The population dynamics of a riparian spider: interactive effects of flow-related disturbance on cross-ecosystem subsidies and spider habitat
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
The transfer of prey resources between ecosystems can have dramatic consequences for both recipient and donor systems by altering food web stability and the likelihood of trophic effects cascading across the ecosystem boundary. Landscape-scale factors influence the importance, direction and magnitude of energy flows, but may also alter the ability of consumer organisms to respond to spatio-temporal changes in allochthonous prey availability. Here, I used flood and drying disturbance gradients to investigate interactions between these two processes on populations of a riparian fishing spider Dolomedes aquaticus (Pisauridae). The abundance of aquatic insects with a winged adult stage, a major component of the diet of D. aquaticus, was markedly higher at less flood-prone rivers and declined with increasing flood disturbance. It was expected that spider populations would be largest at these stable rivers where the aquatic prey abundance was highest. However, a habitat (loose, unembedded riverbank rocks) manipulation revealed that the lack of scouring floods at these sites led to habitat-limited populations, preventing response to the increased prey resource. In fact a peak shaped relationship of spider biomass and abundance was found, with the largest spider populations at intermediately disturbed rivers. In addition, patchy habitat availability was the most likely cause of the small scale (4 m2) aggregation of spiders seen at the most stable and disturbed rivers. These patterns were also associated with strong interactions between the spiders. Stable isotope analysis of field collected spiders and an experimental manipulation of spider densities and food availability indicated that cannibalism rates were likely to be significantly higher at stable and disturbed rivers than those intermediate on the disturbance gradient. Differences in D. aquaticus population size structure and life history traits across the flood disturbance gradient were driven by interactions between resource availability, environmental stability and cannibalism rates. To separate the effects of habitat availability and aquatic prey abundance I used drying rivers, as the amount of aquatic insect prey alters as the water recedes. Desiccation mortality and low aquatic prey biomass most likely caused the spiders' spatial distribution and size class structure to alter in drying river reaches, potentially also leading to differences in cannibalism rates. Overall, cross-ecosystem transfers of prey had large impacts on the distribution, cannibalism rates and life history traits of D. aquaticus but their effects were modified by the nature of the ecosystem boundary. Thus river flow regime controlled the magnitude of the subsidy and its use by a consumer. Hence, cross-ecosystem subsidies will not always lead to larger consumer populations and consumer responses will depend on interactions between large-scale processes.