Australian Natural Resources Atlas

Natural Resource Topics

Soils - Erosion

Australia

The material below is an extract from the Australian Agriculture Assessment 2001 report. For ease of cross reference, figure, table and section references pertain to the chapter structure of this report. The Further Information section provides links to the full graphics version of the material below and the Australian Agriculture Assessment 2001 report.

Erosion and sediment transport

SUMMARY

Predicted sheetwash and rill (hillslope) erosion*

* The assessment of the impact of land use practice on erosion rates could not be undertaken due to a lack of spatial information.

Gully and riverbank erosion

River sediment loads and deposition

MANAGING SOIL EROSION: essential for agricultural sustainability

Soil erosion is a natural process-occurring more in landscapes with high rainfall intensity or steep slopes. The shallow stony soils that cover much of the coastal ranges and the steeper semi-arid lands have been naturally eroded. Where the protective vegetation cover is removed or degraded by clearing, tillage or overgrazing, risks of sheetwash erosion are increased and rill and gully erosion occur. Associated degradation of riparian vegetation has also accelerated erosion of creek and river banks. In arid and semi-arid landscapes, reduced vegetation cover also accelerates wind-borne erosion.

Soil erosion can reduce on-site productivity through loss of fertile topsoil, and associated water-holding capacity and nutrients. Intense erosion also leads to soil structural decline and poor plant growth.

Soil erosion also has the potential for downstream impacts on creeks, rivers, reservoirs, lakes, and estuarine and marine environments. Water-borne erosion increases the supply of sediment to rivers. High concentrations of suspended sediments in rivers can:

Increased supply of sand and gravel from gully and riverbank erosion has led to deposition of sand and gravel beds (sand slugs). Sand slugs smother aquatic habitat. They can prevent fish passage, fill pools and other refugia and are unstable substrates for river bed life.

ASSESSING WATER-BORNE EROSION AND SEDIMENT TRANSPORT

Soil erosion on agricultural land was assessed and placed in the context of river basin sediment budgets. Potential downstream impacts were identified and areas for continuing management attention are highlighted. These build upon major progress made since the 1930s through soil conservation activities to prevent erosion. The assessment includes improved prediction of sheetwash and rill erosion, and the first national assessments of gully erosion, streambank erosion, river sediment loads and deposition of sediment. Most significantly, the work is the first to explicitly relate patterns of sub-catchment erosion to downstream loads and export. The framework should prove valuable for future regional target setting and resource planning.

Assessing water-borne erosion

The assessment of water-borne erosion and river sediment transport covered:

Extent of assessment

The whole continent was assessed for sheetwash and rill erosion. Assessment areas for the other components took place at locations of intensive land use and their surrounding catchments-geographically the catchments of the east coast (extending from Cape York to the Eyre Peninsula), Tasmania and the south-west of Western Australia (Figure 5.1). Complete river basins were used to put intensive land use in the context of their hydrological catchments and to predict river sediment loads. Most river basins also include non-intensive land uses (e.g. forestry and rangelands).

Data

The assessment area was divided into 15 regions based upon drainage divisions and finer boundaries for summarising regional patterns (Figure 5.1). Data for modelling sediment loads in the Timor Sea, Gulf of Carpentaria and Western Plateau Drainage Divisions were inadequate.

Data on erosion rates and river sediment loads are limited with no erosion information available for many basins. Any national assessment must interpolate and extrapolate known data into areas with none. Fortunately, we have a good understanding of the processes of sediment transport and adequate data were available on the factors that control these processes, including comprehensive stream gauging records, digital elevation models, rainfall records and remote sensing of land cover.

The assessment used all available information-collecting further data where practical-and these were inputs to a relatively simple conceptual model* of the main drivers of sediment transport for large catchments. Process knowledge was used to constrain interpolation and extrapolation rather than relying on statistical associations between variables or on geographical interpolation. Statistical models were used for aspects where there was no adequate, large-scale, process understanding (e.g. for extent of gully erosion and width of river channels).

Soil loss and sediment movement predictions are outputs from sophisticated data-mining techniques and were implemented after evaluation. The river sediment model was developed to integrate the erosion process in detail across environmentally diverse catchments. It represents a major advance in available techniques for assessing regional river sediment loads and is also of value for more detailed regional assessment of sediment-related issues (e.g. downstream water quality, river health and catchment restoration).

View Figure 5.1. Grouping of river basins with intensive agriculture for regional reporting.

* A conceptual model is one where simplified constructs of the physical processes are used (in this case, constructs for sediment transport at the large catchment scales).

Sheetwash and rill erosion

Controls on sheetwash and rill erosion are well understood and several models that incorporate these factors exist. The Revised Universal Soil Loss Equation (RUSLE) (Wischmeier & Smith 1978, Renard et al. 1997) was used to predict mean annual sheetwash and rill erosion potential across Australia under the current land use. Methods used to assess sheetwash and rill erosion are given in Lu et al. (2001a, 2001b) and Gallant (2001), including details of verification of the method against 100 observations of soil erosion rates.

The RUSLE calculates mean annual soil loss (tonnes/hectare/year) as:

Annual soil loss = R x K x L x S x C x P

Where:

R is rainfall erosivity factor
K is soil erodibility factor
L is hillslope length factor
S is hillslope gradient factor
C is ground cover factor
P is land use practice factor

RUSLE variables differ considerably across the Australian continent. RUSLE provides a method for estimating spatial patterns of erosion using consistent information for each factor. Limited spatial data for contour cultivation and bank systems, meant that an assessment of land use practice was not possible. Hence, reductions in soil loss due to use of soil and other conservation practices in intensive land use areas were not predicted.

Innovative use was made of time series, remote sensing imagery and daily rainfall when combined with updated spatial data for soil, land use and topography. Problematic to the standard annual application of the RUSLE is the pronounced wet-dry precipitation regime in Australia's tropics and Mediterranean climate areas. To overcome this limitation, the model was applied to RUSLE on a monthly averaged basis (calculating appropriate erosivity and cover factors for each month) to represent the erosive potential of rainfall for each temporally distinct period. It used 20 years (1980 - 1999) of daily rainfall data mapped across Australia (Jeffrey et al. 2001) and 13 years (1981 - 1994) of satellite vegetation data for this purpose. Mapped soil properties from the Australian Soil Resources Information System (Appendix 2), and land use mapping (see Figure 1.2) were used together with new techniques to derive fine scale hillslope topography from the coarse resolution national digital elevation model.

View Figure 5.2. Schematic implementation of the Revised Universal Soil Loss Equation.

Limitations in the resolution of the data sets are evident where predictions of hillslope erosion are clearly overestimated in areas where land use is classified as grazing but grazing pressure/stocking rates are low such as the basins north of Stewart Basin in Far North Queensland (including Jacky Jacky and Olive-Pascoe). However, the modelling predictions do serve to highlight the potential sensitivity of these environments.

Predictions of sheetwash erosion under present land use needed to be put in context of erosion under native vegetation cover. Natural or `pre-European' erosion was predicted using the same procedure, with a cover factor for native vegetation and keeping the other factors of soil erodibility, rainfall erosivity and topography as for the present day. The cover factor for native vegetation before agricultural development was determined by interpolation and extrapolation. This was based on the climatic, topographic and geological characteristics of the current distribution of native vegetation.

Gully erosion

Gully erosion was mapped by predicting the density of gullies across the assessment area. Gully density is measured as the kilometre of gully length per square kilometre of land. It was assessed by interpreting aerial photographs.

Three separate data sources were used across the assessment area:

Streambank erosion

All streams with a catchment area greater than 50 km2 and a length greater than 5 km were mapped across the assessment area. The proportion of stream with cleared native riparian vegetation was determined by intersection of the streams with coverage of native vegetation in 1995 (Australian Land Cover Change project, resolution of 100 m; BRS 2000). These data are only a crude measure of riparian condition as the 100 m resolution fails to identify narrow bands of remnant riparian vegetation in cleared areas and narrow valleys of cleared land penetrating otherwise uncleared land.

Erosion was assumed to only occur on sections of river with cleared riparian vegetation. The mean annual rate of bank erosion was calculated as a function of bankfull discharge to reflect globally observed patterns of erosion rate.

Sediment transport through river networks

The SedNet model (Sediment river Network model) developed for this assessment constructed a mean annual sediment budget sequentially downstream through each link*of a river network. A sediment budget accounts for:

It is essential to predict deposition because only a small proportion of the sediment supplied to river systems is actually exported to the mouth. Separate sediment budgets were constructed for bedload and suspended load because of the quite different transport processes.

Hillslope, gully and bank erosion, and upstream sediment export all deliver suspended sediment (fine-grained sediment suspended in the river water column) to a river link (Figure 5.3) during storms and floods. This contributes to flow turbidity. It also carries substantial amounts of nutrients (see Nutrient loads to Australian rivers and estuaries section).

Floodplains and reservoirs are the main areas of net deposition of suspended sediment. Deposition was modelled using a simple `sediment residence time' approach.

Bedload is the coarse sand, gravel and stones that roll and bounce down the bed of a river. Net deposition occurs on the bed when the loading over time exceeds the sediment transport capacity of the stream (Figure 5.4). Sediment transport capacity is the maximum amount of sediment that a river can carry given its discharge, slope, width and hydraulic roughness. All sediment in excess of capacity is deposited, changing the morphology of the river and impacting on the physical habitat. If the total loading over time is less than the transport capacity, then all sediment is delivered downstream and there is no change to habitat. To predict the location of sand slugs, bedload deposition was expressed as the depth of accumulation since European settlement.

Methods used to construct budgets for river sediment (Prosser et al. 2001) included detailed mapping of floodplains (Pickup & Marks 2001), synthesis of extensive river gauging records, and measurements and modelling of river widths. Predictions have been validated against a range of indicators, including observed river sediment loads.

The SedNet model calculates:

*A link is the stretch of river between tributaries.

View Figure 5.3. Conceptualisation of the suspended sediment budget for a river link.

View Figure 5.4. Conceptualisation of the bedload budget for a river link.

CONTINENTAL SHEETWASH AND RILL EROSION

Northern Australia has higher predicted erosion potential than the south (Figure 5.5)-northern summer rains are highly erosive and can coincide with relatively little ground cover. A zone of high potential erosion also occurs on the western slopes of the Great Dividing Range along the belt of cereal cropping land. Winter is the most erosive season in southern Australia, but where rains fall on well vegetated land erosion potential is negligible.

Results of this modelling represent soil erosion potential because:

View Figure 5.5. Mean annual sheetwash and rill erosion rate.

Each year, 1.2 billion tonnes of soil has the potential to be moved on hillslopes in river basins containing intensive land use. Using the classes represented in Table 5.1, 11% of Australia experiences high sheetwash and rill erosion potential. About 23% of the area is eroded at a rate greater than the continental average of 4.4 t/ha/yr. These statistics show the value to be gained from targeting erosion control to particular problem areas.

The implications of sheetwash and rill erosion for farm productivity have not been assessed.

Table 5.1 Sheetwash and rill erosion rates divided into three classes.
Erosion rate Volume (t/ha/year) Proportion of land (%)
Low < 0.5 39
Medium 0.51 - 9.9 50
High > 10 11

Pre-European sheetwash and rill erosion

Comparisons of current soil erosion potential to that of pre-European settlement conditions under native vegetation cover were needed to understand current issues of land degradation (Figure 5.6). The simplest indicator of potential for accelerated erosion is a map of the ratio of current to pre-European potential erosion (Figure 5.7).

The gross pattern of contemporary soil erosion potential being higher, in the north, reflects the natural distribution of soil erosion across the continent. Within each climatic zone, areas of significantly accelerated erosion potential have occurred-shown by the ratio of present to pre-European potential (Figure 5.7). These are areas where cover has been reduced at least seasonally. While the overall erosion rate is low in the cereal belt of Western Australia, it is still many times higher than the naturally very low rate. Similar results were found for the cropping belt from Victoria through to Queensland and the extensive grazing lands and tropical crop lands of north Queensland. On average, sheetwash erosion has accelerated by three times the natural rate.

View Figure 5.6. Predicted pre-European mean annual sheetwash and rill soil erosion rate.

View Figure 5.7. Present to pre-European sheetwash and rill erosion ratio.

Sheetwash in croplands

The most intensive cropping land uses have the greatest potential to cause accelerated erosion (Table 5.2). Crop lands make up only 8% of the assessment area. Sugar cane and tropical fruit crops are of particular concern, as they are located in areas of high rainfall erosivity. Where they occur on sloping land, soil erosion can only be stopped by retaining good cover at all times. Cereal and legume crops in southern Australia are less susceptible to accelerated erosion because of low rainfall during times of low cover.

Considerable effort to reduce soil erosion potential in croplands includes minimum tillage, stubble retention and contour banking-practices that are widely, but not universally,
adopted. The sugar cane industry reports 80% adoption of minimum tillage, green cane harvesting and trash blanketing, reducing soil erosion rates on sloping land from the order of 100 t/ha/yr to 5 - 10 t/ha/yr. Tillage is still necessary when planting a new crop, creating a risk of accelerated soil erosion at these times.

Riparian filter strips offer a last line of defence to soil erosion-protecting streams from sediment and attached nutrients lost from farm land. Attention has only recently been given to riparian management as a means of mitigating against the impacts of some crop land use practices. Filter strips of less than 5 m width can be effective in protecting against erosion rates of less than 20 t/ha (Prosser & Karssies 2001).

Table 5.2 Summary of erosion by land use for river basins containing intensive agriculture.




Land use Area Total erosion Erosion rate with no conservation practice Rate of acceleration Erosion rate * under best practice
(km2) (t/yr) (t/ha/yr) (t/ha/yr)
Closed forest 22 000 2 552 000 1 1 N/A
Open forest 228 000 6 900 000 <1 1 <1
Woodland (unmanaged lands) 220 000 103 400 000 5 3 N/A
Commercial native forest production 153 000 5 800 000 <1 1 <1
National parks 86 000 76 200 000 9 1 N/A
Cereals excluding rice 180 000 38 933 000 2 10 <1
Legumes 22 000 740 000 <1 3
Oilseeds 6 000 2 382 000 4 10
Rice 1 500 115 000 1 6
Cotton 4 000 2 784 000 7 11
Sugar cane 5 000 18 623 000 40 57 5
Other agricultural land use 2 000 2 329 000 54 34 ~2
Improved pastures 190 000 41 429 000 2 5 N/A
Residual/native pastures 1 673 500 957 939 000 6 3 N/A
Total area of river basins containing
intensive agriculture		 2 793 000 1 260 126 000** 5

* Indicative values obtained from erosion plot studies where available.

N/A Not applicable

** It was predicted that on average there is a potential for about 4.8 billion tonnes of soil to be moved by sheetwash and rill erosion across the Australian continent each year.

Sheetwash in grazing lands

Grazing is the main land use contributing to total soil erosion across assessed river basins, because of the vast areas involved and their location in northern Australia (Table 5.2). Grazed land makes up 75% of the assessment area and is composed of woodlands as well as pastures-the basis of the beef industry in northern Australia. It is predicted that erosion under pasture lands has doubled from natural conditions, with a five-fold increase for improved pastures. Soil erosion under woodlands and native pastures contributes 86% of total assessed area.

Sheetwash erosion is much harder to manage in grazing than in cropping areas because of the:

Structural works and other soil conservation practices are usually impractical in these areas. The greatest scope for reducing soil loss is through improved pasture and stock management aimed at maintaining adequate ground cover at all times (including drought planning, off-stream watering, cell grazing and management of pasture species). These issues are of greatest importance in the northern grazing lands where river suspended sediment loads have most increased and where sediment delivery to the coast is more likely.

Factors contributing to sheetwash and rill erosion

Maps of factors contributing to soil erosion potential illustrate how national patterns were derived (Figure 5.8).

View Figure 5.8A. Factors contributing to sheetwash and rill erosion - rainfall erosivity & woody cover.

View Figure 5.8B. Factors contributing to sheetwash and rill erosion - rainfall erosivity & woody cover.

View Figure 5.8C. Factors contributing to sheetwash and rill erosion - rainfall erosivity & woody cover.

View reporting regions map

GULLY EROSION

Gully erosion is a natural process. Prior to European settlement gullies eroded episodically for a hundred years or so, every few thousand years, and in a few valleys at any one time. The current extent of erosion up valleys is also considerably longer than occurred naturally. The current extensive and relatively synchronous erosion of many valleys is unprecedented for at least the last 15 000 years.

Many of the gullies of today's agricultural landscapes were formed soon after clearing of native vegetation and have since stabilised. They continue to produce poor quality run-off but have ceased to be a major erosion problem, despite considerable lengths of bare, steep banks. Such banks are not good indicators of active erosion, which is better detected by observing changes to gully position over time.

Average gully density within the assessment area was 0.13 km/km2. This includes an approximate total of 325 000 km, that on average had produced 44 million tonnes of sediment each year. It is estimated that gullies have eroded 4.4 billion tonnes of sediment since European settlement of Australia. In most cases, gullies are well connected to streams and rivers, so that the vast majority of the eroded sediment has been delivered into the river network.

The greatest on-site problem of gully erosion is its threat to fences, tracks, roads and buildings.

Amelioration of gully erosion can range from fencing out stock to revegetation, through structural works to stabilising a gully head, to filling the gully and constructing erosion control dams and grassed waterways. These works can range in cost from $2000 to $50 000 per kilometre of gully, depending upon the nature of works. To treat all gullies in the river basins of the assessment area, at an average cost of

View Figure 5.9. Area of moderate and high gully density in each region river basin containing intensive agriculture.

$20 000 per kilometre, would total $6500 m. Such resources are clearly not available, nor are they needed, for gullies naturally stabilise over time. Remedial works should be focused on those gullies that continue to erode and threaten structures, or those that yield considerable sediment or poor quality water. Local observations of gully movements and water quality measurements provide the information required to identify problem gullies.

View Figure 5.10. Gully density.

Table 5.3 Range of gully densities observed.
Erosion class Gully density
(km/km2)		 Mean annual sediment yield
(t/ha/yr)
Low < 0.1 < 0.15
Medium 0.1 - 1 0.15 - 1.5
High 1 - 3.5 1.5 - 5.3

* Gully density can be converted into a soil erosion rate by considering the approximate age of a gully and the volume of soil removed to form the gully. An average gully is 5 m wide and 2 m deep. One kilometre of gully would then produce 10 000 cubic metres (approximately 15 000 tonnes) of sediment for each square kilometre of land. If that volume was eroded over an average gully age of 100 years, the mean annual rate of erosion would be 1.5 t/ha/yr.

Streambank erosion

Degradation of riparian vegetation is a major cause of streambank erosion, since it increases the susceptibility of streambanks to erode during floods:

Four-fold increases to stream width and doubling of stream depth since European settlement is common along cleared creeks and rivers. This can be a significant source of sediment to the river, as well as, threatening valuable floodplain land and infrastructure.

Erosion of banks may not occur immediately following loss of riparian vegetation. Its intensity will also differ between rivers, and in some cases, large floods cause the majority of erosion (see box).

Variations in streambank erosion

Riparian vegetation in the Hunter River system on the New South Wales Coast was cleared late in the nineteenth century, a common practice throughout coastal New South Wales. Major erosion of streambanks did not occur until a sequence of large floods occurred in the 1950s. Floods between 1946 and 1955 caused an average 304 m of erosion along an 82 km stretch of the Hunter River (Erskine & Bell 1982).

In higher energy rivers (e.g. the Bega River), cleared banks were capable of being eroded by even relatively small floods and channel widening occurred in the first few decades of clearing (Brooks & Brierley 1997).

Rivers, such as the lower Lachlan River in central western New South Wales, are of such low energy that they have suffered relatively little erosion despite extensive clearing.

Mapping shows that across the assessment area, 65% of the river length has cleared riparian vegetation, including some 120 000 kilometres of cleared streams. Restoration of native vegetation requires $1.2 billion (at a conservative cost of $10 000 per kilometre for fencing and replanting with volunteer labour). Such investment is not currently feasible and not all areas of cleared riparian lands will be a high priority for restoring degraded streams. However, effort is needed to strategically target restoration works prioritised at regional, State and national levels. Priority could be given to areas where bank erosion creates problems of sediment deposition or loss of land, and where riparian vegetation is limiting ecological health.

View Figure 5.11. Estimated proportion of native vegetation removed along stream banks in river basins with intensive agriculture.

View reporting regions map

SEDIMENT DELIVERY TO STREAMS

Sheetwash erosion, gully and streambank erosion all supply sediment to river networks. River sediment budgets were constructed to route this sediment through each river basin accounting for deposition and calculating net export. Table 5.4 summarises these terms across the assessment area.

Table 5.4 Components of sediment supply (million tonnes per year).
Gross sheetwash and rill erosion 666*
Delivery to stream from sheetwash
and rill erosion		 50
Gully erosion 44
Streambank erosion 33
Total sediment supplied to rivers 127
Total suspended sediment stored 66
Total bed sediment stored 36
Sediment exported from rivers 25
Total of stores and losses 127

Sheetwash and rill erosion dominate total erosion processes, but modelling estimates indicate that less than 8% of soil moving from hillslopes reaches the stream. Field measurements of differences between erosion measured on small plots to that measured at the scale of small catchments and whole hillslopes (Edwards 1993) indicate that much of this soil is deposited a relatively short distance downslope. Sediment delivery from gully and streambank erosion are of comparable magnitude to that from modelled estimates of sheetwash and rill erosion. The certainty of these predictions decreases from sheetwash and rill erosion through gully erosion to streambank erosion. They have enough certainty to demonstrate that each process needs to be considered in regional assessments of river sediment loads.

View Figure 5.12. Estimated amounts of sediment supplied to streams from each erosion process.

* Does not include sheetwash and rill erosion estimates for the Gulf, Western Plateau or Northern Territory as river budget assessments were not undertaken in these areas.

Strong regional differences in dominant sediment source occur across Australia (Figure 5.12):

Patterns within these regions (Figure 5.13) also show that gully erosion processes dominate in the drier parts of Queensland.

View Figure 5.13. Ratio of hillslope to channel (gully and streambank) sediment sources by river link subcatchments.

The strong regional differences between sediment sources are significant. Techniques to prevent gully and streambank erosion are quite different to those applied to sheetwash and rill erosion:

Only 20% of the total load supplied to rivers is actually exported from rivers (Table 5.4). The remainder is deposited within the rivers, on floodplains or in reservoirs. This highlights the importance of assessing river loads for patterns of deposition in order to appropriately target management strategies. For bedloads, it is the sites of deposition that produce the downstream impacts of accelerated erosion.

THE FITZROY BASIN

A Queensland region dealing with soil erosion

The Fitzroy regional case study highlights land management, soil erosion and capacity for change within the grazing industry. Regional development plans need to include options for soil management to improve or maintain ecosystem health while taking into account the social and economic needs, and motivation and capacity of rural industries and other natural resource management stakeholders to implement these options.

Profile of the Fitzroy

Soil erosion in the Fitzroy: national context

The Fitzroy Basin covers approximately 14 million hectares and makes up 8.5% of the national Audit assessment area. Sheetwash (62%) and rill erosion (24%) processes dominate gully and river bank erosion (12%); nationally, 40% of sediment is delivered to streams from hillslope erosion, 34% from gully erosion and 26% from streambank erosion.

The Fitzroy Basin is responsible for 20% of all sediment delivered from hillslopes to streams nationally. Of the 21 million tonnes of fine sediment reaching the coast nationally, 12% (2.6 million tonnes) come from the Fitzroy Basin, and the area has a specific sediment yield of 0.18 t/ha/yr, slightly higher than the national average. Fine sediment loads in the Fitzroy Basin are predicted to have increased 15 times above the natural rate since European settlement, which is well below the national average of 100 times the natural rate.

Gully erosion contributes 0.28 t/ha/yr to streams (national average is 0.26 t/ha/yr). There are significant areas of low to moderate gully density, with 62% of the basin having a gully density of 0.1 to 1 km per kmē, compared with the national figure of 37%. A small area of very high gully density (3 to 3.5 km per kmē) occurs in the Nogoa Catchment. Gully erosion is not considered a great concern for the Fitzroy Basin (<1% of the basin falls into the category of high gully density).

Around 15% of sediment in the Fitzroy Basin is derived from streambank erosion, a natural process which is accelerated in areas of degraded riparian and streambank vegetation and poor stability. The Fitzroy Basin contains 15 500 km of streams, of which around 50% have degraded riparian vegetation (just below the national average). All coarse sediment eroded through gully and river bank erosion remains deposited in downstream tributaries leading to 13% of the river network with in-stream sediment deposition greater than 30 cm (poor in terms of river health but lower than the national average of 16.5%).

While higher loads of fine sediment may arise from cropped lands, erosion management needs to focus on maintaining surface cover on land used both for cropping and grazing.

View Annual sediment from bank, hillslope and gully erosion.

Table 5.5 Water-borne erosion: Fitzroy in context.
Attribute National* Fitzroy basin Fitzroy as % of national
Area (million ha) 167 14 8.5
Stream length ('000 km) 181.5 15.5 8.5
Sediment sources
Bank erosion (Mt/yr) 33 2 6.0
Gully erosion (Mt/yr) 44 4 9.0
Hillslope erosion (Mt/yr) 50 10 20.0
Total (Mt/yr) 127 16 12.5
Sediment delivery 21 Mt/yr 2.6 Mt/yr 12
to coast 0.13 t/ha/yr 0.18 t/ha/yr
In-stream sedimentation >30 cm
Stream length ('000 km) 30 000 2 000 6.5
Percentage of total 16.5% 12%
Degraded riparian vegetation
Stream length ('000 km) 118 600 7 800 6.5
Percentage of total 65% 50%

* The erosion assessment was undertaken for river basins containing intensive agriculture.

DEPOSITION ON RIVER BEDS

Accumulation of sands and gravels on river beds since European settlement of Australia is a major issue for river health.

Many coastal streams of Australia have beds of stones (cobbles), boulders and rock outcrop that are ideal habitat for benthic algae, macroinvertebrates and some fish species. Scour pools and undercut banks, or pools surrounding fallen debris also occur and are important refugia and breeding areas for fish and other biota. Fine sediment and nutrients that accumulate between the rocks or in the deep pools during low flows, are periodically flushed out, cleaning out accumulated debris and re-initiating surfaces for fresh colonisation by algae.

Where the supply of sands and gravels from upstream exceed a river's flushing capacity, this material starts to accumulate covering the rocky surfaces and filling deep pools. The sand and gravel are too unstable for growth of benthic algae and loss of deep pools also means a loss of refugia and breeding grounds. In the most extreme example of deposition, the bed of the river becomes an inhospitable flat sheet of dry sand during low flow.

The semi-arid areas of northern Australia and the western Murray-Darling Basin have naturally sandy river beds as a result of the climate, natural erosion processes and low slopes which result in sand supply exceeding capacity. It should not be assumed that these are impacted streams and the predictions suggest that they have suffered little net accumulation of sand.

View Figure 5.14. River bed sediment accumulation.

Extensive deposition occurs in reaches immediately downstream from areas of high gully and streambank erosion (Figure 5.14).

There are 30 000 km of stream in the assessment area predicted to be impacted by sediment accumulation of greater than 0.3 m since European settlement.

This assessment only includes streams affected by supply of sand from gully and streambank erosion. Another significant source of debris is from alluvial mining for gold or tin, or from the supply of mine tailings to rivers (e.g. in the Ringarooma, George, King, Queen and South Esk rivers in Tasmania; the Tambo River, Ovens River, Yackandandah Creek and Bendigo area in Victoria; the Rocky and Molongolo rivers in New South Wales; and Magela Creek in the Northern Territory).

This assessment presents a snapshot of the current situation. The accumulation is really a pulse of material that will gradually move through the system over time driven by slow movement of sand during flood events. This means that even if source erosion was stopped today, large areas of sand deposition would continue to progress through river systems. Much of the sediment deposited in the upper tributaries of the Murray-Darling Basin was delivered to streams by erosion in the late nineteenth and early twentieth centuries. The sediment will continue to have impacts unless stabilised, extracted or flushed.

SUSPENDED SEDIMENT LOADS

Suspended sediments, mainly clay and silt particles, impact on streams by:

The mean annual yield of suspended sediment (Figure 5.15) is lower in southern Australia than northern Australia–reflecting lower erosion rates in the south. Within a river basin, suspended sediment loads grow as catchment area increases. Displaying loads per unit area (e.g. tonnes/hectare/year) removes this effect and shows where high mean annual loads exist.

View Figure 5.15. Mean annual suspended sediment yield per hectare of catchment.

River discharge was not directly used to predict sediment loads and any such correlations are coincidental. Rivers can carry as much suspended sediment as is supplied, so there is no physical reason why rivers with high discharges have high sediment load (e.g. the Murray River has a similar mean annual flow to the Burdekin River, yet the Burdekin River yields three times more sediment).

Naturally strong differences occur across Australia in soil erosion and these are expected to be reflected in river sediment loads. It is also reasonable to expect that local ecosystems adjust to these loads. Loads of suspended sediment can be expressed as a ratio of the predicted natural suspended loads (Figure 5.16). The natural suspended load (pre-European) is an estimate of sheetwash and rill erosion based on the assumption that there was negligible bank and gully erosion under pre-European conditions.

View Figure 5.16. Ratio of current suspended load to pre-European suspended load.

View reporting regions map

SEDIMENT EXPORT FROM RIVERS

Sediments from rivers are eventually deposited in estuaries and inshore marine environments. Increased rates of deposition can:

Only 20% of sediment delivered to streams in the assessment area were predicted to actually reach estuaries, so it cannot be concluded that increased erosion from a catchment necessarily results in significantly increased sediment export to an estuary or the coast. Erosion from upstream requires efficient delivery of the sediment through the river network to link to an estuary. In many of the bigger catchments, the sediment source is hundreds of kilometres from the coast and has many opportunities for deposition onto floodplains or into reservoirs.

View Figure 5.17. Mean annual sediment export from rivers in each region.

The origin of the sediment in large coastal catchments is of greater importance than the actual export rate. The assessment traced the sediment load of each river reaching the coast to determine which were the contributing subcatchments (expressed as sediment loss [tonnes/ha/year]). Subcatchments making substantial export contributions to the coast are those with high erosion and limited deposition potential between the source and sea–subcatchments close to the coast are more likely to contribute to the coastal yield (Figure 5.18).

Ninety percent of sediment exported from rivers comes from only 20% of the contributing catchment area. While soil erosion is a widespread issue across the assessment area, targeted management can be used to address specific problems. If the goal is to reduce sediment exports from rivers then remedial works can be focused on particular sediment sources, and the land uses and erosion processes found there. Relatively little attention is needed for the rest of the catchment. Sediment delivery to the coast is not the only concern and the same principles can be applied where a reduction in sediment delivery to particular reservoirs, lakes or individual river reaches of high value is required.

View Figure 5.18. Contribution of sediment to the coast.

IMPLICATIONS FOR AGRICULTUREPhoto 5.1

Water-borne soil erosion impacts on river, estuary and marine resources and is therefore a major issue for Australian agriculture and catchment management. It causes unsustainable losses of soil for agriculture that far exceed rates of soil development, by as much as 50 times in some areas.

Hillslope erosion (sheet and rill erosion) remains high in Australia's tropical northern regions during the wet season, and especially in the semi-arid woodlands and arid interior.

In southern and eastern Australia, gully erosion, while inactive in many previously formed gullies, persists as the major erosion process affecting river condition. Sediment from these previously active gullies has affected about 10 000 km of stream length in the Murray-Darling Basin alone. These rivers, now with coarse sand accumulations in stream beds, exacerbate flooding and smother native fish habitat.

Active gully erosion is still occurring in northern Queensland and in south-western regions of Western Australia.

River bank erosion is also a major problem. Extensive lengths (120 000 km) of riparian vegetation along eastern Australia's rivers and streams are degraded and require rehabilitation. Where these landscape resources are intact, they protect the integrity of banks against erosion. Priority areas include much of the Murray-Darling Basin, South Australia and south-western regions of Western Australia.

Sediment delivery to regional streams, rivers and marine estuaries remain high priorities in specific catchments. Deposition of sand and suspended sediments in streams and rivers is greatest in the Murray-Darling Basin, coastal regions of New South Wales, south-east Queensland and the Glenelg region of Victoria.

From a near-shore and estuary perspective, about 90% of suspended sediment loads reaching marine and estuarine environments are derived from 20% of agricultural catchments, particularly in coastal regions of Queensland and New South Wales. On an average annual basis about 25% or 12 million tonnes of sediment delivered to streams is discharged into the Great Barrier Reef lagoon. This is predicted to be approximately three times greater than the natural load, with consequent impacts on estuaries and marine fisheries, seagrasses and near-shore coral reefs.

National, State and regional priorities for natural resources management can now be re-appraised in the light of these findings. The effects on-farm are irreversible and impacts off-farm which will continue for many generations. Catchment management and industry priorities, particularly in terms of implementing improved practice are essential. Total impacts are likely to be equal to, if not greater than, those of dryland salinity. It is an imperative that soil management targets hillslope, gully and river bank erosion in the various regions of Australia. Management approaches will differ across Australia's catchments as processes that supply sediment to rivers also differ.

REFERENCES

Brooks A. & Brierley G. 1997, 'Geomorphic responses of lower Bega River to catchment disturbance, 1851-1926', Geomorphology vol. 18, pp. 291-304.

BRS, 2000, http://www.daff.gov.au/content/publications

Edwards K. 1993, 'Soil erosion and conservation in Australia', in World Soil Erosion and Conservation, D. Pimental (ed.), Cambridge University Press, Cambridge.

Erskine W.D. & Bell, F.C. 1982, 'Rainfall, floods and river channel changes in the upper Hunter', Australian Geographical Studies vol. 20, pp. 183-196.

Ford G.W., Martin J.J., Rengasamy P., Boucher S.C. & Ellington A. 1993, 'Soil Sodicity in Victoria', Australian Journal of Soil Research vol. 31, pp. 869-909.

Gallant J. 2001, Topographic scaling for the NLWRA sediment project, CSIRO Land and Water Technical Report 27/01, Canberra.

Graham O.P. 1989, Land degradation survey of NSW 1987-88: Methodology, SCS Technical Report No. 7, Soil Conservation Service of NSW.

Hughes A.O., Prosser I.P., Stevenson J., Scott A., Lu H., Gallant J. & Moran C.J. (in press) 'Gully density mapping for Australian river basins', Proceedings of the Third Australian Stream Management Conference, Cooperative Research Centre for Catchment Hydrology, Melbourne.

Jeffrey S.J., Carter J.O., Moodie K.B. & Beswick A.R. (2001), 'Using spatial interpolation to construct a comprehensive archive of Australian climate data', Environmental Modelling and Software vol. 16 pp. 309-330.

Jones M-A. 2000, Technical Report 3 -Theme7-Catchment Health-Fitzroy Implementation Project, Queensland. Queensland Department of Natural Resources.

Karssies L. & Prosser I.P. 2001, 'Designing grass filter strips to trap sediment and attached nutrient', in I. Rutherford, F. Sheldon, G. Brierly & C. Kenyon (eds), Third Australian Stream Management Conference, Brisbane 27-29 August 2001, Cooperative Research Centre for Catchment Hydrology, Melbourne.

Lovett S. & Price P. (eds) 1999, Riparian Land Management Technical Guidelines, Volume one: Principles of sound management, LWRRDC, Canberra.

Lu H., Gallant J., Prosser I.P., Moran C., & Priestley G., 2001a, Prediction of Sheet and Rill Erosion Over the Australian Continent, Incorporating Monthly Soil Loss Distribution, Technical Report 13/01, CSIRO Land and Water, Canberra, Australia.

Lu H., Raupach M. & McVicar T. 2001b, 'A robust model to sperate remotely sended vegetation indices into woody and non-woody cover and its large scale application using AVHRR NDU2 Time Series', Technical Report 135/01, CSIRO Land and Water, Canberra, Australia.

Pickup G. & Marks A. 2001, Identification of floodplains and estimation of floodplain flow velocities for sediment transport modelling, NLWRA Sediment Project (CLW 12).

Prosser I.P., Hughes A.O., Rustomji P., Young W. & Moran C.J. 2001, Assessment of River Sediment Budgets for the National Land and Water Resources Audit, CSIRO Land and Water Technical Report (in press).

QDNR 2000, http://www.nrm.qld.gov.au/silo/index.html

Renard K.G., Foster G.A., Weesies D.K., McCool D.K. & Yoder D.C. 1997, Predicting soil erosion by water: A guide to conservation planning with the revised universal soil loss equation, Agriculture Handbook 703, United States Department of Agriculture, Washington DC.

Wallbrink P.J., Olley J.M., Murray A.S. & Olive L.J. 1998 'Determining sediment sources and transit times of suspended sediment in the Murrumbidgee River, NSW, Australia using fallout 137Cs and 210Pb' Water Resources Research vol. 34, pp. 879-887.

Wasson R.J. 1994 'Annual and decadal variation of sediment yield in Australia, and some global comparisons', International Association of Hydrological Sciences Publication No. 224 , pp. 269-279.

Wischmeier W.H. & Smith D.D. 1978, Predicting rainfall erosion losses: A guide to conservation planning, US Department of Agriculture, Agriculture Handbook No. 537, US Government Printing Office, Washington, DC.

Further information

View Soil erosion chapter of the Australian Agriculture Assessment 2001 report.

View the Australian Agriculture Assessment 2001 report.

A range of technical reports have been prepared by CSIRO Land and Water in the development of this work on water-borne erosion and sediment transport: floodplain flow velocities; gully erosion; river sediment budgets; sheet and rill erosion; topographic scaling; and vegetation cover

Floodplains flow velocities

Floodplains may act as sediment sinks during high flow conditions and, as such, may limit downstream transmission of material eroded from hill slopes and upstream gully networks. Floodplain and valley topography will also influence river channel behaviour, by providing limits on lateral migration or channel widening in confined reaches.

Sedimentation on floodplains requires a different approach to that implemented in the hill slope model. On hill slopes and in channels, sediment transport results from overland flow and the transport rate is a function of discharge and energy slope. Uniform flow is assumed, making energy slope the same as terrain slope, and discharge is a function of upslope area. Erosion and deposition are calculated from the difference between upstream and downstream sediment transport capacity. On floodplains, sediment transport is accomplished by overbank flow and the major process is transmission or deposition of sediment. Erosion is limited and is largely accomplished by channel widening or channel lateral migration.

Sediment transport on floodplains involves mainly fine material transported in suspension or as wash load. Only a small proportion of transported bed material reaches the floodplain from the channel during overbank flow and most of that is quickly deposited. Deposition of suspended load occurs in areas of low water velocity, in backwater zones, and in areas of standing water in off-channel storages. Identifying these zones of deposition and estimating their relative intensity requires information on floodplain hydraulics.

Flow on floodplains is often complex and may require hydrodynamic models to deal with complex topography. Use of these techniques is not justified given the limited resolution of information on floodplain elevations over the project area. We have therefore taken a simpler approach in which floodplain flow is assumed to be steady in time, sub-critical or critical in state, and gradually varied in space. This allows us to apply a step-backwater approach to flow profile analysis and floodplain flow calculation that takes into account expansions and contractions in the active flow zone. We also attempt to distinguish between the active flow zone and inundated areas such as backwater zones that do not contribute to downstream flow in the floodplain cross section. Floodplain behaviour is represented by single flow equivalent to the 25 year flood on the annual series.

View or technical report on the " Identification of floodplains and estimation of floodplain flow velocities for sediment transport modelling" by Geoff Pickup and Alan Marks (PDF 0.5 MB)

Gully density modelling

A gully density map for the more closely settled areas of Australia has been generated, covering some 1.7 million kmē. Gully density measurements were obtained from aerial photographs and previous land degradation reports. These data were used to build a map based regression tree models of gully density. The models are based upon environmental attributes available at the continent scale. The model rules were applied across the assessment area to predict gully density in places where no measurements were available. Results show high gully density in the eastern highlands and tropical grazing lands. The rules for prediction are complex with the results being affected by many environmental variables including land use, geology, soil texture, rainfall, indices of seasonal climate extremes and terrain-based attributes such as slope and hill slope length.

View or technical report on "Gully erosion mapping" by Andrew O. Hughes, Ian P. Prosser, Janelle Stevenson, Anthony Scott, Hua Lu, John Gallant and Chris J. Moran (MS Word 0.6 MB)

River sediment budget

This document describes methods used in the National Land and Water Resources Audit rivers and sustainable agriculture projects to represent the erosion of sediment from riverbanks and the propagation of gully, hillslope and riverbank sourced sediment through a river network. Essentially it describes methods for constructing sediment budgets through river networks. The budgets include storage of sediment on floodplains, in the bed of the river and in reservoirs. Calculations are made of the sediment output from each river link and the contribution of sediment to the coast, or any other receiving body, from all sub-catchments. The budgets treat two types of sediment: suspended sediment and bedload. A suit of ArcInfoTM programs are used to define river networks and their sub-catchments; import required data; implement the model; and compile the results. These are referred to collectively as the SedNet model: the Sediment River Network Model.

View or technical report on "Constructing river sediment budgets" by Ian Prosser, Paul Rustomji, Bill Young, Chris Moran and Andrew Hughes (PDF 0.6 MB)

Sheet and rill erosion

A major issue in Australian land management is soil erosion and the consequent reduction of productivity. The off-site effect of soil erosion is the degradation of water quality in streams and water storages. Measurement of soil erosion is time consuming and data on soil erosion rate is limited to a few sites. Those sparse measurements provide little information about the spatial distribution of soil loss rate across the nation. This report describes a spatial modelling framework which is used to predict an Australia-wide sheet and rill erosion. It is based on the Revised Universal Soil Loss Equation (RUSLE) using time series of remote sensing imagery and daily rainfall combining with updated spatial data for soil, land use and topography. The results are presented as a geo-referenced annual averaged soil loss map and its monthly distributions. It is found that the north part of the country has higher erosion potential than the south of the country. The prediction confirms that agricultural land use has higher erosion rate compared with most natural vegetated lands and that erosion potential differs significantly between summer and winter periods.

View or technical report on " Prediction of sheet and rill erosion over the Australian continent, incorporating monthly soil loss distribution" by Hua Lu, John Gallant, Ian P. Prosser, Chris Moran, Graeme Priestley (PDF 1.7 MB)

Topographic scaling

Prediction of hillslope soil erosion at the continental scale is complicated by the scale mismatch between erosion processes and the scale of available data. The best topographic data available across the Australian continent is the AUSLIG 9 second DEM with a resolution of 250 m, whereas topographic control of soil erosion is at a considerably finer scale. This report describes the methods used to derive the slope length (L) and slope steepness (S) factors in the revised universal soil loss equation (RUSLE) using statistical models based on measurements from high resolution digital elevation models.

Predictive variables in the statistical models included measures of relief and topographic position from the 9 second DEM, several indicators of geology and soil type, and various measures of climatic conditions. Hillslope length varies by an order of magnitude, with a mean value of about 200 m, smaller than the resolution of the 9 second DEM. Slope varies over nearly two orders of magnitude, with a mean value of about 13%, and for slopes less than 10% is approximately twice as large as slope measured directly from the 9 second DEM.

View or technical report on "Topographic scaling" by John Gallant (MS Word 0.7 MB)

Vegetation cover

A robust model is proposed to separate Normalised Difference Vegetation Index (NDVI) time series data into woody and herbaceous NDVI using time series decomposition. The model is capable of reducing the transient, aberrant behaviour of NDVI due to sensor errors or atmospheric contamination and estimating temporally varying woody and herbaceous NDVI.

In this study, the separated NDVIs are used to estimate annual averaged woody cover and monthly averaged herbaceous vegetation covers using Pathfinder AVHRR Land (PAL) Global Area Composite (GAC) Advanced Very High Resolution Radiometer (AVHRR) NDVI data from 1981-1994 for Australia. Empirical relationships between woody NDVI and ground-based measurements of leaf area index (LAI) and foliage projective cover (FPC) are derived and compared with existing empirical relationships. The new empirical relationships are critically reviewed in relation to theoretical background and measurements. Finally, the woody cover map is compared with a high resolution woody cover map derived from LANDSAT Thematic Mapper for a 209,310 kmē area located in northern east Australia.

View or technical report on "Decomposition of vegetation cover into woody and herbaceous components using AVHRR NDVI time series data" by Hua Lu, Michael R. Raupach and Tim R. McVicar (PDF 1.2 MB)

Link to the Australian Natural Resources Data Library

Before you download

Most publications are downloadable as PDF files. Adobe Acrobat Reader  is required to view PDF files.

If you are unable to access a publication, please contact us to organise a suitable alternative format.

Key

   Links to an another web site
   Opens a pop-up window