Birds as indicators of agroecosystem resilience and sentinels of market accreditation: A case study of New Zealand South Island sheep/beef farms

New Zealand is unique among industrialized countries in that, in the absence of farming subsidies, farm management is governed largely by consumers in Europe, the United States and Japan, where the majority of agricultural products are exported. The neoliberal context in which New Zealand agriculture is presently situated may provide the most likely vehicle through which biodiversity care on production lands may be introduced. As consumers in key niche markets abroad express increased preference for “green” commodities produced using sustainable and biodiversity-friendly farming practices, New Zealand farmers may be required to alter their farm management to include provisions for biodiversity care, or risk losing access to international markets. There is presenty a great need to identify ecologically relevant indicators of farm environment health and biodiversity that are simultaneously able to guide farm management and resonate with farmers and consumers alike. Chapter 1 describes a conceptual model that proposes the establishment of a ‘Market Flagship Species,’ an innovative surrogate based on ecological and social science research that signals biodiversity-friendly farming practice. Using birds as proxies for wider agroecosystem diversity as well as farmer attitudes and actions toward biodiversity on their farms, Chapters 3-6 take a transdisciplinary approach in testing the efficacy of the model to select appropriate Market Flagship Species. A principal aim of the ecological component of this study was to examine the role of woody vegetation or ‘refuge’ habitat patches embedded within the production land matrix of sheep/beef farms in supporting bird diversity. Accordingly, bird community composition was measured by conducting five-minute point counts exclusively within woody vegetation patches on 12 South Island sheep/beef farms ranging in geographic distribution from Banks Peninsula (Canterbury) to the Catlins (Southland). 9,050 birds were recorded during 1,086 point counts conducted within 246 different patches over the course of 8 field circuits conducted between 2007-2010. Data were collected according to distance sampling methods, but the extrapolation of diversity estimates within individual patches presented a unique methodological problem: Namely, what is the best method to mitigate bias associated with data collected within a series of discrete patches of varying sizes? Chapter 2 describes a methodology for generating unbiased indices of bird species richness based on the exclusion of data beyond a radial truncation distance of 10m. With an assumed detection probability of 1 within a 10m radius, densities of 12 common species occurring within two key habitats were generated. 10m truncation densities closely corresponded to densities generated using distance methods. Although the developed method reduces available data for analysis, it largely mitigates bias associated with detection while ensuring that sampling effort within small versus large patches remains equivalent. This method was therefore used to establish a dataset upon which analyses of Chapters 3 and 4 were derived. Based on the truncated 10m dataset, Chapter 3 uses general linear models (GLMs) to identify associations between species richness of native and introduced bird species and specific woody vegetation habitat paramaters, such as predominant vegetation species, patch area and patch shape complexity. Models indicated native species richness was strongly associated with native tree habitat and patch area (p<0.001). By contrast, introduced species richness was seemingly emblematic of the generalist ecology that characterizes many introduced species and not significantly associated with any measured habitat parameters. Chapter 4 considers emergent ecological themes from Chapter 3 with regard to individual species. Specifically, the analyses of Chapter 3 are complemented by measuring an index of bird abundance, rather than species richness, for 12 candidate indicator species representing six native and six introdcued species. Candidate species were chosen based on their potential utility to serve as ecological indicators and accommodate social elements of the Market Flagship Species model. Non-parametric analyses indicated strong area and habitat effects for the Bellbird Anthornis melanura, Fantial Rhipdura fuliginosa, Grey Warbler Gerygone igata and Rifleman Acanthisitta chloris(p<0.002), intimating a native species preference for large, native tree habitat patches. The introduced Blackbird Turdus merula, Chaffinch Fringilla fringilla, Goldfinch Carduelis carduelia and Yellowhammer Emberiza citrinell similarly showed significant area and habitat effects (p