|dc.description.abstract||The distribution and abundance of fish species is influenced by a range of environmental variables. Information on driving factors in that determines fish habitat use, abundance and distribution is crucial for fisheries and restoration management, especially for species in gradual decline. Habitat use may be structured by interactions between multiple factors including food density, predation risk, competitors, energetic requirements and habitat structure. Habitat selection varies with respect to species, stage of life history and time of day, agonistic interactions between conspecifics may also influence habitat use.
In this study, I selected three streams draining to Otago Harbour in the South Island, New Zealand; Craigs Creek, Deborah Bay Stream and Sawyers Bay Stream. A gradient in riparian vegetation from sparse (Craigs) to dense (Deborah) was the basis of stream selection and they likely differ in food quantity and quality which may affect fish assemblages. I used banded kokopu (Galaxias fasciatus) as an animal model. The fish represent ideal species to study feeding strategies, growth and movement as they appear to form dominance hierarchies that are stable over long periods of time within stream pools. They are also easy to catch, observe and individually tag. High fish density in Craigs Creek and low food density in Deborah Bay Stream and absence of trout in all streams provides an ideal situation to study the role of social hierarchies and intraspecific competition for food and space on patterns of individual growth, residency and movement. Thus the main goals of this dissertation were to elucidate the long-term patterns of banded kokopu habitat selection, growth and movement in relation to their social status, energetic requirements and food supply. Local habitat selection was measured at microhabitat scale to determine if large fish share habitat with or segregate from small conspecifics. Growth of fish across sizes (hierarchical ranks) was investigated to determine whether agonistic encounters influence habitat use and variation in growth among individuals. Food density and quality (energy availability) and metabolic rates of fish (energy requirements) were measured to assess whether stream pools were capable of providing sufficient energy. Fish movement patterns were monitored to determine whether the fish would change their habitats if energetically unsuitable.
This thesis tests several hypotheses. I assumed that a social dominance hierarchy exists among banded kokopu with larger fish outcompeting smaller fish to gain access to the most profitable patches and this would impact smaller individuals’ growth. I also hypothesised that the banded kokopu population would exhibit evidence of energy limitation based on an assessment of energy consumption in relation to energy supply in the form of drifting invertebrate prey. Fish of low social status would exhibit more movement in an effort to improve their social rank and gain access to more food in new pools.
The results of my first study of banded kokopu habitat use showed that water depth, focal velocity, adjacent fastest velocity and substrate were the factors most strongly associated with banded kokopu position. Patch use by fish was clearly size-related with small fish occupying shallow pools with fast velocities over fine substratum and large fish selecting deeper pools with slower water velocities and coarse substratum. Although the different patterns of habitat use by individuals may be due to ontogenetic shifts in habitat preferences, evidence from multiple published studies suggests that observed pattern of microhabitat partitioning is likely due to agonistic intraspecific interactions and a social hierarchy among individuals of varying size.
Social hierarchy in banded kokopu assemblages appears to influence individual feeding and growth. The growth of high ranked fish in the social hierarchy (large fish) was correlated with food abundance and biomass (energy available in habitat). This emphasises the ability of large fish to outcompete small ones given their high status in social hierarchy.
The banded kokopu diet study showed that the fish favour food items with high caloric values. Invertebrate drift was insufficient to meet banded kokopu energetic requirements; hence fish must access other sources of energy in the habitat such as terrestrial invertebrate inputs and benthic prey. The fish appeared to be likely energetically constrained in all seasons as they could obtain no more than 70% of their total energy expenditure from drifting invertebrates alone, which based on observations of their feeding behaviour, was their primary food source.
Akaike’s information criterion (AIC) showed that the best candidate model used to predict fish abundance was the one that included food density, pool area, undercut banks and water velocity. Pool size determines the number of small fish, and the energy available determines total fish biomass in pool and also growth of large fish. The growth of subordinate fish is determined by an interaction between food supply and the large dominant fish in the social hierarchy.
Dominant fish exhibited higher growth rates and a reduced tendency to move compared to subordinate fish, suggesting that high-ranked fish monopolise food intake to some extent in pools. In contrast subordinate fish changed pools more often, most likely as consequence of intraspecific competition that resulted in lower growth rates. The growth of fish improved after relocation to a new pool only if a higher social status was gained in the new habitat, suggesting that fish movement was intraspecific competition-oriented rather than food-related.||