New approaches to modelling drivers of species distribution and abundance in the Southern Ocean
Inferring ecological patterns from marine survey data is difficult due to the large spatial and temporal scales at which processes operate and the challenges associated with collecting comprehensive and balanced survey data. In this thesis I use large scale survey data and cutting edge modelling tec...
| Main Author: | |
|---|---|
| Format: | Text |
| Language: | unknown |
| Published: |
Macquarie University
2022
|
| Subjects: | |
| Online Access: | https://dx.doi.org/10.25949/19427189.v1 https://figshare.mq.edu.au/articles/thesis/New_approaches_to_modelling_drivers_of_species_distribution_and_abundance_in_the_Southern_Ocean/19427189/1 |
| Summary: | Inferring ecological patterns from marine survey data is difficult due to the large spatial and temporal scales at which processes operate and the challenges associated with collecting comprehensive and balanced survey data. In this thesis I use large scale survey data and cutting edge modelling techniques to examine the drivers of species distribution in the Southern Ocean at three trophic levels - primary producers, grazers and top predators. I develop a model to predict phytoplankton abundance in a 3D environment from temperature, salinity and depth. This framework is widely applicable to other marine settings regardless of their survey design and provides a robust method for dealing with complex data sets. An important grazer on phytoplankton, Antarctic krill (Euphausia superba), has previously been regarded as passively drifting with large scale current systems. I provide quantitative evidence that they actively swim, demonstrating that krill consistently aggregate around resources over an immense survey area spanning 1.3 million km2. Krill distribution is patchy, and predators must locate these dynamic swarms across vast expanses of ocean. Islands may provide predictable and reliable feeding areas due to the Island Mass Effect. I find that krill swarms at the Balleny Islands, a Southern Ocean archipelago, are three times more numerous than in the adjacent open ocean, and are also denser and more compact. Around the islands, humpback whales (Megaptera novaeangliae) aggregate in areas of high productivity, medium krill density and waters greater than 350m deep. Two chapters of this thesis required manual processing of active acoustics data for detecting krill, which is time consuming and suffers from a lack of reproducibility. To automate this process, I developed an R package which drastically reduces processing time and is useful for any scientists using acoustic data. This thesis fills knowledge gaps about the mechanisms structuring the distribution of animals in the Southern Ocean and the statistical methods and software library developed are applicable to many other problems arising in complex environments. |
|---|