Estimating the proportion of Archey’s frogs in the Wharekirauponga mine vibration footprint 24 Appendix: Log-linear Capture-recapture Modelling in Rcapture Models were fitted using the function closep.t in the R-package Rcapture[16, 17]. The fitted models can incorporate up to three sources of variation among capture probabilities: a temporal effect (t), heterogeneity between units (h) and a behavioural effect (b). Temporal effects result in capture probabilities varying among capture occasions. Heterogeneity causes capture probabilities to vary among individuals. A behavioural effect means that first capture changes the behaviour of an individual, so the capture probabilities before and after the first capture differ. These sources of variation lead to eight fundamental closed population models: M0 (no source of variation), Mt, Mh, Mth, Mb, Mtb, Mbh, and Mtbh. Three different methods are used for heterogeneity models (Chao, Darroch and Poisson2), providing twelve different models (Table 13). The results of closed-population modelling of capture-recapture data from Hotham (2019) are provided in Table 13 with models listed in order of their Akaike’s information criterion score (AIC) for three different data sets. Columns in Table 13 are: N: the estimated population size. SE: the standard error of the estimated population size. Deviance: the model’s deviance. df : the number of degrees of freedom. AIC : the Akaike’s information criterion. BIC : the bayesian information criterion. InfoFit : a numerical code giving information about error or warnings encountered when fitting the model: 0 - no error or warning occurred while fitting the model; 1 - a warning indicating that the model fit is questionable (e.g. the algorithm did not converge, non-positive sigma estimate for a normal heterogeneous model or a large asymptotic bias). Population estimates used in Table 6 were obtained from abundance estimates for the models with the lowest AIC value in Table 13. In all cases the model including a behavioural effect (Mb) had the lowest AIC and BIC values.
RkJQdWJsaXNoZXIy MjE2NDg3