|dc.description.abstract||Land use conversion, especially to intensive agriculture, has been identified as the main driver of loss of biodiversity and ecological degradation in freshwater ecosystems worldwide. Intensive agriculture introduces multiple stressors to streams, including elevated levels of fine sediment and nutrients. Therefore, resource managers need to understand the effects of these stressors on aquatic ecosystems. In the present study, I combined two stream surveys and a novel reach-scale sediment addition/removal experiment in two New Zealand streams to investigate effects of dairy farming prevalence in the stream catchment (surveys) and surplus deposited fine sediment due to intensive land use (experiment) on stream physicochemistry, invertebrates and fish. One particular focus of my research was on an economically and recreationally important fish species, the introduced brown trout (Salmo trutta L.).
In the surveys I hypothesized that, in parallel with rising prevalence of dairy farming in the catchments of the studied streams, (1) dissolved nutrients and deposited fine sediment levels will increase, (2) macroinvertebrate community health metrics will decline, and (3) densities of exotic trout (and possibly also native fish) will decrease. In the most novel objective of my second stream survey, I also predicted that (4) available invertebrate food for brown trout will decrease, resulting in less prey biomass eaten and poorer fish condition.
My survey results demonstrated that increasing dairy farming prevalence was indeed associated with rising levels of nutrients and fine sediment, whereas invertebrate community health metrics and the prevalence of certain pollution-sensitive invertebrate taxa declined. Also as predicted, densities of brown trout and native fish declined as dairying became more prevalent, with no trout found in streams where dairy farms covered 50% or more of the catchment area. Interestingly, however, trout food intake and condition were unrelated to land use intensity in the catchment and trout and native fish became rarer.
In my experiment, I added or removed fine sediment in 50-m reaches of two streams; the first time such a reach-scale manipulation focusing on fish has been performed. Here I hypothesized that (1) adding sediment will have negative effects on stream health metrics and will initially increase drift rates of sediment-intolerant taxa (which will subsequently become rarer in the benthos and also in the drift), whereas removing sediment from the streambed will have the opposite effects, (2) sediment addition will negatively affect fish density in general and brown trout density and condition in particular while sediment removal will result in the opposite patterns, and (3) sediment addition will adversely affect available food for trout (via lower densities of preferred prey items, such as sediment-intolerant mayflies, in the drift) and thus trout condition, whereas sediment removal will have the opposite effects. To test these hypotheses, I used a repeated-measures design with a Before-After-Control-Impact (BACI) approach and three experimental treatments. To simulate an increase in farming intensity coinciding with higher sediment inputs to streams, I added natural fine sediment to one reach per stream. To simulate a decline in farming intensity, I removed sediment from another reach per stream by water-blasting. An unmanipulated 50-m reach per stream served as a control. Response variables including nutrients, fine sediment, benthic and drifting invertebrates and fish density (estimated by spotlighting) were measured repeatedly, and all reaches were electrofished once at the end of the 2-month experiment.
In the experiment, sediment addition reduced total taxon richness of benthic invertebrates while sediment removal increased it, resulting in richness being highest in removal, intermediate in control and lowest in addition reaches. In contrast, the studied invertebrate drift metrics showed no clear responses to my manipulations. Further, removing deposited fine sediment improved the in-stream habitat for trout and resulted in higher trout density, whereas adding fine sediment had the opposite effect. These results imply that fish communities in my study streams generally preferred habitats without high levels of deposited sediment. Besides the observed patterns for fish density, trout condition also responded to my manipulations, with trout in sediment addition reaches being in poorer condition than those in removal or control reaches at the end of the experiment.
One of the two key management implications of my thesis is that continuing land use conversion to intensive agriculture, especially to dairying with poor management practices, is likely to result in high fine sediment and nutrient inputs to agricultural streams, which can negatively affect both stream invertebrate and fish communities including brown trout. Conversely, reducing fine sediment inputs to farmland streams and minimizing erosion in tributary catchments are likely to improve trout fisheries in these streams.||