Differences in behavior in both sex.

[Pablo Lau]

Hi Murray and forum,

How to deal with models when one of the sexes is disproportionately captured by sampling? (e.g., when females travel off-road routes, where camera traps are usually placed). Is it reasonable to fix pmix with the sex ratio commonly encountered in the specie, ej: fixed= =list(mix=0.6?. Thanks in advance.

[Pablo Lau]

Another question: We obtain higher density values in models that include sex as a covariate of g0 and sigma, compared to models that do not model sex. What could explain this trend? The data come from several camera trap studies focusing on jaguars, in which females tend to be under-sampled due to their more cautious behaviour, which involves avoiding roads and exposed trails, compared to males. I thought that models without sex overestimate detection overall, consequently underestimating the number of individuals required to explain the detections and producing lower estimates of density. Is this argument reasonable?.

II would be very grateful to read your thoughts on this topic. Thanks..!

[Eric Howe]

Good day,

Yes, if there is heterogeneity in detectability among individuals (including between sexes), models that ignore that heterogeneity may yield negatively biased estimates of abundance as you describe. 

Note that if the sex with higher sigma (male jaguars?) has lower g0, the a0 parameterization could allow you to fit a simpler model that accounts for the differences.

https://www.otago.ac.nz/density/pdfs/secr-parameterisations.pdf 

https://esajournals.onlinelibrary.wiley.com/doi/abs/10.1890/13-1497.1

Best,

Eric