pond
natural=water
water=pond
Freshwater pond
PAB
Palustrine aquatic bed
Pond
natural=water
water=pond
Riverine
R
Riverine wetland and deepwater
River or stream channel
natural=water
water=pond
In addition to the reclassification, I would advocate for preserving the ATTRIBUTE tag as a new tag we could invent. Something like, wetland:cowardin=*.
According to the above:
“This data is subset of the USFWS Natl Wetlands Inventory.”
Which means you could go directly to USFSW and get the data and not have to worry about the license as USFWS data is public domain.
Just because there is more data, doesn’t mean it is more accurate. It might be, but we shouldn’t assume so. I would be interested in an actual assesment of its accuracy.
That is very important. It is important not to wholesale wipe out existing data and replace it with this data. For one thing, the existing data may have some tagging that the new data does not. e.g. the name of a wetland. We also want to preserve the history of the objects currently in the OSM database when appropriate. In JOSM you can use the replace geometry function to preserve the history.
Your table is very helpful, because I was using a similar one, but without the OSM equivalent, so it was mostly me trying to figure out which code means what.
I’m using a different file that I don’t remember where I found it.
The data found on the USFWS website here National Wetlands Inventory (usgs.gov) is taking a much larger surface area, but most shapes from this one and Carteret county GIS are the the same. I think USFWS is adding docks, E1 and some more details, but they are generally the same.
In my example above, I used the 2021 data which included docks area and E1, but it looks like they removed them from the 2022 update.
They also have the ditches from the farm land up north, but I they are already in OSM, so I not do anything with them.
I will concentrate on the things that aren’t already there, update/merge current shapes and delete the ones that are not relevant anymore.
E1ABL which may be irrelevant for OSM. Not too sure about this one. I may only work with what’s on the Carteret County GIS which may make my work easier.
I will also simplify the ways and shapes for OSM, because some are too detailed.
Tidalflat works for those submerged ones. Underwater mapping in OSM is not well established. You could review those manually and decide to delete if they’re usually fully submerged.