Preliminary Steps

We assume that the user has started a new session, loaded a data file and selected the Occupancy option. In the ‘Data status’ section of the main window, the ‘Modify’ button allows to specify the characteristics of your model. The number of age classes should be fixed to 1 here. Age is the time elapsed since first detection, which is equivalent to time if all sites start being monitored at t = 1. If site-specific covariates are to be used, then the number of individual covariates should be amended accordingly (recall that the sites are the equivalent of individuals in capture-recapture analyses). The number of events is the number of observations (e.g., undetected and detected), while the number of states should always be the number of states to be used (e.g., unoccupied, occupied by non-breeders, occupied by breeders) plus one. Indeed, because E-SURGE was initially developed to estimate demographic parameters, it always considers the absorbing state ‘dead’ which is useful in a capture-recapture context to have individuals die but of little interest in an occupancy analysis.

In the ‘Advanced numerical options’ section, the 'Compute C-I' option is deactivated by default to save time by avoiding calculating the parameters’ confidence intervals of each model. Just tick this option to get standard errors and confidence intervals, e.g., for the best fitting model. Yet in the same section, in the pull-down menu ‘Initial values’, E-SURGE offers the possibility to use several sets of initial values randomly chosen or use estimates of a simpler model (from last model) as initial values, which might be useful to avoid picking up local minima in the deviance, an issue often encountered in Hidden Markov Models.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License