Definition of some 2 species interaction factors

questions concerning analysis/theory using program PRESENCE

Definition of some 2 species interaction factors

Hello all,
Can someone please provide a definition for the parameters listed below which are components of the two species models. Most of the paramaters in the two species models are defined on page 232 of MacKenzie et al. (2004) however there are a couple that are not and these are:
rA (which is a component of all three paramaterisations)
rB (which is a component of the first paramaterisation)
And can someone please confirm that PsiBa refers to the probability of occupancy of species B, given the site is not occupied by species A.
Further to the above, the definitions of probability of detection parameters(excluding the two above of which I am unsure) relate to the probability of detection during the jth survey. One of my key questions is do predators affect the detection probability of prey at camera sites over each two week sampling period. In addressing this question using the 2 species models in presence, does the species interaction factor (SIF) generated from detection probabilities refer to species interactions for each survey (in my case each 24hr camera sampling period) or can it incorporate the detection history for each species over a season (season in my case being a two week sample)?
Bill

Posts: 10
Joined: Tue Jan 31, 2012 11:18 pm

Re: Definition of some 2 species interaction factors

Hi Bill,

rA = probability of detecting species A, given both species are present
rB = probability of detecting species B, given both species are present
psiBa = probability of occupancy for species B, given species A is absent

I'm not sure I follow your subsequent question, but I will try to answer the best I can. When you incorporate survey-specific covariates to detection, then PRESENCE will provide multiple detection probability estimates to correspond with each survey. If you do not include survey-specific covariates in your models (i.e., detection is held constant among surveys), then PRESENCE will provide a single estimate for each site for the entire season. Furthermore, if you do not include site-specific covariates, then PRESENCE will provide the same estimate for all sites, and only the first estimate will print to the output file to keep things simple.

Hope this helps,
Tim
tajensen

Posts: 11
Joined: Thu Oct 13, 2011 12:37 pm

Re: Definition of some 2 species interaction factors

Bill,
rA and rB detection probabilities are defined at bottom of page 236 of MacKenzie et al (2006).

In terms of the detection, one species could affect the other in 2 ways (within the model). The first is, given both species are present at a site, if 1 species is seen in a survey, does that affect the probability of also detecting the other species in the same survey. In terms of the model, a SIF that is different from 1 would indicate some sort of interaction. You could define the model such that the SIF is the same for all surveys or varies in time or according to a covariate. The second way is if the first species is present at a site, is the probability of detecting the second species different compared to a site where only the second species is present. Here you're asking whether rB = pB. You could imagine the first situation represents a short-term avoidance of each other (for example), while the second might represent a longer-term behavioral response to the presence of the predator, or could be something like lower prey abundance at a site with the predator vs a site where the predator is absent.

These options aren't mutually exclusive either, and you could have both occurring as well (which is the most general model)
Cheers
Darryl
darryl

Posts: 493
Joined: Thu Jun 12, 2003 3:04 pm
Location: Dunedin, New Zealand

Re: Definition of some 2 species interaction factors

Wonderful, sincere thanks to both of you - I have a much better handle on my results now.
Bill

Posts: 10
Joined: Tue Jan 31, 2012 11:18 pm

Re: Definition of some 2 species interaction factors

I also have a question about interpreting SIFs. I recently ran a species interaction model for mink and muskrat and we hypothesized that muskrat would avoid mink. However, our SIF ( 1.8 ) suggests that muskrat and mink occur together more often than expected by chance. This has caused us to question the initial hypothesis that muskrat avoid mink and we are now considering the possibility that mink key in on muskrats. However, if this alternate hypothesis is true, would we need to rework the models so that muskrats are the dominant species (attracting mink)? Or would it be okay to infer that mink are keying in on muskrats without running model sets with muskrat as the dominant species? Most two species models in the literature involve competition, not predation, so I’m not certain what is considered appropriate for species interaction models involving predators and prey.
Thank you so much! This forum has been extremely helpful to me.
Caleb
CalebBomske

Posts: 5
Joined: Fri Jan 21, 2022 1:57 pm