Land Theme Report
Australia State of the Environment Report 2001 (Theme Report)
Prepared by: Ann Hamblin, Bureau of Rural Sciences, Authors
Published by CSIRO on behalf of the Department of the Environment and Heritage, 2001
ISBN 0 643 06748 5
Accelerated erosion and loss of surface soil
Environmental indicators reported on in this section as originally listed and defined in Hamblin (1998):
| Environmental indicator | |
|---|---|
| L1.1 | Change in total exposed soil surface contributing to erosion, as a percentage of land area per landcover region, stratified by major land use |
| L1.2 | a | b | | Total grazing pressure relative to net primary productivity (biomass) by landcover regions and AERs |
| L1.2A | a | b | c | | Domestic vertebrate grazing pressure per landcover region and AER |
| L1.2B | a | b | | Non-domestic vertebrate herbivores per landcover region and AER |
| L1.4 | a | b | | Surface soil loss index |
| L1.5 | Gullying index per major catchment |
| L1.6 | Change in dust storm index relative to number of high wind events by AERs and landcover regions |
| L1.8 | Implementation of new drought policies |
| L1.9 | a | b | | Percent of land managers using agreed Best Practice by land use and/or catchment |
| L1.9B | Area of forested lands in which the legal framework encourages best practice codes of forest management, and the conservation of special environmental values |
Water erosion is the principal natural process shaping land outside glacial regions, and wind erosion is associated particularly with semi-arid and arid environments. The difficulty in reporting for the State of the Environment is how to identify the human-induced accelerated erosion.
Ideally the extent and impact of human-induced erosion would be measured by comparing monitored catchments in the same geological formations and climates that are subjected to different land uses, using tracers or other distinguishing signatures to identify the causes of the erosion. There has been an increase in research and monitoring in this area during the 1990s, but Australia is too vast and its climate is too variable in many regions for this to be a comprehensive activity. Consequently, instead of measuring the extent of actual pressure, we have to make do with estimating the risk of anthropogenic pressure using surrogates that measure the potential for erosion.
Accelerated water and wind erosion occur wherever the soil surface is bare of vegetation, and where rainfall or wind events can be very intense. Human-induced erosion occurs either as a result of clearing vegetation (by wholesale clearance of woody vegetation, and by bare-soil cultivation) or by overgrazing stock, leading to soil surface exposure. Typical features that distinguish accelerated from natural erosion are listed in Table 2.
| Natural erosion processes | Resultant features | Human intervention | Resultant features of accelerated erosion |
|---|---|---|---|
| Wind erosion and deposition | |||
|
arid regions |
dunes, dune fields deflation pans stone deserts, stone fretting wind-pruned vegetation dust storms |
vegetation clearing cultivation overgrazing by stock |
trapped deposits in fences, and roads downwind bare ground wind scars in crops |
| coastal dunes | dust storms | high air-dust level in non-arid areas | |
| Water erosion and deposition | |||
|
land dissection by streams and rivers |
valley floors and slopes lakes and river channels |
vegetation clearing cultivation |
steep-sided gullies rapid headward gully extension loss of A horizon from soil profile |
|
sedimentary deposits and rock formation |
alluvial plains, terraces, fans, etc. | overgrazing by stock |
increased suspended solids and turbidity in streams, lakes |
Pressures
Bare soil surfaces exposed to erosion [L Indicator 1-1]
If bare areas expand as the result of human activities, accelerated erosion pressure increases. The extent of exposed soil can be detected by remote sensing at local and regional scales. The impact of specific erosion events (cyclones) has been reported from north-western Western Australia (Wallace and Campbell 1998) (see the gullying and sheet erosion).
Monitoring the continental extent of bare ground exposed to erosive processes has not been attempted at a small scale, and it would be too time-consuming and costly at present. However, broader, synoptic assessment of continental changes, was used for this report using the 1 km2 resolution of NOAA-AVHRR data, and its application through an extension of the Normalised Difference Vegetation Index was developed (Cridland and Fitzgerald 2000). This provides for the first time a continental estimate of standing biomass and sheet erosion.
Preliminary continental-scale application
The area of exposed soil surface (distinguished by having a reflectance significantly greater than vegetated areas) was identified for each four major land use categories in IBRA regions over the period 1991-2000, and the trend has been statistically determined in a novel method using the NDVI database (Fitzgerald and Cridland unpublished). Agricultural regions were classed as:
- intensively cultivated and irrigated,
- used for broadacre cropping and pastures, or
- relatively undisturbed but used for pastoralism.
Non-agricultural regions include vacant crown land, which is large tracts of land in the arid deserts, gazetted conservation estate, defence lands and forestry areas. Regions that have had seasonal rainfalls in the lowest decile of the long-term average were identified as areas where bare ground may have occurred as a result of drought, irrespective of management.
Limitations to this method include the coarse scale of 1 km2 resolution available to the NOAA satellite. The scale of resolution offered by Landsat imagery is required if the impact of unsealed roads, earthworks and open-cast mine sites are to be monitored. Landsat or airborne scanner technology is also needed to assess local effects of conservation and remedial measures, such as removal of stock, control of feral herbivores, adoption of conservation tillage and planting of trees. Figures 2 and 3 have been derived from this analysis.
Figure 2: Areas of increasing and decreasing bare ground compared with lowest previous value (1991-2000) located in conservation estates.
Source: Fitzgerald and Cridland 2001, commissioned for SoE 2001(unpublished)
Figure 3: Areas of increasing and decreasing bare ground compared with lowest previous value (1991-2000) located in little altered but grazed agricultural lands.
Source: ABS (2003c)
Over the period 1991 to 2000, the area of bare ground (see Figures 2 and 3) has increased in conservation reserves in three biogeographic (IBRA) regions, all in Western Australia: Esperance Plains, Warren and the Swan Coastal Plain. (The apparent increase in the area of bare ground in the West and South Coast of Tasmania is considered to be an artefact of effect of cloud on the remote sensing method.) Table 3 summarises the changes in the four land use categories for IBRA regions where statistically significant changes have occurred. The increase in bare ground in non-agricultural regions could have occurred as a result of drought, fire, pests or grazing by feral herbivores.
| Biogeographic region | Conservation estate (and forestry, defence etc) |
'Natural' grazed lands |
Broadacre crops and pastures | Intensive agricultural, irrigated lands |
|---|---|---|---|---|
| Esperance Plains | + | + | + | + |
| Warren | + | + | + | |
| Swan Coastal Plain | + | + | ||
| Coolgardie | + | |||
| Geraldton Sandplains | - | |||
| Victoria Volcanics | + | + | ||
| Tasmanian Central Highlands | - | |||
| Tasmanian West | + | |||
| Darwin Coastal | - | - | - | - |
| North East Tablelands | + | |||
| Sydney Basin | + | |||
| Brigalow Belt North | + | |||
| Wet Tropic Coast | - | - | - | |
| Central Mackay Coast | - | |||
| Einasleigh Uplands | - | |||
| Tiwi Cobourg | - | |||
| Desert Uplands | - | |||
| MacDonnell Ranges | - | |||
| Channel Country | - | - | ||
| Darling Riverine Plains | - | - |
A Normalised difference in vegetation index.
B Regression trends significant at 95% and 90% t -test.
C See Baxter and Russell (1997).
Source: Fitzgerald and Cridland 2001, commissioned for SoE 2001(unpublished).
There are some interesting decreases in bare ground in regions that have a large proportion of their area in rough grazing, where increased ground cover may well be related to reduction in grazing pressure. Of particular interest is the relationship between this evidence of reduction in bare ground in the biogeographic regions such as the Channel Country and Einasleigh Uplands and the contraction in the extent of accelerated wind erosion in the same regions.
Interpretation
Where there is a change in the bare ground in the conservation estate as well as in other land uses, it is probable that the reason is climatic. This is the case with the increase in bare ground in Esperance Plains and Warren, in the south-west of the continent, and the decreases in bare ground in Darwin Coastal, Wet Tropic Coast and Channel Country IBRAs in the north and centre of the continent.
Decreases in bare ground area (shaded green and blue in Figures 2 and 3) are larger in extent than the increases in bare ground. Decreases are concentrated particularly in those areas of northern, central and eastern Australia that are recovering from the severe droughts of the early 1990s. This supports the evidence (derived from a different analysis undertaken by ERIN) of the trend to increased 'greening-up' of vegetation to seasonal rainfall in the northern half of the continent. The bare-ground analysis, using the basal area (i.e. vegetation that contains enough green to be detected all year round), suggests that most of the increase in vegetation is coming from thickening of woody canopies, not from an increase in the seasonal (grassy understorey) flush.
The areas in the Channel Country and Finke biogeographic regions where there has been a decrease in bare ground correspond surprisingly well with areas where there has been a reduction in accelerated wind erosion, noted in the latter part of the decade in the study reported on later in this section.
It is possible that increases in bare ground at a smaller scale may be associated with land clearing and overgrazing (Brigalow Belt North, Coolgardie, North East Tablelands) or with peri-urban expansion and agricultural land intensification (Sydney Basin, Victorian Volcanics, Swan Coastal Plain). However, more detailed analysis and ground surveys of these areas would be needed to establish whether this is the case, and whether there has been a consequential loss of ecosystem function.