22 – The challenge of monitoring coastal marine mammals^297extension of the closed robust design (CRD) model (Pollock 1982; Kendall and
Nichols 1995) and the multistate model for recapture data (Arnason 1972; Brownie
et al. 1993). The MSCRD model provides estimates of:
● (^) apparent survival between primary samples (probabilities of being alive and
present in the sample area)
● (^) movements between sites and temporary emigration between primary samples
● (^) abundance at each primary sample location.
The CRD model deals with only one site at a time (Bynoe Harbour, Darwin
Harbour and Shoal Bay are each considered separately or together as one site). In
contrast, the MSCRD model can simultaneously provide these estimates for
multiple states (here multiple sites, Bynoe Harbour, Darwin Harbour and Shoal
Bay). A full description of the project including its methods and results is available
in Brooks and Pollock (2015). Abundance estimates for humpback dolphins are
plotted in Fig. 22.4. Survey work is shown in Fig. 22.5.
Wet and dry season effects were often observed on the apparent survival and
temporary emigration estimates. Estimates of apparent survival and temporary
emigration often had wide confidence intervals, which, together with the paucity
of knowledge of biological survival rates for Australian coastal dolphin species,
indicate that the estimates of rates of permanent emigration should be interpreted
Fig. 22.4. Estimated number of humpback dolphins present in each primary sample with 95% confidence
interval on all sites combined (CRD model), and for each site (Bynoe Harbour, Darwin Harbour and Shoal Bay:
MSCRD model). UCI, LCI = upper and lower bounds of the 95% confidence interval, respectively. Figure
adapted from data presented in Brooks and Pollock (2015).