Unequal time intervals

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

Unequal time intervals

Postby Geraldine » Fri May 12, 2006 3:54 pm

Hi everyone,
I’m trying to use M-Surge with unequal time intervals and encounter several problems. Capture sessions were made in May and August each year so that time intervals are 4 and 8 month long respectively (i.e., summer and winter). I have eleven occasion of captures, (1 3 5 7 9) corresponding to summers (May to August, 4 months) and time (2 4 6 8 10) to winters (August to May, 8 months). Also, times (1 2), (3 4), (5 6), (7 8) and (9 10) each correspond to a population year.
1) I wanted to test for an effect of year on survival but I can’t find how to do it with unequal time intervals. With equal time intervals I simply test it by writing t(1 2, 3 4, 5 6, 7 8, 9 10) but since times 1 and 2, 3 and 4, etc. correspond to different time intervals (i.e. 4 and 8 months) I can’t specify that they are similar using this notation: t(1 2, 3 4, 5 6, 7 8, 9 10). Is there another way to test for a year effect?
2) If I make my model selection without taking into account the fact that I’m dealing with unequal time intervals, the best model I obtain can be written Phi (time(2, 4, 6, 8, 10, 1 3 5 7 9)+age) p (time(1, 3, 5, 7, 9, 2 4 6 8 10)+age). Can I use this final model to recalculate the monthly survival rates for each time period by just switching to the “unequal time intervals” option? I tried it and the monthly survival rates I obtain are different from those I obtained with the model that did not take unequal intervals into account. For example, with equal time interval I obtained 0.521 for the summer survival of young and if I recalculate it with the same model but taking into account the unequal time intervals I obtain a monthly survival rate of 0.889, which gives me a summer survival rate of 0.8894=0.625. Also, if I compare my capture rates they are different between the model with equal and unequal time intervals: I thought that unequal time intervals were considered only when estimating survival probabilities?
Thanks if you can help me with this or indicate me some references that would give some details about using unequal time intervals.
Geraldine
Geraldine
 
Posts: 2
Joined: Fri Apr 21, 2006 9:20 am

Postby CHOQUET » Mon May 22, 2006 4:46 am

Hello Geraldine,

When the option unequal time intervals is used with in your case values of 4 and 8 respectively for winter and summer period, estimated survivals in M-SURGE are monthly values.
1) So, to test equality between season, simply use the intercept.
2) As this option change the umbrella model, you have to do again all the model selection.
3) For additive model like <<t(2, 4, 6, 8, 10, 1 3 5 7 9)+age>>, M-SURGE doesn't necessarily start from a good starting point, and very often the fitting failed. I recommend to used rendom initial values in this case.

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


Return to analysis help

Who is online

Users browsing this forum: No registered users and 9 guests