search

trias-project / wrims-checklist

🐡 World Register of Introduced Marine Species (WRIMS)
MIT License
0 stars 0 forks source link

select Belgian locations #3

Closed LienReyserhove closed 5 years ago

LienReyserhove commented 6 years ago

WRIMS containes data about 1376 unique localities. We need to filter out the Belgian ones. However, filtering on localities containing "belg-" is not sufficient.

An example of some localities:

It's a tedious job to check them all. However, for each, we have coordinate information available, although not in the distribution extension but on http://marineregions.org. For instance, this is the information for Spuikom: http://marineregions.org/mrgid/5477

Maybe we could use the coordinate information to filter out the Belgian localities?

peterdesmet commented 6 years ago

As long as:

We will be able to filter on filter on non-native distributions for belgium, using:

LienReyserhove commented 6 years ago

Countrycode is not provided yet. So this still needs to be done. To avoid that someone needs to interpret all the localities, I made the suggestion of using the coordinates. But if you know the data it shouldn't be much effort to just add the countrycode I assume.

bart-v commented 6 years ago

We (WRIMS/Aphia) can automatically populate the field "countryCode" via the marineregions.org gazetteer (all localities have a hierarchy). There is no need to do this manually.

LienReyserhove commented 6 years ago

Thank you for this! When could you provide us with this update? If I have this information, I can have a better idea which values for occurrenceStatus, invasion stage and establishmentMeanswe need to integrate (see #2 )

bart-v commented 6 years ago

The next update (2018-05-01) should have it.

LienReyserhove commented 6 years ago

Thanks for the update! There are now 963 records for Belgium. One question: a large majority of the new country field remains unpopulated. E.g. taxonID 100872 in the distribution extension has locality = North Sea. Obviously, this species does not occur exclusively in Belgium, but what to do with it? Do we only integrate species for which have have country information or do we broaden the scope (which will make our task a lot more complex)... I suggest the first option...

bart-v commented 6 years ago

I think you will need to go through all distinct localities and see which ones can be attributed to Belgium sensu lato.

You can try to use the marineregions.org REST API to do this (semi)automatically, but I'm afraid this will not work out well.

peterdesmet commented 5 years ago

Do we only integrate species for which have have country information ...

Yes, that is the safest option and we'll see how far we get with that.

LienReyserhove commented 5 years ago

Decided to go for option one