Set `manipulation-type`

[peterdesmet]

manipulation-type is an deployment-level field we always try. It has controlled values. The values are:

value	definition	remark
none	The animal received no treatment other than tag attachment and related measurements and sampling (if applicable).	The default.
manipulated-other	The animal was manipulated in some other way, such as a physiological manipulation.	We use this in the GPS-data for animals treated at VOC.
relocated	The animal was released from a site other than the one at which it was captured.	We've never used this before. It also implies a bit that capture and release info are different and available. It can definitely differ for CR-Birding info, but we don't have the two locations.
confined	The animal's movement was restricted to within a defined area.	Likely not applicable
domesticated	The animal is domesticated, for example, is a house pet or part of a managed herd.	Not applicable
reintroduction	The animal has been reintroduced as part of wildlife conservation or management efforts.	Likely not applicable

@EricStienen it would especially be useful to identify sick/wounded birds that were treated before/during ringing (to set manipulation-type = manipulated-other). How can we recognize this in the data? I see 2 potential fields observation_condition (which can contain "released after revalidation") and observation_location (which can contain "VOC"). Here are the values for those 2 fields for captures only. Does the suggested manipulation make sense? Would you use both fields or only observation_condition?

observation_condition	location	records	manipulation
alive and healthy	Kapellen, VOC	141	manipulated
alive and healthy	Kappellen, VOC	23	manipulated
alive and healthy	Merelbeke, VOC	6	manipulated
alive and healthy	Oostende, VOC	100	manipulated
alive and healthy	Raversijde, VOC	1575	manipulated
alive and healthy	Raversijde, VOC: dak	4	manipulated
alive and healthy	name does not contain "VOC"	11857	not manipulated
released after revalidation	Berendrecht, Opstalvallei, VOC	21	manipulated
released after revalidation	Oostende, VOC	8	manipulated
released after revalidation	Raversijde, VOC	29	manipulated
released after revalidation	name does not contain "VOC"	678	manipulated
sick or wounded	Raversijde, VOC	1	manipulated

[EricStienen]

seems OK.

Only the 678 released after revalidation where the name does not contain VOC worry me a bit. Can you list some of these ringcode so that I can check them?

[peterdesmet]

Looking more into this: while there are 678 observations with release after revalidation where the observation_location does not contain "VOC*", there are far fewer records where the user_name does not contain "VOC*". There are only 2:

• 2 by INBO at Raversijde, Strand
• 78 by Reinoud Allaert at Nieuwpoort, IJzermonding

I think those are fine.

[EricStienen]

Yes indeed. It is not the location that lmust contain VOC, but it is the observer!