Abstract
Species assemblages, which are groups of related species sharing space and time, are useful surrogates for monitoring environmental changes. Marine species have been shown to modify their distribution patterns due to changes in ocean conditions. Establishing baselines, however, is an essential first step to effectively monitoring these shifts. Seabirds have been used as sentinel species to detect changes in ocean ecosystems for decades. The Australasian region is a hotspot of diversity and abundance of seabirds. Surprisingly, however, very little is known about the distribution of seabirds in this region, especially at the assemblage-level.
In this thesis, I aim to identify and describe species assemblages in Australasia, using cutting-edge multivariate statistical methods. I analysed three datasets, each with distinct spatial, temporal and methodological characteristics: (i) an existing dataset from eastern Australia [Chapter 2], and new data from (ii) Otago, Aotearoa New Zealand [Chapter 3], and (iii) Northland, Aotearoa New Zealand [Chapter 4].
In Chapter 2, biogeographic models were fit to identify and describe seabird assemblages over waters of the East Australian Current (EAC) at the mega/macro-scale, from the Coral Sea to the south of Tasmania. I fitted seasonal models using presence–absence and relative abundance (count) data, including oceanographic and physiographic parameters as explanatory variables. All models suggested two macro-scale assemblages (‘northern’ and ‘southern’), except for the autumn presence–absence model that identified three groups. Sea surface temperature and/or salinity were always retained as predictors of assemblages, primary variables for characterising the EAC. In addition, the spatial predictions of the models were consistently associated with the behaviour of the EAC, showing a ‘transition zone’ between assemblages at ~34°S, where the EAC detaches from the shelf break and turns eastward to form a trail of eddies (the ‘Tasman Front’).
In Chapters 3 and 4, I investigated oceanographic, physiographic and temporal drivers of seabird assemblages. Model-based ordinations revealed a strong influence of seasons in shaping seabird assemblages in Aotearoa New Zealand, where season alone explained species composition better than oceanographic parameters. The frequency of occurrence and relative abundance of species across seasons clearly showed that there was a high number of migratory and wide-ranging dispersive species in both study areas. Therefore, the seasonality of these species’ movements directly influenced species occurrences and, thus, the makeup of the assemblages. Furthermore, off Northland, the influx of migratory species massively influenced estimated seabird biomass (an almost eight-fold increase in biomass from winter to summer). This result shows that seabirds may affect the ecosystem functioning of the region, for example by dispersing essential micronutrients in the top layer of the water column. In addition, based on nine years of seabird observations, I demonstrated that nearly 40% of the species are changing their frequency of occurrence off Otago, likely reflecting large-scale changes in regional oceanography.
Chapters 2–4 provide the first quantitative assessment of seabird assemblages in those regions. To test whether seabird distribution may be altering due to global warming, continued monitoring of the ‘transition zone’ identified off New South Wales, Australia, and off Northland and Otago, Aotearoa New Zealand, would be valuable. The results
presented herein can now be used as baselines to monitor for possible changes.
Prompted by the different survey methods used in Chapters 2–4, in Chapter 5, I zoomed out from ‘seabird assemblages’ and focused on the ‘methods’ to count seabirds at-sea during vessel-based surveys. I conducted a literature review for the period of 1971–2020. I gathered over 800 papers and extracted data from 100 randomly sampled studies to investigate methodological choices over time. Strip-transects were overwhelmingly the preferred method, but some differences in protocols emerged. To understand the pros and cons of each method and protocol, I summarised 50 years of literature. Key points to consider when planning and executing vessel-based surveys are: (i) identify the spatial and temporal scales of the research question; (ii) understand oceanographic processes operating at the identified spatial and temporal scale; (iii) weigh the pros and cons of each method and protocol to address your research question confidently; (iv) consider the logistics of surveys (number and frequency of surveys required, number of observers, how to measure distances from vessels, use of software to aid recording of data). Finally, I provide a best-practice protocol for counting seabirds at-sea from vessels, which embraces most logistical and analytical constraints. The recommended method essentially merges the strengths of strip-transects and distance sampling protocols: record seabirds continuously, also noting their behaviour, and use multiple strips to allow for calculating detectability and therefore correcting counts. The recommended protocol increases the potential for data integration across datasets and allows a broader range of analyses.