17840
Reporter: kbraak
Assignee: mblissett
Type: Bug
Summary: Inconsistent geolookup for French territories
Priority: Blocker
Resolution: Fixed
Status: Closed
Created: 2015-09-23 14:19:39.63
Updated: 2016-04-21 12:00:03.928
Resolved: 2016-04-21 11:32:44.912
Description: In our occurrence processing, the geolookup does not return France for all French territories.
According to POR-2707, it returns France for the following territories:
French Guyana, Guadeloupe, Martinique, Mayotte, and Réunion
It does NOT return France for French Polynesia though, highlighting a major inconsistency in the way we handle French territories. ]]>
Author: trobertson@gbif.org
Created: 2015-09-23 14:51:29.354
Updated: 2015-09-23 14:58:25.385
Disclaimer: I don't claim to be an expert on administrative levels.
Using a coordinate for Tahiti you see the services return the ISO code PF [1] since "overseas countries" are treated as independent entities in the ISO 3166-2 standard [2]. Administratively speaking, I understand French Guyana, Guadeloupe, Martinique, Mayotte, and Réunion to be classed as overseas departments or territories, which is considered as a different classification level to an overseas country.
It can of course be modified, but this is all based around ISO 3166 (as is almost all GBIF reporting) and comes with the open geographic databases we use (the exclusive economic zone and the natural earth data).
The French Polynesia page is shown: http://www.gbif.org/country/PF/summary
[~thirsch@gbif.org] - this is not a technical decision, but something I think needs your decision. We can add a clause simply for FP and just react to requests as they come in if preferable. If I recall correctly, that is how we have treated some specific islands in the past.
[1] http://api.gbif.org/v1/geocode/reverse?lat=-17.7&lng=-149.5
[2] https://en.wikipedia.org/wiki/ISO_3166-2:PF
Author: thirsch@gbif.org
Comment: I agree there are policy considerations here and the lead needs to come from us (i.e. management) based on our understanding of current technical practices. As verbally discussed with the informatics team I will take some more soundings and come up with what I think is the preferred solution and we will then see how feasible it is to implement. Although there is no one solution that suits everyone, I think we should as far as possible avoid going into individual exceptions as the range of different categories of political status for overseas territories is almost endless (5 categories for France alone!) and the principle should be to have a single system using neutral standards/authorities that can be transparently described in our data views.
Created: 2015-09-23 16:46:12.781
Updated: 2015-09-23 16:46:12.781
Author: mblissett
Created: 2016-04-21 11:32:44.942
Updated: 2016-04-21 11:32:44.942
This is now consistent to the ISO standard. When alternative meanings are required, they will need to be built at a higher level (for reporting statistics etc).
Réunion: http://www.gbif-uat.org/country/RE/summary
Wallis and Futuna http://www.gbif-uat.org/country/WF/summary
Saint Martin http://www.gbif-uat.org/country/MF/summary
Saint Pierre and Miquelon http://www.gbif-uat.org/country/PM/summary
New Caledonia http://www.gbif-uat.org/country/NC/summary
Mayotte http://www.gbif-uat.org/country/YT/summary
French Southern Territories http://www.gbif-uat.org/country/TF/summary
French Guiana http://www.gbif-uat.org/country/GF/summary
French Polynesia http://www.gbif-uat.org/country/PF/summary
Martinique http://www.gbif-uat.org/country/MQ/summary
Guadeloupe: http://www.gbif-uat.org/country/GP/summary
France: http://www.gbif-uat.org/country/FR/summary (includes Clipperton Island, which is in line with ISO 3166. There's an exceptionally reserved code for the island, but it's not a proper part of the standard.)
Saint Barthélemy http://www.gbif-uat.org/country/BL/summary (missing EEZ, POR-3084)