Question about occupancy study design

[mspinola]

Dear list members,

I am designing a study to assess occupancy of meso-mammals in areas with and without logging. The forest tracts are not too large and some species could have home ranges larger than the forest tracts.
I am planning to use small rectangular plots (3-5 square meters) to detect species tracks (plus other methods, as hair traps and waxtags). I have 15 rectangualr plots per forest tract and the idea is to allocate the rectangular plots 500 m apart. Is the spacing among plots appropriate?
What are my sites, the forest tracts or the small rectangular plots?
How do I need to interpret psi, as occupancy or use? I read MacKenzie and Rolyle (2005), but I am not clear about the interpretation in my case.
Is there any other consideration that I need to be aware?
Thank you very much in advance.
Best,

Manuel Spínola

[darryl]

Manuel,
What is your objective for this study? How do you want to use/interpret the psi parameter? That should you guide in how to define a 'site'. You might also want to read the post form the last topic in the list on spatial correlation, some of the issues might be pertinent for your situation.
Darryl

[mspinola]

Thank you very much Darryl.
My objective is to assess the impact of forest logging on meso-mammals and my state variable will be occupancy (or use). I think that my site will be the forest tracts with and without treatment (logging), but I have only 3 forest tracts with logging and 3 forest tracts without logging (control). I cannot get more forest tracts for logistic reasons. Are the nunber of sites too low or what matters is the number of sampling units on each site (forest tracts)? In each forest tracts I will set 15 sampling (observational) units that will be visted at least 5 or 6 times (days) or even more. The space among those units will be 500 m apart.
Best,

Manuel

[darryl]

Manuel,
When I said 'objective' I was assuming that you would have something more definitive about how exactly you're going to use this information; at least a statement about how precisely you might want to estimate any potential impact.

Sorry, I may have been confused in what you were calling a site. These days i try to refer to the landscape unit at which you want to establish the presence or absence of the species as the sampling unit, which in our 2002 paper we called a site. It sounds like you intend this to be the small rectangular plots? Your sampling unit must then be repeatedly surveyed to address the detectability issue.

However, it sounds like your species are very wide-ranging (even larger than the forest tracts). Are they very territorial, or do they have overlapping home ranges? What sort of effect would you expect logging (if any) to have on your species? Would they just start avoiding those areas?

In terms of your sample size questions, really you need to either look through the literature on analytic results, or do some simulations to assess whether the sample sizes are going to be big enough. Two starting places are our book, and the recent paper by Bailey et al. (2007) Ecological Applications.

Finally, (and this a general reminder to all readers) when you start asking very specific questions about your study design (or analysis for that matter), then you're starting to push the boundaries of the intention of this forum (see the FAQ). While most of the people I know that work in this area don't mind offering a bit of free advice and direction, any post that would require much more than 5-15 mins of thought for a reply is likely to go unanswered. If after doing all of your own research, have followed up on suggestions from here (or other sources) and still need help, then perhaps you need to start thinking about hiring or collaborating with someone more closely. I recall a quote from this forum a number of years ago (possibly posted by Evan) "Free advice is worth every cent you paid for it."