[Truncated abstract] Most of Australia, and large areas of many other continents, is drained by intermittent rivers and streams, however comparatively few biogeochemical studies have been completed for these systems. Intermittent, dryland streams are highly dynamic environments subject to unpredictable and sporadic flow. Natural disturbance from lengthy drought periods and sudden floods are typical for these systems. Without adequate baselines for natural disturbances, it is difficult to quantify other effects from anthropogenic disturbance such as dewatering, land clearing, and urbanisation, or climate change. This thesis presents work from a four-year study examining the biogeochemistry of nitrogen (N), phosphorus (P) and carbon (C) in soils and sediments of two intermittent streams (Barnett Creek and Pirraburdoo Creek) in the Pilbara region of north-west Australia. The Pilbara is an area of ancient geology and highly weathered environments that is undergoing rapid development yet is poorly understood from an ecological perspective. The principal objectives of this thesis were to determine: i) how flooding affects the spatiotemporal patterns of nutrients in intermittent stream landscapes; ii) the role of flooding in N and C mineralisation and microbial dynamics; and iii) the connections between benthic algae, microbes and nutrient availability in channel sediments. To address these objectives, three field studies and two incubation experiments were conducted. Field studies at Barnett Creek indicated that flooding reduced the spatial heterogeneity of available soil nutrients and microbes in the stream landscape, and that topography (relative elevation) in the stream landscape was of less importance in influencing nutrient and microbial patterns than flooding or landscape position. ... Field studies at Pirraburdoo Creek indicated that microbial biomass and activity increased in benthic algal mats during mat senescent stages, and decreased after flooding when mat biomass peaked. Benthic algae grew rapidly in gravel run environments after flooding, while declining in pools, and demonstrated moderate N limitation and strong P limitation. Pools had two to eight times greater NO3-N, three to five times more total N, and two to three times more labile P, OC and total C than either pools after flooding, or runs before or after flooding. Hence, the pools at Pirraburdoo Creek represented a local, interflood store of nutrients in otherwise nutrient-poor landscape, when connectivity to upstream reaches or upland environments was weak or non-existent. This thesis provides the first detailed analysis of soil and sediment biogeochemical responses to flooding for intermittent streams in the Pilbara region and for semi-arid Australia. Further pressing questions raised by this work include: What is the key pulse size and frequency for maintaining Pilbara riparian communities as well as soil microbial function? How do the spatio-temporal nutrient and microbial patterns observed persist over (i) multi-decadal scales, (ii) mega-spatial (larger landscape to regional) scales, (iii) different flood frequency-magnitude regimes, and (iv) different stream sizes? Stream biogeochemistry is a burgeoning field, and it is therefore reasonable to expect such existing gaps in knowledge may be addressed in the near future.

[Truncated abstract] Understanding the ecohydrological dynamics of native vegetation can provide a benchmark for future efforts to restore landscape hydrology and allow predictions of potential landscape responses to climate uncertainty and associated changes in vegetation cover. The key drivers of evapotranspiration (Et) involved in maintaining a hydrological balance that minimises deep drainage in semi-arid ecosystems operate at a range of scales, and in this thesis I assessed the water relations of functionally and taxonomically diverse plant communities in south-western Australia from the leaf-level to ecosystem scale. For three key communities; heath shrubland, mallee (small multistemmed eucalypt) -heath, and open eucalypt woodland, populating a typical catenary sequence of soil types along a slope, I addressed the following questions: 1) What are the predominant water use strategies of wheatbelt native plant communities and what underlying trade-offs determine the distribution of plant water use strategies along the topographical gradient? 2) What are the roles of soil water and hydraulic limitation in controlling the spatial and temporal dynamics of transpiration in different functional types? 3) What is the magnitude and partitioning of total Et in the woodland community and what processes determine Et fluxes on a seasonal and annual basis? 4) What are the seasonal differences in Et among contrasting community-types and how do these patterns relate to canopy attributes and transpiration capacity along the topographical gradient? A key philosophical step in working with species-rich communities was to develop the concept of 'hydraulic functional types' (HFTs) to identify groupings of species using associations of physiological and morphological traits that define their hydrological functioning. .... However, as shallow soils dried during spring and summer, Et fluxes were significantly lower at the heath site (0.35 versus 0.66 mm day-1 for the woodland in February), demonstrating that the seasonality of Et fluxes differentiates communityscale contributions to regional water balance. Land-surface exchange of water over native vegetation is by no means uniform, but varies according to the spatial and temporal availability of water along topographical gradients. In general, shallow soils present fewer opportunities for water use partitioning and favour drought hardiness and a transpiration response that tracks recent rainfall patterns, whereas deeper soils promote greater differentiation in water use strategy and support canopies responsive to atmospheric demand. This thesis provides a unique description of ecosystem water balance in a global biodiversity hotspot by viewing complex vegetation mosaics in terms of their relevant hydrological units. This information is fundamental to sustainable agroforestry and revegetation efforts and our ability to gauge possible changes in vegetation structure and function under a changing climate.