David Dunkerley

Runoff and runon areas in a patterned chenopod shrubland, arid western New South Wales, Australia: characteristics and origin

Strongly developed vegetation banding in desert chenopod shrubland occurs on hillslopes having gradients of as little as 0·5 degrees and displays a stepped microrelief of about 10 cm. Surface runoff is shed from the bare surfaces in rainstorms of as little as 4–5 mm, and infiltrates readily within the vegetated bands. The banding thus functions as an efficient system for water redistribution, the landscape being divided into multiple bare runoff (water source) and vegetated runon (water sink) zones. Patterns of stone distribution across a study hillslope suggest that the vegetation banding is at least Holocene in age. The patterned shrublands thus represent an enduring component of this arid rangeland environment, and one whose unusual microhydrology should be preserved by informed management.

Measuring interception loss and canopy storage in dryland vegetation: a brief review and evaluation of available research strategies

The interception storage capacity has been measured for a range of dryland plants. Interception losses over time, however, arise in rain events that deliver either less or more than the canopy capacity. The fate of water in these cases depends on the efficiency with which the intercepted water is returned to the atmosphere by evaporation from the plant canopies. Two primary methods to estimate interception losses are (i) calibrated process-based models of interception and evaporative loss and (ii) direct measurement. Models have been applied only rarely to dryland plant communities, and direct measurement techniques are in need of additional testing and refinement. Most published estimates of interception loss in dryland plant communities therefore appear to be based upon inadequate data and methods. Research needs in this area are highlighted. Copyright

Flow behaviour, suspended sediment transport and transmission losses in a small (sub-bank-full) flow event in an Australian desert stream

The behaviour of a discrete sub-bank-full flow event in a small desert stream in western NSW, Australia, is analysed from direct observation and sediment sampling during the flow event and from later channel surveys. The flow event, the result of an isolated afternoon thunderstorm, had a peak discharge of 9 m 3/s at an upstream station. Transmission loss totally consumed the flow over the following 7·6 km. Suspended sediment concentration was highest at the flow front (not the discharge peak) and declined linearly with the log of time since passage of the flow front, regardless of discharge variation. The transmission loss responsible for the waning and eventual cessation of flow occured at a mean rate of 13.2% per km. This is quite rapid, and is more than twice the corresponding figure for bank-full flows estimated by Dunkerley (1992) on the same stream system. It is proposed that transmission losses in ephemeral streams of the kind studied may be minimized in flows near bank-full stage, and be higher in both sub-bank-full and overbank flows. Factors contributing to enhanced flow loss in the sub-bank-full flow studied included abstractions of flow to pools, scour holes and other low points along the channel, and overflow abstractions into channel filaments that did not rejoin the main flow. On the other hand, losses were curtailed by the shallow depth of banks wetted and by extensive mud drapes that were set down over sand bars and other porous channel materials during the flow. Thus, in contrast with the relatively regular pattern of transmission loss inferred from large floods, losses from low flows exhibit marked spatial variability and depend to a considerable extent on streamwise variations in channel geometry, in addition to the depth and porosity of channel perimeter sediments. | The behaviour of a discrete sub-bank-full flow event in a small desert stream in western NSW, Australia, is analysed from direct observation and sediment sampling during the flow event and from later channel surveys. The flow event, the result of an isolated afternoon thunderstorm, had a peak discharge of 9 m 3/s at an upstream station. Transmission loss totally consumed the flow over the following 7.6 km. Suspended sediment concentration was highest at the flow front (not the discharge peak) and declined linearly with the log of time since passage of the flow front, regardless of discharge variation. The transmission loss responsible for the waning and eventual cessation of flow occurred at a mean rate of 13.2% per km. This is quite rapid, and is more than twice the corresponding figure for bank-full flows estimated by Dunkerley (1992) on the same stream system. It is proposed that transmission losses in ephemeral streams of the kind studied may be minimized in flows near bank-full stage, and be higher in both sub-bank-full and overbank flows. Factors contributing to enhanced flow loss in the sub-bank-full flow studied included abstractions of flow to pools, scour holes and other low points along the channel, and overflow abstractions into channel filaments that did not rejoin the main flow. On the other hand, losses were curtailed by the shallow depth of banks wetted and by extensive mud drapes that were set down over sand bars and other porous channel materials during the flow. Thus, in contrast with the relatively regular pattern of transmission loss inferred from large floods, losses from low flows exhibit marked spatial variability and depend to a considerable extent on streamwise variations in channel geometry, in addition to the depth and porosity of channel perimeter sediments.

Banded vegetation near Broken Hill, Australia: significance of surface roughness and soil physical properties

Abstract Selected physical properties of the soils developed within a strongly-banded grassland in arid New South Wales were investigated to reveal their possible significance for the hydrologic and erosional behaviour of the mosaic landscape. Detailed surface microtopography, surface roughness, soil bulk density and the unconfined compressive strength of the soils were determined using linear transects across the banded mosaic. The landscape is shown to consist of a tier of concave-upward microtopographic elements. Results indicate that the cross-pattern (downslope) variation in the soil parameters is systematically related to position within a particular component (grove, intergrove, etc.) of the mosaic. Compressive strength and bulk density increase downslope across intergroves, peaking at very high levels within the zone of forbs, while groves display lower but more uniform values. Surface roughness increases downslope through the intergrove and the zone of forbs at the upslope margin of a grove, reaching its maximum within the grove. Mosaic components thus cannot be treated as uniform in their soil properties, and single samples from within a component are shown in general to be inadequate. The mapped pattern of soil properties implies a very stable configuration for banded mosaics. Surface runoff is increasingly hindered during flow from the intergrove onto the grove. At the same time, soil resistance to entrainment increases in opposition to the shear forces generated by the runoff. In concert, these tendencies imply that little sediment transport is possible across the mosaic. The resulting landscape stability appears to confer robustness to the mosaic in the face of stresses such as drought and pastoralism, when plant cover may be temporarily thinned or absent. After drought, for example, water ponding would again begin at the downslope margin of the concave topographic elements, fostering re-establishment of the groves.

The influence of organic litter on the erosive effects of raindrops and of gravity drops released from desert shrubs

Abstract This paper reports on a laboratory study of the effects of water drop impacts on litter and sand splash beneath desert shrubs. Individual drops of 5.7 mm diameter were released from heights of 0.5, 1.0, and 1.5 m (selected to encompass the height range of typical desert shrubs) onto targets of bare or partially litter-covered, saturated fine sand. The natural litter, largely derived from the saltbush Atriplex vesicaria , was collected from desert shrubland sites in western New South Wales (NSW). The drop impacts caused both sand and litter particles to undergo splash displacement. The mass of sand splashed was found to increase with drop fall height, while mass of litter particles splashed did not vary significantly with fall height. Weights of sand moved by airsplash were significantly diminished by surface litter applied at the rate of 200 g/m 2 . These findings indicate that gravity drops released from desert shrubs may provide both an erosive force beneath these plants, and a means for dispersing litter from the plant base into the surrounding landscape, where litter may continue to affect hydrologic and erosional processes. By restricting splash of mineral particles, litter acts to limit soil splash from beneath shrubs, and in this way may contribute to the persistence of plant mound microtopography that is common in desert shrublands. Under open-field conditions, large raindrops delivered in convective showers must cause similar airsplash transport of litter particles, thus playing a role in the distribution of litter within the landscape.

Plant canopy interception of rainfall and its significance in a banded landscape, arid western New South Wales, Australia

The canopy interception storage capacity of three plants in an arid, banded (runoff-runon) landscape was determined in the field from weight gain under artificial rain. The plants, two shrubs and a grass, were the bladder saltbush Atriplex vesicaria (Heward ex Benth.), the pearl bluebush Maireana sedifolia (F. Muell.) P. G. Wilson, and perennial Mitchell grass Astrebla lappacea (Lindl.) Domin. For saltbush and Mitchell grass, canopy water storage capacity (1.3 mm) correlated strongly with projected canopy area, while for the dense canopy of the bluebush (2.0 mm capacity), plant weight was a better predictor. In view of canopy cover amounts in each plant community, and the mean rain day rainfall of 5.7 mm, estimated annual water losses amount to 32% for Mitchell grass communities, 5% for saltbush communities, and <1% for bluebush communities. Canopy interception loss is shown to be a significant influence on the water redistribution that supports banded landscapes in the arid Broken Hill area.

Frictional retardation of laminar flow by plant litter and surface stones on dryland surfaces: A laboratory study

The relative contributions of plant litter and surface stones to the frictional retardation of laminar flows were investigated in a series of laboratory experiments on a glued sand board. A computer-controlled measuring gantry was used for precise recording of flow depths, and retardation was quantified using the Darcy-Weisbach friction factor f. Surface stones had relatively little effect on flow retardation or flow speeds but increased flow depths in comparison with the same discharge across a stone-free surface. In contrast, plant litter produced larger reductions in flow speed and increases in flow depth, associated with large increases in flow retardation. These effects were most marked at the highest loading tested, 80 g m -2. Results show that the retardation caused by a given surface cover fraction of litter significantly exceeds that caused by the same cover of surface stones, emphasising the need for measures other than surface cover to be employed in work on overland flow. Neither nonsubmerged obstacles such as surface stones nor plant litter contribute increased roughness with rising discharge in the laminar regime. Instead, values of f decline as Reynolds number rises. These findings contrast notably with relationships established for turbulent flow and suggest that for correct modeling of overland flow in drylands, the temporal and spatial ranges of laminar flows must be distinguished from those where turbulent flows occur.

Banded chenopod shrublands of arid Australia : modelling responses to interannual rainfall variability with cellular automata

Abstract Cellular modelling is used to investigate the behaviour of well-developed banded chenopod shrublands in western N.S.W. Australia, where annual rainfall exhibits high year-to-year variability. Rainfall records for the past 100 years show that the study region has a mean annual rainfall of about 215 mm/a and coefficient of variability of annual totals of 40–50%. Synthetic sequences of yearly rainfalls matching these characteristics are used in modelling. It is found that interannual rainfall variability like that of the modern environment does not inhibit the development of vegetation bands. However, band spacing is increased, while band regularity, and the long-term mean plant cover, are significantly diminished. Rainfall fluctuating about a mean of 215 mm/a results in an 8% diminution in plant cover, compared to an unvarying rainfall of the same average amount. Vegetation bands widen by downslope expansion in wet intervals and isolated band fragments may also link across the slope. In dry intervals, the lower margin retreats upslope and bands may re-fragment. No net upslope band migration is seen in the model during wet episodes, and this is consistent with historical aerial photograph evidence of band fixity in western N.S.W.

The influence of hillslope gradient, regolith texture, stone size and stone position on the presence of a vesicular layer and related aspects of hillslope hydrologic processes: a case study from the Australian arid zone

Abstract A near-surface layer of vesicles may be found in many dryland soils, often below a stony desert pavement. During and after rain the uppermost part of the regolith develops a surface seal, trapping air which is compressed and which may also subsequently expand as a result of solar heating. This results in the deformation of the soft regolith and the formation of a vesicular layer just below the surface. The work reported here was undertaken to determine the influence of slope gradient, regolith texture, stone size, and depth of stone embedment on the presence of a vesicular layer in arid Australia. This appears to be the first investigation of the role of topography in the formation of soil vesicularity. The four factors listed typically vary systematically with position on a hillslope. The data collected indicate that the development of a vesicular layer is inversely related to hillslope gradient and size of the stone, positively related to the amount of silt and clay in the regolith, and weakly related to the depth to which the stone is embedded. Field evidence indicates that the vesicular layer becomes much more extensive on the lower parts of hillslopes in the study area, and thus so too must the choke on water penetration that such layers create. The spatial distribution of vesicular layers thus constitutes a factor that has traditionally been overlooked in studies of dryland hillslope hydrology.

Ecogeomorphology in the Australian drylands and the role of biota in mediating the effects of climate change on landscape processes and evolution

Abstract Australian dryland landscape has developed under the influence of aridity, low relief, tectonic stability and biota adapted to nutrient and water scarcity. The biota in general, but notably the plants, mediates the impact of water scarcity and of climate change on ecohydrological and geomorphological processes. It reduce the partitioning of rain into overland flow, and so limit soil erosion, notably through the development of patch structures that partition the landscape into local runoff sources and runon sinks. In large rain events, when flow does reach ephemeral streams, channel-associated plants again modify flow conditions, reducing flow speeds and flow competence. Given the diverse influences of biota on landscape processes, it is argued that it likewise moderated the effects of Quaternary and Holocene climate change. Field evidence from Australian and other drylands suggests that the effect of changing land surface properties on runoff and erosion may exceed the effect of moderate climate change. Knowledge of the role of dryland biota and its role in land surface change is therefore a prerequisite to understanding the responses of landscapes to climate change, to understanding the complex spatio-temporal variability in landscape development, and to developing the ability to correctly interpret the alluvial record of changing geomorphological processes in terms of changes in climate and other external drivers.