Understanding sigma

[Lachlan McRae]

Hi Murray,

I'm hoping to clarify exactly what a Sigma value means. As far as I know, Sigma is defined as the spatial scale of detection function (in meters) that determines the rate at which the probability of detection decreases with distance and is related to the home-range area of individuals. Does a Sigma value of 80m mean that on average the detection probability of individuals is halved (50%) when they are 80m away from their home-range centre? I'm studying short-eared possums (Trichosurus caninus) btw.

Cheers,

Lachlan

[Murray Efford]

I'd hesitate to interpret sigma in such concrete terms. It is the spatial scale parameter of a particular detection model. I imagine you are using a halfnormal model. The probability is for capture in a particular trap in the absence of any competing traps. Then I get

r <- 80                               # distance from activity centre

sigma <- 80          

p <- exp(-r^2/(2*sigma^2))   # half normal, ignoring g0 for probability relative to centre

p

[1] 0.6065307

Which is not 0.5. Use a different detection function and you'll get a different p result for given r, even when both give similar density estimates.

[Lachlan McRae]

Thanks Murray!