Using covariates from final capture event

questions concerning analysis/theory using program MARK

Using covariates from final capture event

Postby BigBen197 » Mon Sep 19, 2022 11:58 pm

Context: I am investigating whether the relationship between fish length and apparent survival differs between two areas. These areas have been sampled bimonthly for two years.

Question: Would it be unsuitable to use individual length at last recapture instead of length at first capture? My thinking is, the last known length is closer to the time of any assumed mortality event, and the relationship between initial length and survival may be confounded by different individual growth rates (especially for individuals recaptured throughout the study).

Cheers,
Ben.
BigBen197
 
Posts: 1
Joined: Sun Sep 18, 2022 11:52 pm

Re: Using covariates from final capture event

Postby jlaake » Tue Sep 20, 2022 9:02 am

Yes it would be unsuitable because if there is a predictive relationship, survivors would presumably grow longer whereas those that die earlier would be shorter. Therefore you are measuring a covariate to predict survival (length) that is affected by survival process. See the problem? This is a common problem and it requires a growth model to be incorporated into the estimation model and recaptures of survivors provides information on growth. Individual variation in growth would complicate the model and the covariates usefulness. In summary, survival covariates have to be measured prior to the survival process for all live individuals. Since we don't recapture all individuals at each occasion, a growth model would have to predict the length at each future occasion. Hopefully this makes sense.
jlaake
 
Posts: 1417
Joined: Fri May 12, 2006 12:50 pm
Location: Escondido, CA

Re: Using covariates from final capture event

Postby ehileman » Tue Sep 20, 2022 11:43 am

In case it's helpful, here are two recent papers involving snakes that incorporate individual growth models into capture-recapture models as Jeff described. In Henderson et al. (2021), shameless plug, we used a maximum likelihood approach that assumes no error in body size measurements and predicted sizes of unobserved snakes. Rose et al. (2018) used a slicker Bayesian approach that accounts for uncertainty in body size.

Henderson, R.W., E.T., Hileman, R.A., Sajdak, B.C., Harrison, R., Powell, and D.R., Bradke. 2021. Effects of body size, diet, and transience on the demography of the arboreal boid snake Corallus grenadensis on Carriacou (Grenada Grenadines, West Indies). Population Ecology. 63:177-188.

Rose, J.P., G.D., Wylie, M.L., Casazza, and B.J., Halstead. 2018. Integrating growth and capture–mark–recapture models reveals size‐dependent survival in an elusive species. Ecosphere.9:e02384.
ehileman
 
Posts: 51
Joined: Sat Nov 26, 2011 6:40 pm
Location: West Virginia University

Re: Using covariates from final capture event

Postby jlaake » Tue Sep 20, 2022 1:02 pm

Would you mind sharing a copy of your paper with me please? I'm retired and don't have access to most journals. I tried to email you separately but the address that shows up on phidot was rejected. Did you write your software so it could be used generically by others? jefflaake@gmail.com

Jeff
jlaake
 
Posts: 1417
Joined: Fri May 12, 2006 12:50 pm
Location: Escondido, CA

Re: Using covariates from final capture event

Postby ehileman » Tue Sep 20, 2022 3:19 pm

jlaake wrote:Would you mind sharing a copy of your paper with me please? I'm retired and don't have access to most journals. I tried to email you separately but the address that shows up on phidot was rejected. Did you write your software so it could be used generically by others? jefflaake@gmail.com

Jeff


Hi Jeff,
I sent you the paper. As you'll see, we separately fit Fabens using the R packages FSA and nstools, recursively estimated predicted snout-vent lengths for unobserved individuals, and then modeled them in MARK using the standard approach for temporal individual covariates. Nothing too fancy.

P.S. Thanks for the heads up regarding the defunct email address. I'll have to reach out to Evan to see if he can change it on his end as I was unable to update it through the user control panel.

Cheers,

Eric
ehileman
 
Posts: 51
Joined: Sat Nov 26, 2011 6:40 pm
Location: West Virginia University


Return to analysis help

Who is online

Users browsing this forum: No registered users and 10 guests

cron