I am trying to import a shapefile (kml) that contains our university’s three campuses for a bioblitz that will take place this September (Universite de Moncton). However, I am running into issues. I managed to create the shapefile in QGIS and the polygons look fine both in QGIS and Google Earth. However, once I visualize the shapes in iNaturalist, the shapes do not fit with what I have imported. They have been simplified in a way that is not true to the campus area anymore. Any ideas on how to fix this? Thanks!
iNat does intentionally simplify shapes. Can you post screenshots of what yours look like?
Here’s a picture showing the 3 campus (on the left what I see in QGIS and google, on the right what I see in iNat)
btw thanks for a really fast reply @jwidness !
the boundaries you loaded are fine. the simplification only occurs when iNat delivers your place boundaries as vectors. probably you see the extreme simplification here because you’ve created a single place that contains 3 distant locations. (personally, i would have created each of the campuses as their own place in the system.) but the simplification of the polygon vector does not affect the ability to filter for observations in your defined boundaries.
you can see a better representation of the actual boundaries loaded in the system if you request the place polygons as map tiles. here’s what i see in my custom map (when i enable the iNat Place layer):
I see. So it’s just a visualisation issue, the underlying fil is correct if I understand correctly?
I’ll try to upload the campuses separately - I think you are right that it would make more sense.
Thanks a lot for the really fast help!!! Very appreciated!
I haven’t seen a problem with iNat removing points as I see in your example. That’s not right!
I agree with the recommendation to break them out into separate places. You can then create a Project to aggregate all of them.
Example: The most complex “Place” I’ve created was a missing US EcoRegion: 84c, Barrier Islands/Coastal Marches. It covers a narrow band of the North America Atlantic Coast, from Montauk Point on Long Island in the north, to Cape May, New Jersey in the south.
I downloaded the Shapefile, imported into QGIS, edited for only EcoRegion 84c, and exported to KML. Even when I broke up the KML under iNat’s 1MB limit, I got errors (which iNat doesn’t explain) trying to import it into a place.
These boundaries are complex, and I didn’t want to spend the time “simplifying” them. By trial and error, removing shapes from the KML and trying to import, I found the offending shapes: two adjacent islands in a bay whose shapes overlapped. When I broke each out into their own KML, iNat accepted it.
I ended up with four places:
- Cinder Island
- North Cinder Island
- Nassau County (The Cinder islands are in a bay in Nassau County)
- Suffolk County
Although the Places had to be separate, I was able to “combine” them in a single Project. I still have to add the areas for NYC and New Jersey.
The Project Description summarizes some of the issues I had getting even these into iNat. The second bullet is a problem you might be encountering with your shapes:
Due to limitations with importing KML to create a new iNaturalist Place, this Project includes several different places, all of which together comprise the entirety of EcoRegion 84c:
- EcoRegion GIS data are only available in ArcGis Shapefile format (*.shp files). They must be imported into a GIS tool as SHP, then exported to KML. The open-source QGIS program does this.
- Converting from SHP to KML appears to introduce errors in the resulting KML, such as overlapping boundaries, e.g.: Cinder Island and North Cinder Island in Nassauc County, NY. Importing such KML to create an iNaturalist Place will fail.
- KML files must be 1MB or less in size. This precludes EcoRegions with complex boundaries from being imported as a single file. They must be broken up into multiple files, e.g.: by State and/or County, to get them down to a manageable size.
it’s not removing points from the actual stored place. it’s just delivering a simplified version of the polygon vector that only affects the display of the place in this particular case.
i don’t think that’s the right use of of the word “simplifying” in this context. if there were overlapping boundaries, i would have just corrected the boundaries. there’s no need to make Cinder Island and North Cinder Island their own places unless you have an actual use case for them as separate places.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.