N-mixture model performance and optimal survey design

Choose a scenario of local abundance in sites \(\lambda\) and individual availability probability \(\phi\) to see single and double observer models performance (relative RMSE curves) and optimal survey effort allocation (optimal number of visits).

   

 

Improving accuracy in abundance estimation by using a double-observer protocol

Increase in model accuracy by using a double-observer protocol when compared to a single observer model (as a baseline).

Figure: Increase in model performance with the proportion of a double-observer protocol in image review

 

Reducing fieldwork effort with double observer review

Proportional reduction in fieldwork effort (from \(B=2000\) flights) that can be achieved with a full double-observer protocol, while matching the same accuracy of the single observer model (for \(p=0.8\)).

 

Proportion of the total effort (budget) of the double observer model (multinomial N-mixture) needed to obtain a similar performance to the single observer model (binomial N-mixture)

 

 

For more details on the simulation methods and results check out the original paper.