Potential nest area bias in habitat selection

[Chris MacColl]

Hello CTMM team, 

Ive assessed habitat selection in an endangered bird of prey using RSFs via the CTMM package. We had 13 range-resident individuals which had a bias toward nesting females (n=10). We developed RSFs for each individual-year (n = 37) so we can assess dynamic environmental processes taking place in the home range including annual fire activity, habitat loss, habitat restoration, etc. 

Female movements are concentrated to the nest for approx. 3 months of the year during incubation & early nestling phase before gradually roaming wider across the home range for another ~month as the chicks become bigger and less vulnerable. Thereafter once chicks fledge and gain independence they are freed from the nesting area and utilise the full home range extent although the core home range (50% wAKDE) is still centred around the nesting area, the heart of the breeding territory. 

Some co-authors have raised concerns that our RSF results may be biased by this extended nest-site concentration. My understanding is that RSFs measure relative selection and avoidance given availability, so the nest focus should not necessarily introduce bias. I would, however, appreciate feedback on whether this is a significant issue that requires correction. Options include (as I understand it) masking the nest area in the AKDE (e.g. a 1 km hard boundary buffer) or explicitly modelling breeding vs. non-breeding selection. My preference is to minimise subsetting or filtering data further unless results are demonstrably biased by nest-site utilisation.

Thank you very much for your thoughts. 

Chris

[Christen Fleming]

Hi Chris,

Large changes in selection would bias the results. I would model breeding versus non-breeding selection either by segmenting the track or including time-dependent modifiers to the selection coefficients. You can do this by annotating the tracks with a season categorical and then having a formula like ~ season:landcover . If the range contracts during nesting season, then you could also add a term like season:I(x^2+y^2) to model that, though I haven't tested spatial interactions of this form (if it doesn't work, contact me and I will fix it).

Best,

Chris