I collected data on small terrestrial mammals in two protected areas (PAs). In each PA I used Sherman's live trap to trap small mammals from three different land-use types. In each land-use type 20 experimental plots were sampled once with 40 traps. I also collected some environmental and habitat data at every other trap station.
In occupancy analysis what is the number of sites and occasions?
Hi Abena,
If you consider the two PAs separately, you would have three sites per PA with 20 replicates (n or spatial replicates); consider both at the same time and you would have six sites with same 20 replicates. But considering (I am guessing) that the land-use types in both PAs were the same, you should run the analysis separately for the two PAs and then compare the results. As for occasions, how long were the traps placed for?
Good that you have environmental variables, you can use them as covariates in your occupancy models.
No, the land use types are your replicates, so long as the PAs are distant from each other (otherwise the land use types are pseudo-replicates). You therefore have two replicates of each land use type. The 20 plots are simply sub-samples and 40 traps is simply the sampling method. But it doesn't matter too much. At least you can get local information and possibly more if your PAs area spatially isolated from each other :)
It depends on how you want to roll this up. Presuming that you checked traps on multiple occasions through time, you could do an occupancy analysis at the scale of individual traps. However, depending on the movement ability of your small mammals relative to the distance between traps within a plot, it might make more sense to do an occupancy analysis at the scale of the plots. Here, the plot is occupied if one or more traps have a captured animal at a given time step. But now you have richer data than 0/1 because you could actually have anywhere between 0 and 40 "hits". There are extensions of the classic occupancy analysis that can accommodate this extra information that are called "multi-state occupancy models."
Since you have habitat data, you can either include that directly as a (continuous) covariate or you could use the land use categories as a discrete covariate.
Since you are asking what the number of occasions is, I suspect that you did not actually check your traps multiple times through time...
Shah Nawaz Jelil Linas Balciauskas The protected areas are over 150km apart. in each of the 3 landuse types I have 20 plots across the PA. each had 40 Sherman traps set for only one night in one season. captures were identified and morphometrics taken. vegetation data at the point of capture (as well as other randomly selected points) was taken.
Abena, then I believe you should run the analysis separately for both PAs. How far were these plots placed in each PA?
Again to be perfectly clear, each of your plots had 40 traps? In all the 20 plots?
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping