Occupancy & ground versus canopy sampling of trees

[hf.hwa]

Hello - We are sampling for a forest pest, which offers some challenges to estimating its occupancy since it often initially infests tree canopies.

We used the following sampling design: Three independent observers visited trees on the same day and noted detection / non-detection of the pest on branches that can be reached from the ground. An additional observer sampled the canopies (all branches that could not be reached from the ground) of a subset of trees using a lift and noted detection / non-detection. For logistical reasons we could not survey the canopies of all trees nor could we perform multiple surveys of any of the canopies. The canopy survey was independent as the observer had no knowledge of the findings of the ground surveys. Thus, a subset of trees was surveyed four times, 3 from the ground, and once from the air.

We would like to estimate, using our subset of canopy surveys, the true infestation rate of trees given that some infestations will go undetected in the canopy (and, of course, at the ground).

Is it appropriate to run a single-season, single-species model using the 3 ground surveys & compare this to a model including the 4th survey of the canopy? Any other ideas?

Many thanks for your help!
Matt

[darryl]

I think that should be ok, just make sure you have different detection probabilities for ground and canopy surveys.