Temporal covariates and seasonal survival

[tamarlok]

Hello,

I'm trying to model seasonal survival as a linear function of a temporal covariate, but I fail to write the correct GEMACO.
For simplicity (the reality is much more complicated...), let's assume I have a 5-year study with 11 capture occasions (1 every half year), resulting in 10 intervals of half a year.

To start, I made a very simple seasonal model where survival in the first half of the year differs from the second half, but without temporal variation in these seasonal survival estimates between years:

GEMACO: season1+season2
where season1 = t(1:2:11) and season2 = t(2:2:10)

Now I would like to model survival in the second half of the year as a linear trend over the 5 years. I tried:

season1+season2.[i+t*x]

with x being either:
1 1 2 2 3 3 4 4 5 5
or
1 2 3 4 5

But I get errors all the time. What am I doing wrong?

I hope anyone can help me!

Cheers, Tamar

[Guillaume Souchay]

Hi Tamar,

I think that your problem is your use of shortcut.
You specified that season1 and season2 can't vary over year and you want to test a temporal variation on something without temporal variation ...

With your shortcut, if you model survival as season1+season2, you will have only 2 parameters.
You can not input your temporal variation on this model. If you want to do it, you have to restart with a model which enables overyears variations. You need to have as many occasions as the number of temporal covariates.
So, with a temporal covariate of 5 values, you need to have 5 temporal occasions ...
You may read (again) the E-Surge manual and the section on temporal covariates.

Kind regards,
Guillaume