Livestock production and stream health in the Great Lakes Basin: An agroecosystem health approach.

摘要

An agroecosystem health framework is presented as a new approach for assessing livestock impacts on stream health in the Great Lakes Basin (GLB). Agroecosystems are defined as complex, open, self-organizing systems, occurring within a nested hierarchy. They are characterized by biophysical and socio-economic dimensions, and by complex interrelationships and feedback loops within systems, between dimensions, and between levels of the hierarchy. Agroecosystem health is assessed in terms of integrity, efficiency, adaptability, and effectiveness, using indicators from different scales and dimensions of the hierarchy. A theoretical model of the links between livestock production, stream health and farm economics identified three pathways of effect: manure, grazing, and feed crop production.; A conceptual model case study was conduced on Carroll Creek, in Ontario. Livestock access to the creek and exponentially decayed livestock density were inversely associated with several stream health indicators, including an Index of Biotic Integrity of fish (IBI) and a stream habitat index.; Agricultural census data were re-aggregated into watersheds, yielding livestock and economic indicators for a GLB watershed study. A need for universal stream health indicators was identified in the lack of consistency between monitoring protocols from six GLB jurisdictions. Fish and benthic macroinvertebrate indices were only weakly associated, suggesting these communities reflect different aspects of stream health. Two indices were applied to all available GLB data: a taxonomic order-level Hilsenhoff Biotic Index (invertebrates) and a modified GLBIBI (fish), which was validated against existing fish indices. Associations between these indices and livestock indicators were tested using multi-level mixed effects regression models. Swine density and corn area were associated with Hilsenhoff score (higher score indicates more tolerance to pollution). GLBIBI was negatively affected by poultry density and soybean area, and the effect of corn was equivocal when poultry density was controlled. Significant statistical interaction occurred between State of origin of the data and some covariates.; Indicators of manure impact (livestock density) and cropping effects (grain area) seemed more strongly associated with biotic indicators of stream health than grazing indicators, especially at the watershed scale. The findings suggest that biophysical integrity of streams is decreased in areas of intensive swine, poultry, and cash cropping in the GLB.