Biodiversity, ecosystem service provision and human well-being are inextricably linked. The current rate of biodiversity loss worldwide is impacting on ecosystem service provision with negative implications for human well-being. Little quantitative information is available about the provision of most ecosystem services by most ecosystems, the effect of management on the ability of vegetation to provide services, or trade-offs in service provision with land use. This information is particularly important in agricultural landscapes where the extent of landscape change is affecting biodiversity and ecosystem service provision substantially and thus agricultural sustainability.

This study quantified the provision of carbon storage, erosion mitigation and biodiversity conservation services by five vegetation communities (river red gum Eucalyptus camaldulensis riparian forests, coolibah E. coolabah woodlands and open- woodlands, myall Acacia pendula tall shrublands and tall open-shrublands, black box E. largiflorens woodland and open-woodland, and mixed grassland – low open-chenopod shrubland) common on the lower Namoi floodplain in northern New South Wales, Australia. Sites represented the full range of structural and compositional variants encountered within each vegetation type over the 7100 km2 study region, from heavily grazed derived grasslands to old-growth woodland or forest evidently little affected by anthropogenic disturbance.

The environmental conditions dictating the location of each vegetation type in the landscape were investigated. The distribution of vegetation types depended predominantly on soil type, flood patterns and the interaction between the two. Woody and non-woody vegetation was mapped across the study region using unsupervised classification of ten single-date SPOT 5 scenes with 85% accuracy. Woody vegetation covered approximately 7% of the lower Namoi floodplain.

Carbon storage was measured or estimated for soils, woody vegetation, dead standing vegetation, coarse woody debris, herbaceous vegetation, litter and roots. River red gum sites were the most valuable vegetation type for carbon storage, having up to 4.5% carbon content in the surface 0–5 cm soil depth increment, with total site carbon storage averaging 216 t C ha–1. The most carbon-dense site was east of Narrabri and dominated by river red gum. Grasslands were the least carbon-dense with 40.0 t C ha–1. The greatest proportion of carbon in river red gum sites was in woody biomass, but in all other vegetation types and especially grasslands, the top 0–30 cm of the soil was the most C-rich component of the ecosystem. Woody biomass C was positively correlated with C derived from dead standing wood, coarse woody debris and litter, but not herbaceous biomass C, which was negatively correlated.

Herbaceous vegetation cover, litter cover and macroaggregate stability as determined by the topsoil C:N ratio was used to rank sites for erosion mitigation service provision. Erosion mitigation value was assessed in terms of aggregate stability, which was determined by a relationship between mean weight diameter of aggregates and soil C:N ratio, as well as dominant cations on the clays. Soils with higher C:N contained more stable macroaggregates, and tended to be dominated by river red gum. High aggregate stability in river red gum sites was attributed to large inputs of eucalypt litter and coarse

ii

woody debris. Highest microaggregate stability was also observed in river red gum sites and attributed to the dominance of Ca2+ rather than Na+ on clay exchange sites.

Vascular plant and bird conservation value of sites was determined by ranking sites according to the number of rare (i.e. infrequently observed) species present. For birds, species richness was also taken into account. River red gum sites were ranked highest for vascular plant and bird conservation value because they contained the highest abundance of rare species of both vascular plants and birds. However, river red gum sites also contained the greatest number of introduced plant species presumably as a result of flood mediated dispersal of propagules. All vegetation communities were included among the sites of highest conservation value for both vascular plants and birds. However, in the top 30% (16 of 54) of sites ranked according to conservation value, only five sites were valuable for both plant and bird conservation. River red gum sites had the most structurally complex vegetation, which coupled with their proximity to water, encouraged high bird species richness and abundance. Woody plants were the most influential vegetation component determining bird conservation value, but different vegetation types were preferentially used by different bird species, implying that the full spectrum of vegetation types is required to maximise bird and plant conservation at the regional scale.

Increasing grazing intensity severely diminished both plant and bird conservation value at river red gum and coolibah sites as a result of the loss of rare species. Grazing also detracted from carbon storage, both directly through biomass consumption and indirectly through associated management (such as ring-barking to increase herbaceous

iii

biomass production and clearing). The functional richness (i.e. the number of different life-forms of vegetation types) was more influential than species richness in terms of ecosystem service provision. Shannon–Wiener diversity of vegetation communities showed no relationship with ecosystem service provision. No trade-offs were evident between the three ecosystem services measured in this study, but conservation value and carbon services declined under increasingly intense grazing. Increasing woody vegetation biomass and cover resulted in decreased herbaceous biomass production, leading to the trade-off between nature conservation and carbon sequestration on the one hand, and livestock production on the other. There are few ungrazed sites in the study region, hence natural capital may be diminished still further with continued grazing of almost the entire landscape.