E-SURGE, constraint like p1=a*p2

[pascaline lg]

Hi,
I am studying a species which is ringed and monitored in two areas. I have recoveries of individuals that come from ringers and from public. We assumed that ringers mostly reported recoveries from their monitored areas and that public reported only recoveries from birds that have emigrated from monitored areas to unmonitored areas and lived there until death. In this context it seems that E-Surge is appropriate, but it is the first time I use this software so I have some questions!

The idea is to define 4 states + extra-state of E-Surge:{area1, area2, dead in monitored area, dead elsewhere, extra-state}.
And 3 events +not seen :{not seen, seen in area 1, seen in area2, seen dead}

1- Is it possible to use the extra state of E-Surge as alive elsewhere (i.e. put no null transition in the last row of transition matrix using equal survival constraint)?
2- In the event matrix, for the column corresponding to the event seen dead we will have 2 probabilities: one due to ringer from monitored areas and one due to public from elsewhere. We would like to put a constraint like P from ringers= alpha*P from public. How can we do that? in GEMACO or IVFV?

Thank you very much for your help,
Pascaline

[CHOQUET]

Hello Pascaline,

Some remarks:

> 1- Is it possible to use the extra state of E-Surge as alive elsewhere
> (i.e. put no null transition in the last row of transition matrix using
> equal survival constraint)?

Just ask to E-SURGE that the number of states is 5 instead of 4.
You can choose which state is 'alive elsewhere'. For example, the third.
In the event matrix, the probability from state 3 to event 'not seen' will be always 1.

> 2- In the event matrix, for the column corresponding to the event seen > dead we will have 2 probabilities: one due to ringer from monitored
> areas and one due to public from elsewhere. We would like to put a
> constraint like P from ringers= alpha*P from public. How can we do
> that? in GEMACO or IVFV?

This is a more difficult point not because of the constraint
which can be defined with a two steps process for events in E-SURGE, but because
you should defined clearly the model.

In fact, doing a mixture of capture and recoveries in several sites is not straightforward.
You have to read at least papers from:

Encounter history modeling of joint mark-recapture, tag-resighting and tag-recovery data under temporary emigration
Barker RJ, Burnham KP, White GC
STATISTICA SINICA Volume: 14 Issue: 4 Pages: 1037-1055 Published: OCT 2004

Barker RJ, White GC, McDougall M
Movement of paradise shelduck between molt sites: A joint multistate-dead recovery mark-recapture model
Barker RJ, White GC, McDougall M
JOURNAL OF WILDLIFE MANAGEMENT Volume: 69 Issue: 3 Pages: 1194-1201 Published: JUL 2005

and more recently

Duriez O., S.A. Saether, B.J. Ens, R. Choquet, R. Pradel, R.H.D. Lambeck, M. Klaassen. (In press).
Estimating survival and movements using both live and dead recoveries: a case study of Oystercatchers confronted with habitat change. Journal of Applied Ecology


Sincerely,

Rémi