SE calculation for derived parameters from model averaging

[Eurycea]

I understand that for derived parameters, the delta method is used to calculate the SE. Is this also the case for derived parameters from model-averaged estimates? PRESENCE only outputs a VC matrix for singular models, and I'm not sure how to apply this in a model-averaging scenario.

[Eurycea]

D. MacKenzie responded to me off list so I thought I'd post the response:

You'd have to calculate the model averaged VC matrix yourself by using the VC matrices from the individual models. That said, I generally suggest that people calculate the things of interest within each model first, then model average the metric from the different models. That way any correlations and covariances associated with the calculation of the metric are all taken care of within the individual model structures, and avoid you trying to calculate a model-averaged VC matrix.
Cheers
Darryl