INTRODUCTION

Bottlenose dolphins (Tursiops spp) are one of the most studied genera of cetacean mainly due to their widespread distribution and ease of accessibility, and as such, much is known about their biology, social behaviour and population dynamics (Wilson et al. 1999b, Connor et al. 2000b, Reynolds et al. 2000b). Standard methods for determining the abundance of dolphin populations have been developed and used routinely. However, we should continue to refine these methods based on our current and growing understanding of dolphin biology and behaviour. This will ensure that abundance estimates and population parameters are reliable and accurate (Silva et al. 2009) and applicable to the Indo-Pacific bottlenose dolphin (Tursiops aduncus) where few abundance estimates are available.

2.2.1 Dolphin population dynamics

It is becoming clear that many coastal bottlenose dolphin populations include individuals with varying patterns of residency and home range size. Some individuals reside in an area, others pass through only occasionally (Möller et al. 2002, Fury & Harrison 2008, Silva et al. 2009) and others temporarily emigrate leaving the area for a period of time (e.g. seasonally), but return more frequently than transients. The varying degrees of residency have a large impact on abundance estimates at any given time and need to be recognized and incorporated into study design and data analysis through models.

Resident individuals may be particularly vulnerable to anthropogenic pressures such as habitat degradation, environmental contaminants, marine debris, anthropogenic noise, vessel traffic, illegal feeding, harassment, ingestions of fishing gear, competition with fisheries for prey and entanglement in fishing gear (Nowacek et al. 2001, Watson-Capps & Mann 2005, Bejder et al. 2006b, Wells et al. 2008, Kucklick et al. 2011). Accurate assessments of abundance, distribution and life history parameters such as survival, recruitment and residency are all critical to understanding a population and its use of a particular habitat as well as recognizing the impacts of anthropogenic or natural pressures.

2.2.2 Estimating abundance – capture-recapture

For most dolphin species, particularly Tursiops, photographs of the dorsal fin are used for individual identification (Würsig & Würsig 1977). Fin shape and distinguishing marks, nicks and scars allow long-term identification and markings on the surface of the fin allow short-term identification (Scott et al. 2005). Initial cataloguing of each identified fin is the capturing process, with later resightings of the fin used as the recapture events. As dorsal fins may change over time it is necessary to record these changes and track individual fin modifications to avoid mis-identification (Yoshizaki et al. 2009).

While the capture-recapture methods have remained the standard for identifying animals, there are a variety of sampling protocols for collecting the captures. These involve standard transect line surveys conducted over zones of the study area using varying temporal and spatial sampling designs. To take into account residency patterns and the opportunities available to capture all individuals within the population, careful thought must be given to the sampling regime, including sampling periods for and between captures and re-captures.

2.2.3 Population models

Models used in capture-recapture studies are statistical models used to calculate abundance estimates over multiple sampling periods (Williams et al. 2002b). Models are useful because they allow calculation of estimates under complex parameterizations. The sampling design must meet all of the model assumptions and the parameters must make biological sense. These parameters should be customized as either constant or time varying according to the characteristics and life history of the study population. The traditional definition of a population is “a group of organisms of the same species occupying a particular space at the same time” (Krebs 2001). In this study I considered the population to include any bottlenose dolphins within the defined study area (Figure 1.1) during the study.

Closed versus open models

There are two general types of models that have been used to describe bottlenose dolphin populations: closed and open. The closed population model is used when there is no change to population abundance during the sampling period. Closed

models are best suited to species where sampling can be conducted over a short period (e.g. a week) to avoid violation of this assumption.

Open population models have been used more recently in cetacean studies (Straley et al. 2009). They allow for increases (births, immigration) and losses (mortality, emigration) to the population (Williams et al. 2002b). However, multiple movements in or out of the population, known as temporary emigration, is accommodated in the Robust Design model (Kendall & Nichols 1995, Kendall et al. 1997, Pollock 1982).

The Robust Design model combines both open and closed population models (Pollock 1982, Williams et al. 2002b). It incorporates open sampling events called primary periods within which are a number of closed secondary periods (Pollock 1982). The Robust Design relies on these secondary periods being closed so that closure can be assumed within each primary period (Kendall 2004). The capacity of the Robust Design model to allow for temporary emigration is a very useful feature when estimating abundance for species that exhibit movement in and out of the study area.

The aim of this study was to use a sampling design and population model that addresses the complexity posed by a coastal dolphin population that includes both resident animals and temporary emigrants, and produce an accurate abundance estimate for baseline information. In this study, I employed the Robust Design model to examine annual and seasonal dolphin abundance and the implications that residency and temporary emigration have on the accuracy of abundance estimates.

2.3 MATERIALS AND METHODS