Alternate sessions (breeding and wintering), some questions

questions concerning analysis/theory using programs M-SURGE, E-SURGE and U-CARE

Alternate sessions (breeding and wintering), some questions

Postby simone77 » Mon Apr 30, 2012 5:23 am

Hi,

I am dealing with an analysis of local survival and dispersal of a bird population marked in site A in spring, and resighted in winter in site A, site B and site C.
I have 16 breeding season when chicks (only chicks) are marked and 16 correspondent wintering seasons.
When chicks are captured (and marked), some of them are sexed, so that I have sex uncertainty in my data set and this is also why I chose to use E-SURGE.

I have some difficulties on how to build the models.

I have defined 8 states (apart from dead state):

1-female chick site A (--> breeding seasons)
2-male chick site A (--> breeding seasons)
3-female site A (--> wintering seasons)
4-male site A (--> wintering seasons)
5-female site B (--> wintering seasons)
6-male site B (--> wintering seasons)
7-female site C (--> wintering seasons)
8-male site C (--> wintering seasons)

Moreover, there are four other possible events:

9-unknown sex chick site A (--> breeding seasons)
10-unknown sex site A (--> wintering seasons)
11-unknown sex site B (--> wintering seasons)
12-unknown sex site C (--> wintering seasons)

My data set is arranged in headed format (to allow this high number of events) and the first column represent the 1st breeding season, the second the 1st wintering season, the third the 2nd breeding season, the fourth the 2nd wintering season and so on.
In the 4th (t 7) and 10th (t 19) breeding seasons there was no breeding at all and no chicks were marked. Also, individuals in site B and C were not resighted before 3rd wintering season (t 6) and individuals in site B were not resighted in the 4th and 5th (t 8, 10) wintering season.

These are my matrices in GEPAT:

Initial State:
Image
Really I would have set 9 columns (as there are 9 states including the dead state) but it appears a window saying "Dimensions of pattern matrices not allowed!!!"


Survival:
Image

Transition:
Image

Capture:
Image

State Assignment:
Image

Sometimes, like in this case, due to particular data structure or due to presence/absence of some specific data, I don’t understand how to deal with events for the parameters estimate both in GEMACO and in IVFV fixing parameters.

Two examples:
1) about the particular data structure, given the above-explained structure of data, events 1, 2 and 9 are present only in odd sessions and, viceversa, the others only in even sessions.
2) about the presence/absence of some data, events 5, 6, and 11 are not present in t(2,4,8,10) and events 7, 8, and 12 are not present in t(2,4).


QUESTIONS:

So, given these shortcuts:
“brt”=[t(1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31)]
“wrt”=[t(2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32)]
“brts”=[t(1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31)]
“wrts”=[t(2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32)]

1. In Initial State, does it make sense to define a sentence like this in GEMACO:
brt + wrts
and fix the last parameter in IVFV (corresponding to “wrts”) to zero?

2. In Survival, how would I define a state and time effect?
f(1,2).brt + f(3_8).wrt + others

and fix to zero (i) the last parameter corresponding to “others” (that should resume f(1 2).wrts + f(3:8).brts) and (ii) the parameters corresponding to “wrt” sessions with any data as for example f(1,2).t(7,19) when there were no captured chicks and f(5_8).t(2,4) and f(5,6). t(8,10) when there were no data about those events?

3. In Capture, again, time and state effect (age=1 in data input):
firste + nexte.f(3_8).wrt + others

and fix to zero (i) “others” as it should resume f(1,2).wrts + f(3_8).brts, and (ii) the parameters corresponding to “wrt” sessions when there were no data about events 5_8?

4. In State Assignment, again, time and state effect:
f(2,3).brt + f(4_9).wrt + others

Here the numbers are shifted of one right due to the structure of the matrix in GEPAT (1_8 == 2_9)
and fix to zero (i) others and (ii) parameters corresponding to “wrt” sessions when there were no data about events 3_8


Answers to these questions would help me a lot to better understand how E-SURGE works
Thanks for any help,

Simone
simone77
 
Posts: 197
Joined: Mon Aug 10, 2009 2:52 pm

Re: Alternate sessions (breeding and wintering), some questi

Postby CHOQUET » Fri May 04, 2012 5:08 am

Hello,

I found the overall quite complicated. So in the first step, I will not reply fully to all.
Moreover, I think that you can deleted the chck states because an individual becomes
automatically an adult after one year and because all individuals are marked as young.

QUESTIONS:

So, given these shortcuts:
“brt”=[t(1,3,5,7,9,11,13,15,17,19,21,23,25,27,29,31)]
“wrt”=[t(2,4,6,8,10,12,14,16,18,20,22,24,26,28,30,32)]
“brts”=[t(1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31)]
“wrts”=[t(2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32)]

1. In Initial State, does it make sense to define a sentence like this in GEMACO:
brt + wrts
and fix the last parameter in IVFV (corresponding to “wrts”) to zero?

YES

2. In Survival, how would I define a state and time effect?
f(1,2).brt + f(3_8).wrt + others
and fix to zero (i) the last parameter corresponding to “others” (that should resume f(1 2).wrts + f(3:8).brts)

It seems to be OK

and (ii) the parameters corresponding to “wrt” sessions with any data as for example f(1,2).t(7,19) when there were no captured chicks and f(5_8).t(2,4) and f(5,6). t(8,10) when there were no data about those events?


3. In Capture, again, time and state effect (age=1 in data input):
firste + nexte.f(3_8).wrt + others
and fix to zero (i) “others” as it should resume f(1,2).wrts + f(3_8).brts,

OK

and (ii) the parameters corresponding to “wrt” sessions when there were no data about events 5_8?


4. In State Assignment, again, time and state effect:
f(2,3).brt + f(4_9).wrt + others

Take care, no data means no capture.

Here the numbers are shifted of one right due to the structure of the matrix in GEPAT (1_8 == 2_9)
and fix to zero (i) others and (ii) parameters corresponding to “wrt” sessions when there were no data about events 3_8


Rémi
CHOQUET
 
Posts: 211
Joined: Thu Nov 24, 2005 4:58 am
Location: CEFE, Montpellier, FRANCE.

Re: Alternate sessions (breeding and wintering), some questi

Postby simone77 » Fri May 04, 2012 9:30 am

Hi Rèmi,

Thanks for answering and please excuse me for the amount of questions.
Perhaps I can reformulate the questions and make them much shorter.

Resuming them in general terms I was asking two questions:

1. (You have already answered that) - In case some parameters, due to data structure (for instance first captures made only at odd sessions), are not estimable how should I treat them?
I understand that I can set them to zero in the IVFV.

2. In case there are no data available for some parameters, how should I treat them? For instance, events 5, 6, and 11 are not present in t(2,4,8,10) and events 7, 8, and 12 are not present in t(2,4), or as another example, there are no data for some transitions (i.e. 3->5, 3->7, 5->3, etc. - I have checked the matrix of data and there are no 3 by a five or by some 0's and a five).

This case is more difficult to understand to me. As you have highlighted, no data means no captures, in other words we are dealing with incomplete detections and no captures may depends merely on stochasticity.

2.1 Anyway, would you set to zero the capture probability of those events in those sessions?
2.2 What about the State Assignment probability for those specific events (in those sessions)?
2.3 In the second example (no transitions data), would you set to zero those specific transitions?

3. Given the seasons are alternate breeding-wintering-breeding, is there a way to calculate the p of survival of an individual age(2) between wintering seasons? is it possible to define it in GEMACO by keeping the same GEPAT structure?

Also, thanks for making me see that chick state is uninformative; I can use just these detectable states:
1-female site A
2-male site A
3-female site B
4-male site B
5-female site C
6-male site C

Simone
simone77
 
Posts: 197
Joined: Mon Aug 10, 2009 2:52 pm

Re: Alternate sessions (breeding and wintering), some questi

Postby CHOQUET » Mon May 07, 2012 7:55 am

2. In case there are no data available for some parameters, how should I treat them? For instance, events 5, 6, and 11 are not present in t(2,4,8,10) and events 7, 8, and 12 are not present in t(2,4), or as another example, there are no data for some transitions (i.e. 3->5, 3->7, 5->3, etc. - I have checked the matrix of data and there are no 3 by a five or by some 0's and a five).

Define two shortcuts NAdata for non available data and data for available data

NAdata = t(2 4 8 10).to(5 6)
data = t(2 4 8 10).to(1:4)&t(13 4 6 7 9 10:40)

then put your effect on data only i.e.

firste+nexte.[data.t.f + NAdata]


This case is more difficult to understand to me. As you have highlighted, no data means no captures, in other words we are dealing with incomplete detections and no captures may depends merely on stochasticity.

2.1 Anyway, would you set to zero the capture probability of those events in those sessions?

YES but E-SURGE is able to estimate it to zero

2.2 What about the State Assignment probability for those specific events (in those sessions)?

State assignment is conditional to capture. So there is no problem for this parameter.

2.3 In the second example (no transitions data), would you set to zero those specific transitions?

This is more complicated, if there is no capture then it may be because individuals didn't gone to
the site (transition=0, capture not equal to zero) or that the field work has not be done (transition not equal to zero, capture=0)

see Balkiz, O., A. Béchet, Rouan, L., Choquet, R., Germain, C., Amat, J.A., Rendon-Martos, M., Baccetti, N., Nissardi, S., Ozesmi, U., Pradel, R. (2010). "Experience-dependent natal philopatry of breeding Greater flamingos." Journal of Animal Ecology. 79(5): 1045-1056



3. Given the seasons are alternate breeding-wintering-breeding, is there a way to calculate the p of survival of an individual age(2) between wintering seasons? is it possible to define it in GEMACO by keeping the same GEPAT structure?

Yes as the product of the breeding survival and the wintering survival
CHOQUET
 
Posts: 211
Joined: Thu Nov 24, 2005 4:58 am
Location: CEFE, Montpellier, FRANCE.

Re: Alternate sessions (breeding and wintering), some questi

Postby simone77 » Mon May 07, 2012 10:24 am

Thank you very much Rémi
simone77
 
Posts: 197
Joined: Mon Aug 10, 2009 2:52 pm


Return to analysis help

Who is online

Users browsing this forum: No registered users and 7 guests