Using tooth growth chronologies to investigate responses of marine mammals to variability in the marine environment

In a changing climate, it becomes increasingly important to understand how species respond to variability in their environment, and how their responses might influence overall population dynamics. Such information is important for the assessment of impacts on species, and for modelling variability in ecosystems, however obtaining this information for long-lived, cryptic marine mammal species is difficult. Growth layers (termed growth layer groups: 'GLGs') laid down annually in the teeth of marine mammals can provide a proxy record of the energy budget of individuals; the size of each GLG is a factor of the metabolic energy available for the deposition of the structures associated with each. Chronologies developed from time series of measurement of GLGs, similar to those developed from tree-rings or otoliths, have the potential to be compared with environmental variables, thereby providing insights into environmental drivers of the energy budgets of marine mammals. Although a number of attempts at developing chronologies have been conducted on pinnipeds, they have all varied in their methodology and complexity and have in most cases, been developed for single species, limiting both their repeatability and comparability among studies. Further, the life spans of study species and subsequently their associated chronologies have been relatively short, reducing the ability of these methods in their applicability for investigating the responses of individuals to longerterm environmental variability. The responses of cetacean species to variability in their environment are poorly understood in the southern Australian/New Zealand region, particularly for wideranging pelagic odontocete species such as long-finned pilot whales (Globicephala melas) and sperm whales (Physeter macrocephalus). This thesis therefore aims to: use dendrochonology (tree-ring science) techniques to establish a standardised methodology for development of tooth growth chronologies, and test the efficacy of these chronologies for investigating responses of long-finned pilot whales and sperm whales to environmental variability. These aims are achieved utilising long-finned pilot whale teeth sourced from New Zealand and Australia and sperm whale teeth sourced from two regions in Australia: Western Australia and Tasmania. Investigation of the relationship between tooth growth chronologies and environmental variables comprised comparisons at two scales: i) Broad-scale annually averaged climate indices of relevance to the region (Southern Oscillation Index: SOI, Southern Annular Mode: SAM, Indian Ocean Dipole: IOD and the Fremantle sea level: FSL) ii) Seasonally averaged and spatially gridded environmental measures (sea surface temperature: SST and zonal wind speed). By modifying the dendrochronology 'list method', I established a method to crosscheck GLG identification within individuals and refine age estimates. I then established a repeatable approach to the measurement of GLG widths, suitable for both acid etched teeth (sperm whales) and thin sectioned and stained teeth (pilot whales). Dendrochronology detrending techniques were then trialled on time series of GLG measurements to standardise time series and produce chronologies for each individual. Cubic smoothing spines provided the best fit to the GLG width data for both species with both linear models and negative exponential curves found to be unsuitable. Chronologies derived from each of the individual whales from each sample site were then averaged within calendar years to produce master chronologies at the whole group (composite chronology) and each sample group level, thereby enhancing common environmental signals and reducing the influence of individual variability. Sample site chronologies for pilot whales were 10 and 12 years for the New Zealand and Australian strandings, respectively. A 70 year time series spanning 1935 ‚- 2004 was established for sperm whales by combining tooth samples from the 1960s (whaling archives), 1990s and 2000s (strandings). Regional sample site chronologies for sperm whales were 30 years in length for the southwest Australia samples, 22 years for the Flinders Island stranding, and 38 and 20 years for the 1998 and 2004 Strahan strandings, respectively. (i) Broad-scale annually averaged climate indices of relevance to the region: Generalised additive models (GAMs) were used to investigate relationships between broad-scale climate indices and tooth growth. GAMs including the sperm whale composite master chronology as the response variable resulted in poor explanatory power of environmental predictors. Subsequent environmental comparisons were therefore conducted on sample site chronologies and revealed stronger and contrasting relationships among sample groups. A positive relationship with the SOI at a one year lag was observed for the New Zealand pilot whale chronology. The Australian pilot whale chronology demonstrated a positive relationship with the SOI and a negative relationship with the IOD. Environmental conditions associated with mutual interactions of the SOI and IOD result in changes in wind and increases in storm activity across the southern Australian region, which may drive increased productivity through mixing and upwelling of productive deep waters. These relationships suggest that the more westerly population of pilot whales is influenced by a mixture of both the SOI and IOD, with low or no influence of the IOD on populations to the east that utilise New Zealand waters. A negative relationship with the IOD was identified for sperm whales from southwest Australia, while modelled relationships suggested a combination of a positive relationship with the SAM and negative relationship with the IOD were important drivers of tooth growth and energy budgets for the Flinders Island group. In contrast, a positive relationship with the SAM as a single predictor variable was observed for the 1998 Strahan group. Both negative IOD and positive SAM conditions lead to increased westerly winds and storm activity in the southern Australian region, resulting in higher productivity via increased upwelling and mesoscale activity in frontal zones. A positive relationship with the SOI and varying relationship with the IOD (decreasing with negative IOD values and increasing with positive IOD values) were observed for the 2004 Strahan group. Positive SOI conditions are associated with increased storm activity and variations in strength and SST of major currents in the Australian region, leading to regional increases in biological productivity. The effects of the SOI in combination with the IOD are likely to vary depending on the connectivity of SOI and IOD events. Differences in environmental drivers among sample sites suggest spatial variability in foraging patterns among sperm whale social groups, leading to differential responses in energy budgets, most likely associated with spatial variability in prey availability. (ii) Seasonally averaged and spatially gridded environmental measures: Correlations between the regional chronologies and seasonal averages of gridded sea surface temperature and zonal wind speed were carried out across a domain bounded by 30¬∞S ‚- 60¬∞S, 94¬∞E ‚- 190¬∞E. Spatial maps of correlations were then produced to identify potential areas associated with higher and lower tooth growth (energy budgets). Areas associated with higher tooth growth for pilot whales from Australia consistently occurred in frontal regions south of Tasmania across all seasons, whereas areas associated with higher tooth growth for New Zealand pilot whales varied among seasons. These differences in spatial correlations suggest consistent utilisation of highly productive frontal regions by the more westerly population of pilot whales, whereas the population in the east may alter foraging movements in response to seasonal variability in regions of high prey biomass. Areas associated with higher tooth growth for sperm whales were consistent with regions of known sperm whale habitat (e.g. southwest of Western Australia and frontal regions south of Australia). Both regions are influenced by variability in the SOI, the IOD and the SAM, with effects on SST, wind direction and storm activity resulting in changes to food webs. Spatial correlations for the southwest, Flinders Island and 2004 Strahan groups consistently occurred in the region southwest of Western Australia in austral winter, but regions associated with higher tooth growth differed across other seasons. In contrast, the 1998 Strahan group showed a consistent pattern of higher tooth growth associated with frontal regions south of Australia across all seasons. Differences in spatial correlations among sample sites suggest previously unknown differences in habitat preferences between sperm whale groups in the southern Australian region. Conclusions: Through the adaptation of dendrochronology techniques to time series of GLG widths a standardised, transparent and repeatable methodology broadly applicable to long-lived mammal species has been developed. For the first time, relationships between variability in annual GLG widths and broad-scale climate indices have been identified in Southern Hemisphere odontocetes. In doing so, this study has provided a step change in our ability to quantify how a changing environment might influence populations of these and other marine mammals in the future.