Can I use "pcountOpen" models to investigate (partial) altitudinal migrations in bats?

[Carlos Henrique Russi]

Hi,

I have a database of ~30 locations with different altitudes, where bats were sampled with mist nets about 12 times a year (average of 3 samples per season).

Based on exploratory analyses, I suspect that some species perform partial migrations, that is, during winter, the abundance decreases in the higher areas while it increases in the lower areas. The opposite occurs in summer, when the abundance of these species decreases in the lower areas and increases in the higher areas.

So, I thought about using "pcountOpen" models to explore the hypothesis of partial altitudinal migration while considering detectability. 

The idea is to use elevation as a predictor of initial abundance (lambda) and the interaction between elevation and season of the year as a predictor of recruitment rate (gamma) and apparent survival (omega).

The model would be described in 'Unmarked' like this:

pcountOpen(~elev, ~elev*season, ~elev*season, ~1, umf)

I would like to know if my line of reasoning is correct or if there are better ways to test my hypothesis in 'Unmarked'

[Marc Kery]

Dear Carlos,

I think this is a neat idea that should work in principle. The only problem may be that the Dail-Madsen model can very hard and time-consuming to fit when you put covariates on its parameters. Also, your sample size of 30 may be a little smallish ? But best to just try it out.

I might perhaps also simplify the problem by not adopting the fully demographic version of the model, with recruitment and apparent survival, but the one where these two parameters are combined in a population growth rate. You can then of course also model this latter as a function of elevation and season and hopefully find that the growth rate increases and then decreases again at higher elevations over the course of a year.

Best regards  --- Marc

---

From: unma...@googlegroups.com [unma...@googlegroups.com] on behalf of Carlos Henrique Russi [carlosh...@gmail.com]
Sent: Wednesday, July 21, 2021 16:32
To: unmarked
Subject: [unmarked] Can I use "pcountOpen" models to investigate (partial) altitudinal migrations in bats?

--
You received this message because you are subscribed to the Google Groups "unmarked" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unmarked+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/b7be6283-4a6c-4b1d-ad84-0fbcf75ebc85n%40googlegroups.com.

[Carlos Henrique Russi]

Hi, Marc

Thanks for your answer! 

I'm running some simulations and the model is really a little slow, but that's okay as long as I reach my goal. 

However, I was a little confused about the correct way to simplify the model as per your suggestion. Should I change the "dynamic" argument to "=trend" and ignore the omega parameter? 

Eg: pcountOpen(~elev, ~elev*season, ~1, ~1, umf,dynamic='trend') 

Or is there another way to proceed? 

Also, if you know of any paper that investigated spatiotemporal dynamics with the Dail-Madsen models, and could you refer me, I would be very grateful.

[Marc Kery]

Dear Carlos,

yes, I think you have to set "dynamics = 'trend' ". I am not sure, but I would think that then you don't get any omega parameter, but only the trend parameter, which replaces recruitment and apparent survival in the general, demographic version of the model.

Not sure about any references for spatiotemporal dynamics modeling using the DM.... though there ought to be some.

Best regards  -- Marc

---

From: unma...@googlegroups.com [unma...@googlegroups.com] on behalf of Carlos Henrique Russi [carlosh...@gmail.com]

Sent: Wednesday, July 21, 2021 19:11
To: unmarked
Subject: Re: [unmarked] Can I use "pcountOpen" models to investigate (partial) altitudinal migrations in bats?

To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/42ac3f6a-b1f7-4157-8874-580c69dc7c96n%40googlegroups.com.

[Carlos Henrique Russi]

ok, Marc

I think I now have a good starting point to test my hypotheses. Thank you very much for your contribution.