how to create and use footprintWKT

[sarahcd]

Are there any best practices for defining and creating footprintWKT values? The DwC footprintWKT term seems like it has a lot of potential, e.g. to represent the bounds of a track/dataset, home ranges, or possibly to help make tracking datasets discoverable by allowing spatial searches using a footprint rather than millions of observations. Ideally these best practices could address

• how to report footprints for long-distance migrants
• where to describe how a footprint or home range was calculated
• how to address outliers that might or might not have been identified by the owner
• how to address low-accuracy tracking methods

[peterdesmet]

For the actual value, you have to close the polygon with the first coordinate again and have them in clockwise order. So:

POLYGON ((-112.762 43.841, -112.547 43.841, -112.762 43.763, -112.547 43.763))

Should be:

POLYGON ((-112.762 43.841, -112.547 43.841, -112.547 43.763, -112.762 43.763, -112.762 43.841))

You can test this at https://arthur-e.github.io/Wicket/sandbox-gmaps3.html

[wardappeltans]

@peterdesmet @sarahcd see the OBIS manual http://iobis.org/manual/darwincore/#location. If you do not close the footprintWKT you should use LINESTRING.

[peterdesmet]

I think footprintWKT has potential for a bounding box indeed. Some thoughts:

• I would only populate a footprintWKT for a deployment (which groups many occurrences), not a capture/recovery event (which contains one occurrence)
• I think the footprint should just be for guidance, so I find it less important how it was calculated (i.e. I would calculate it from all min/max coordinates of occurrences except outliers)
• Long distance migrants will have a big bounding box 😄

Thanks @wardappeltans, good reference. If we want to express a bounding box, we should use POLYGON imo.

[peggynewman]

I agree with all of this and have been barking about using footprintWKT to summarise deployment events. I suspect that it would be important to document how the footprint was calculated, especially if we use different home range algorithms to summarise. Alpha hulls might be a useful way of capturing the long ranging migrants.

[sarahcd]

I updated the Mahoney use case to close the polygons and list only for deployment events as recommended by @peterdesmet.

[peterdesmet]

Thanks @sarahcd, the polygons are still not in clockwise order however. E.g. for F53:deployment1:

POLYGON ((-112.078   38.510, -111.932 38.510, -112.078 38.230, -111.932 38.230, -112.078 38.510))

results in ⏳:

Should be:

POLYGON ((-112.078 38.510, -111.932 38.510, -111.932 38.230, -112.078 38.230, -112.078 38.510))

See https://arthur-e.github.io/Wicket/sandbox-gmaps3.html 😄

[sarahcd]

Thanks @peterdesmet for catching that, and illustrating why that test site is helpful :) Made the fix.