Non-perennial rivers are valuable water resources that support millions of humans globally, as well as unique riparian ecosystems. In Australia, the Earth’s driest inhabited continent, over 70% of rivers are non-perennial due to a combination of ancient landscape, dry climates, highly variable rainfall regimes, and human interventions that have altered riverine environments. Here, we review Australian non-perennial river research incorporating geomorphology, hydrology, biogeochemistry, ecology, and Indigenous knowledges. The dominant research themes in Australia were drought, floods, salinity, dryland ecology, and water management. Future research will likely follow these themes but must address emerging threats to river systems due to climate change and other anthropogenic impacts. Four high level opportunities for future research are identified, namely: (1) integrating Indigenous and western scientific knowledge; (2) quantifying climate change impacts on hydrological and biological function; (3) clarifying the meaning and measurement of “restoration” of non-perennial systems; and (4) understanding the role of groundwater. These challenges will require inter- and multi-disciplinary efforts supported by technological advances. The evolving body of knowledge about Australian rivers provides a foundation for comparison with other dryland areas globally where recognition of the importance of non-perennial rivers is expanding.

Lake Eyre Basin (LEB) rivers are important but under-researched. This study integrates published and new information into a systematic geomorphological overview, subdividing the LEB into nine landscape zones: basin-margin Uplands with relatively high-energy rivers; Rocky Fringe, piedmont with short rivers and floodouts; Western Rivers, notable for neotectonism and stream capture; low-relief Northern Plains drainages and floodouts; Central Plains dunefield; Channel Country-Vertic Downs, cracking-clay interfluves and valley macrochannels; Channel Country-Stony Domes, rocky or sandy interfluves and complex macrochannels; Lakes & Dunes maintained by saline groundwater; and playas in the Mega-lakes. Zone boundary GIS datasets are available. LEB rivers are shown to be diverse and distinctive, indicating a need for locally relevant knowledge in management and monitoring. Fluvial landform suites include sand-bed rivers, macrochannels hosting coexisting anabranching and anastomosing systems in vertic floodplains, floodouts, low-angle alluvial fans, and three types of waterhole. Channels may be discontinuous or not present; floodplains can dominate flow transmission. Contributing hillslopes are heterogenous in their capacity to generate runoff, while drainage networks may have low to high degrees of integration. Management implications include the importance of in-river terrestrial ecosystems and “patchy” drainage networks, and the inapplicability of some morphometric practices.

Two spectacular cliff lines occur along Australia's south and west coasts: the Great Southern Scarp (new name) and the Zuytdorp Cliffs. Detailed measurements of their length show that they are exceptionally long and unusually continuous. The Great Southern Scarp is the southern edge of the Nullarbor Plain's Bunda Plateau; it is cut into shallow marine sediments and extends for 820 km. Once a continuous sea-cliff, local uplift has isolated two sections from the ocean, so it now comprises the Bunda Cliffs (210 km of coastal cliff, 180 km of which is uninterrupted cliff line), Hampton Range (inland; 300 km), Baxter Cliffs (160 km of coastal cliff) and Wylie Scarp (inland; 160 km). The Zuytdorp Cliffs are coastal cliffs cut into eolianite, and extend for 210 km with an uninterrupted section of 120 km. The length of the Great Southern Scarp and the Zuytdorp Cliffs results from an unusual combination of circumstances. They are both composed of poorly jointed, relatively homogenous biogenic calcarenites, presented to high-energy ocean waves by regional uplift. The carbonates are sufficiently well cemented to maintain a steep vertical cliff face, but susceptible to disintegration under direct wave attack. Cliff retreat has been fairly uniform because of the broad spatial scale of both lithology and erosion. The arid climate, absence of nearby non-karstic catchments and karstic nature of the cliffs' hinterlands has discouraged integrated drainage development, so no significant fluvial systems dissect the cliffs; this is a key factor in cliff edge preservation. In the case of the Great Southern Scarp, these processes have formed the longest continuous cliff line in Australia and probably the world. Referenced against criteria from Australia's National Heritage List, the cliffs have potentially international/national levels of significance for the rarity of their scale, their demonstration of landscape evolution, and their spectacular beauty.

This paper presents a summary of an internal report on the potential National Heritage values of Australian geomorphology within a defined area. The desktop study sought information on landforms which best demonstrate the evolution of the characteristic landscapes now found in Australia's drylands. Australian dryland landscapes owe much to continental stability, which allows inheritance of landscape features and preserves the effects of previous climates. Aridity is only the latest and not always the most powerful of a long series of influences. Twenty-eight places with high potential to pass heritage criteria were identified across eight geomorphic themes (astroblemes, sand deserts, karst, arid coasts, tectonic landforms, uplands, regolith, watercourses). Some were significant within a single theme (the Acraman Impact Structure, the Nullarbor Plain); others had value across many themes/sub-themes (two Cooper Creek sites, the Neales Catchment, Rodinga Range). Many had sufficient information to identify particular locations (e.g. Gosses Bluff, the Zuytdorp Cliffs), but those without were assigned broad indicative areas, from which suitable locations can be found (e.g. the Eastern Goldfields Palaeodrainages, the Simpson Desert, the Black Soil Plains). Knowledge gaps are identified, especially the remote Great Sandy Desert and Great Victoria Desert dunefields and the Davenport–Murchison Ranges. Clusters of potential heritage occur along the Amadeus Basin ranges, and within the Lake Eyre Basin.

Understanding the processes governing sediment distribution is important for predicting reservoir, seal, and migration pathway properties. Cross-discipline collaborations across geomorphology and sedimentology can develop process-based facies models, which are adaptable to project-specific contexts (tectonic, climatic, provenance). Examples are given from Australia's Lake Eyre Basin, a presently underutilised modern analogue for fluvial processes in a dryland setting. The Lake Eyre Basin (LEB) catchment is arid to semi-arid, with extremely variable flow regimes. Drainage is centripetal towards Lake Eyre. The LEB is a continent-scale Australian intracratonic sag that overlies the Eromanga, Pedirka, Cooper, and Galilee Basins. Sediments have been accumulating since the Palaeocene; in the LEB's centre and inner north-west, where synsedimentary domal uplift and trough subsidence occurs, the Cainozoic attains thicknesses of 200->300 m. At the field and reservoir scale, the LEB hosts a variety of dissimilar river types, including anabranching and single-thread channels of low to high sinuosity, floodplains of sand or of mud, and bedloads of mud aggregates, sand or gravel. Within the rivers, there is a variety of landform suites related to flows moving from constricted to unconstricted settings. As geomorphic entities, they include 1) valleys with channel-floodplain suites, 2) floodouts (transition from channelized to unchannelised flow down-valley) and 3) distributary channel systems. As sedimentary packages, some will form valleyor basin-fill, and some will be low-angle alluvial fans. Few have been investigated from a sedimentologic viewpoint; others are better known from the geomorphic literature; many are undescribed. Their individual characteristics relate to differences in fluvial processes, sediment provenance, local base level, and post-pluvial history. Differences are expressed in different channel planform and bedload (sand body distribution and connectivity) and floodplain composition and creation (seal and migration barrier qualities). We will present examples from a range of these, in this initial collaboration between a basin-scale process geomorphologist and a reservoir-scale focused process sedimentologist, in the anticipation that this will enable the depositional systems within this large dryland basin to be explored in a new and useful light.

The invasion of northern Australia by the poisonous cane toad is well recognised, as is its devastating impacts on numerous local native species. However, there is little recognition that the toads are spreading into south-western Queensland. Utilising local knowledge, a limited survey was undertaken within the Cooper Creek catchment to locate the invasion front. Dispersal during 2010–11 floods has established cane toads as far south as Jundah. Integrating this information with landform mapping indicates that cane toad invasion can continue south-west down the Cooper Creek. Though arid, Cooper Creek’s geomorphology renders it partially independent of local climate, and permanent and semipermanent waterholes (including RAMSAR-listed wetlands) are found downstream from Windorah and into the Strzelecki Desert. Natural landforms provide potential daytime shelter and breeding sites, and additional suitable habitat created by human activity is also widespread. Even unsuccessful attempts at breeding may be detrimental to regional ecology, especially fish populations, at critical stages of their boom/bust cycle. We conclude that there is no reason why cane toads cannot penetrate further down the Cooper Creek, threatening wetlands in north-eastern South Australia. Published models of cane toad expansion, which conclude that north-eastern South Australia is too dry for cane toad populations to establish, are based on climatic parameters that significantly under-represent true habitat availability.

Geological heritage is insufficiently recognised in Australia; it should be considered in its own right, not as an addendum to other heritage values. The lack of a suitable robust and repeatable methodology has seriously constrained the assessment of geological sites suitable for the National Heritage List (NHL). A desktop assessment of Australian desert landscapes required intrinsic natural values of a diverse group of sites, spread over a vast area, to be assessed against NHL criteria. The Earth Sciences Comparative Matrix (ESCoM) was developed for this study. In the ESCoM, sites are grouped in process themes. Each is assessed against NHL criteria then compared with other similar places, according to degree of unusualness, integrity, and authenticity. A site scoring well across multiple themes has increased heritage significance. The overall values of a site are quantified, leading to a qualitative judgement on whether it achieves the threshold of outstanding heritage value. Examples of assessment using this method are given. In this methodology, significance determination is based on rigorous comparisons of specific values. It is semi-quantitative, repeatable, and robust. It differs from other geoheritage assessment methods in its combination of process-based groupings (facilitating the separation of site type from heritage criteria), matrix structure (minimising complexities of scale or diversity), and use of numerical rankings as an aid in decision-making. While the study for which ESCoM was developed was focused on landforms, it can be used for other types of geoheritage (e.g. fossils, tectonic processes), with modification of matrix theme headings.

This study examines landscape rehabilitation treatments installed 20–40 years ago in the Western Catchment of NSW. Treatment outcomes were assessed using geomorphic criteria, because geomorphic processes are fundamental to ecological permanence. Contour furrowing creates artificial runoff-runon sets which intercept runoff (resistance to flow by windrows microrelief and surface roughness) and promote infiltration (artificial permeability by ripping). As originally conceived, after windrows subside, flow resistance would be afforded by surface roughness under belts of vegetation. This study shows that rehabilitation treatments have a more complex relationship with the landscape than this would suggest, and that the final effect of the treatment depends on the geomorphic processes natural to the site. Treatment design should therefore be site-specific. The relevant aspects of treatment design are site location, runoff : runon ratio (expressed as furrow spacing and furrow length), furrow placement, and post-treatment management. Some long-term successes are documented. In ironstone ridge country affected by impermeable hard-setting soils, furrowing creates artificial permeability, allowing plant germination; plant material in the soil reverses hard-setting and establishes self-sustaining permeability. In stony gilgai country furrowing through vegetated patches can aid in re-establishing vegetation, but furrowing through stony runoff patches only diminishes, rather than improves, landscape function. Other landscape types will have different key attributes. In all cases, selection of appropriate sites for rehabilitation treatment is of primary importance. The 1990s NSW Soil Conservation Service best-practice included a specialised furrower, surveying techniques for accurate furrow placement along the contour, staggered gaps along each furrow line to reduce risks of gullying by windrow breakthrough, and post-treatment management of total grazing pressure. New guidelines for treatment design developed from this study include determining for each site the optimum runoff:runon ratio (which varies according to climate, gradient, vegetation, and regolith), and matching furrow spacing and furrow/gap length to local runoff:runon ratios. In stony gilgai country, furrow placement should be along the contour but within non-stony patches; elsewhere, placement should be rigorously along the contour. In ironstone ridge country, a greater runoff:runon ratio, commensurate with the area’s apparently larger patch scale, can be achieved by having more gap than furrow along each furrow line. No single rehabilitation technique will fit all landscape types, and these guidelines will ideally be developed further with investigation of other landscapes.

Fowlers Creek is a mud-aggregate fluvial system. Floodplain muds dominate the river's deposits and consist of silt, fine to very fine quartzose sand, and clay. Up to similar to 80% of the silts and clays are bound into sand- and silt-sized aggregates and comprise a substantial component (> 42%) of the floodplain muds. Mud-aggregate sediments behave like sands during transport, and as a result, muds can be deposited under conditions of greater flow velocity than would otherwise be the case. Newly deposited floodplain muds are loose and easily entrained, but older floodplain muds are cohesive, and the distribution of modem and older floodplain muds influences erosion patterns across Fowlers Creek.In the lower order streams of the Fowlers Creek uplands, alternate reaches of shallow rectangular channels and unchannelled floodplains collectively form discontinuous ephemeral streams. These landform sequences consist of gullies, coalescing downstream to arroyos, which terminate in distributary intermediate floodouts. At Fowlers Creek, floodouts are preferentially located at tributary junctions, reflecting their origin during very large floods. At floodouts, low slope and high vegetation density promote sheetflow infiltration and landform stability. Their efficiency in retaining runoff make floodouts drought refugia; they are an important ecological element in this and area.The higher order channel of the mid-uplands is a mobile, low-sinuosity, single-thread arroyo, incised into wide muddy unstable floodplains. Fluvial processes are dominated by episodic flood-driven channel avulsion, and variability in stream energy and boundary resistance contributes to a non-equilibrium fluvial style. Frequent reach-scale channel relocation is accompanied by the burial of the abandoned channel in floodplain muds and both erosion and aggradation in downstream floodplains. (c) 2006 Elsevier B.V. All rights reserved.

Dryland rivers in which fine sediments travel as aggregates are increasingly recognized in modern and ancient fluvial systems. Fowlers Creek, Australia, is an ephemeral dryland mud-aggregate river whose sediments provide insights into the dynamics of mud-aggregate floodplains, the origin of massive mudrocks from and depositional environments, and the nature of planar bedding. Fowlers Creek's flow conditions were inferred from relationships between landforms and the sediment texture, bedforms, and sedimentary structures remaining after flow ceases.Floodplain muds, consisting of fine sand and sand-size mud aggregates, are distributed over the floodplain in suspension. As flow decelerates they are deposited as bedload. The shallow depth, high sediment load, and low aggregate particle density promote flow conditions ranging from high in the lower flow regime to upper flow regime, producing coexisting ripples, scours, flat beds, and clay layers. With time, visible signs of aggregate structure are lost, leaving a massive cohesive mud; consolidation is not achieved by burial. Aggregates reappear when muds reenter fluvial transport. In unchanneled reaches, sheetflows deposit sediment with a pervasive horizontal fabric. Channel sediments (coarse sands and gravels) are a minor component of Fowlers Creek's deposits. Widespread lower-flow-regime conditions produce planar beds and 2-D dunes, usually deposited without internal stratification or in horizontal laminae. Lower-flow-regime planar bedding is also observed in fine silty sands. Near the close of flow, rapid shallowing may move channel conditions from lower to upper flow regime, or from lower to higher positions in the lower flow regime bedform stability spectrum, leading to unusual bedform associations.In the rock record sediments from a river like Fowlers Creek would be characterized by structureless gravelly sands (channel facies), massive red mudstones (floodplain facies), and sediments with horizontal fabric but poorly expressed bedding (sheetflow facies).

Large faults expressed at the ground surface often have clear effects on drainage networks. Smaller scale tectonic activity, or that which occurs beneath sedimentary cover, can have more subtle effects that arise from changes to down-valley slope. These include an increase meander sinuosity or the development of anabranching. Fowlers Creek arises in the northern Barrier Range (NSW) and its terminal floodout flows across the flat surface of the Bancannia Trough. In the floodout’s proximal and medial reaches the channel is sinuous but generally without active meander development. Active meandering occurs in only in two reaches, and is associated with higher slopes and repeated avulsions, the most recent being during the 20thC. It is suggested that small-scale uplift may have occurred, increasing the down-valley slope and therefore promoting meander development and avulsion. More closely spaced topographic data is necessary to support this hypothesis, however Fowlers Creek is one of nine creeks flowing east from the Barrier Range which show planform irregularities several kilometres from the rangefront. This supports the possibility of a concealed tectonic feature, and suggests that it is large-scale and linear. Major structures (e.g. mines, waste dumps) may be built in sparsely monitored remote locations where earthquake hazard estimation can be difficult. In such places, maps of fluvial geomorphology can be useful indicators of neotectonic activity.