Modelling the distributions of Australian shrublands and shrub species: the role of climate and soil properties

Shrubs, short multi-stemmed species, are an important plant growth form that can play a key role in biogeochemical cycles, stability of soil and prevention of soil and water erosion, provision of forage for livestock, and are a source of wood and non-woody products and foodstuff for many populations. However, there is a lack of knowledge regarding the influence of environmental variables on shrub distributions, and shrubland remains undefined as a plant functional type in most global vegetation models. Broadly, the goal of my thesis was to assess relationships between the distribution of Australian shrubland and shrub species and environmental properties, specifically climate and soil characteristics. The objectives of my thesis were three-fold. I initially assessed how the climate envelope of shrubland differed from other major vegetation types (forest, woodland, grassland) as well as differences between the six major shrubland classes (Acacia, Chenopod, Heathland, Mallee, Tall shrublands, “other” shrublands). Using generalized linear models I found that shrublands separate from other major vegetation types along a seasonal soil moisture gradient, with shrublands being the dominant vegetation type in areas with lower moisture. I then used MaxEnt, a species distribution model, to assess drivers of thedistributions of 29 shrub species that together represented dominant members of each of the six shrubland classes. In particular, I sought to determine whether the inclusion of soil characteristics, along with climate variables, improved models of species distributions. I found that whilst models calibrated with soil and climate were not significantly more powerful than those calibrated with only climate variables, for some species projections of the distribution of suitable habitat differed substantially across these models. This led to regional differences in projected species richness, highlighting the value of exploring a broader range of predictor variables when developing models, rather than relying solely on climate. Finally, I examined spatial changes to the distribution of suitable habitat for the 29 shrub species that may occur due to climate change. Given uncertainty in the direction of future precipitation changes, I compared distribution patterns that may result under a hot, dry future versus a warm, wet future. In general, the size of suitable habitat was projected to decline for most species, with greater contractions in central and western regions of Australia and some extensions in temperate regions. Importantly, for some species the direction and magnitude of projected changes varied between models calibrated with only climate variables versus those calibrated with climate and soil variables. The net impact of this meant that different regional patterns in species richness may be projected as a result of model calibration and future climate scenario. This thesis has identified the climate envelopes of shrublands and dominant Australian shrub species; highlighted the importance of considering soil properties when modelling plant species distributions; demonstrated potential impacts of climate change and how patterns of species richness may vary depending on whether the future is warm and wet or hot and dry; and revealed uncertainty in projections of future suitable habitat that may occur due to selection of predictor variables.