* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Monitoring soil erosion in South Africa at a regional scale
Soil horizon wikipedia , lookup
Soil respiration wikipedia , lookup
Terra preta wikipedia , lookup
Canadian system of soil classification wikipedia , lookup
River bank failure wikipedia , lookup
Soil food web wikipedia , lookup
Crop rotation wikipedia , lookup
Soil salinity control wikipedia , lookup
Soil compaction (agriculture) wikipedia , lookup
Soil microbiology wikipedia , lookup
No-till farming wikipedia , lookup
Sustainable agriculture wikipedia , lookup
Soil contamination wikipedia , lookup
Surface runoff wikipedia , lookup
Monitoring soil erosion in South Africa at a regional scale ARC-ISCW report no GW/A/2011/23 Project GW 59/004 Task 50 FEBRUARY 2011 REPORT FOR THE COUNCIL OF GEOSCIENCE GEOHAZARDS PROJECT BY Le Roux, J.J. AGRICULTURAL RESEARCH COUNCIL INSTITUTE FOR SOIL, CLIMATE AND WATER PRIVATE BAG X79 PRETORIA 0001 Tel: 012 310-2684 Fax: 012 323-1157 E-mail: [email protected] Executive summary Soil erosion is a major soil degradation problem, confronting land and water resource management throughout South Africa. Soil erosion involves the loss of fertile topsoil and reduction of soil productivity, as well as serious off-site impacts related to increased mobilization of sediment and delivery to rivers. Before prevention of soil erosion or remediation can be undertaken, however, the spatial extent of the problem should be established. The Council of Geoscience proposed the development of a decision support or risk-management system to assess the impacts of different geohazards including soil erosion (in association with the Agricultural Research Council – Institute for Soil, Climate and Water). Remote sensing techniques are very promising for assessing and monitoring soil erosion, especially with the development in sensor technology and the availability of space-borne data with improved spectral, spatial and temporal resolution. After a brief description of soil erosion impacts and cost implications, this report reviews mechanisms leading to sheet-rill and gully development including rainfall erosivity, topographical variables, parent material-soil associations and land use-cover interactions. The second objective of this report is a brief description of the methodologies for regional water erosion assessment using satellite remote sensing addressing mainly erosion detection of eroded areas and features; followed by the assessment of off-site impacts and erosion controlling factors, and finally data integration for erosion modelling. Third, the report describes two most recent spatial inventories namely a gully location map of South Africa created by visual interpretation and vectorization from SPOT 5 satellite imagery, and a water erosion prediction map of South Africa created by simplification of the Universal Soil Loss Equation (emphasizing sheet-rill erosion). Results from the latter mentioned spatial inventories confirm that erosion is a major soil degradation problem in South Africa. In terms of gully erosion, the Northern Cape (160 885 ha) and Eastern Cape (151 759 ha) provinces are the most severely affected, followed by KwaZulu-Natal (87 522 ha), Free State (64 674 ha), Limpopo (58 669 ha), Western Cape (25 403 ha), Mpumalanga (17 420 ha), North West (10 782 ha) and Gauteng (110 ha) provinces. In terms of sheet-rill erosion, surface areas classified as having a moderate to high erosion risk (where the average annual soil loss rate exceeds 12 t/ha.yr), the Eastern Cape (6 188 581 km2) is the most severely affected province, followed by the Free State (5 503 875 km2), Northern Cape (5 407 138 km2), Limpopo (4 422 688 km2), Mpumalanga (2 578 363 km2), KwaZulu-Natal (2 2 138 038 km2), Western Cape (2 065 319 km2), North West (1 485 263 km2), and Gauteng (775 688 km2) provinces. Finally, the report emphasizes the importance of spatially modelling areas that are currently erosion free, but under threat and in need of area-specific management e.g. protection of the current vegetation cover. The assessment of areas that are susceptible to erosion is important in policy terms because it indicates those areas which are inherently susceptible to erosion (potential risk), but which are presently protected by vegetation. Due to limited financial resources it will not be feasible, for example, to rehabilitate large gullies with large and expensive structures at a catchment scale. In addition to rehabilitating existing gullies, currently vegetated gully-free areas susceptible to gully development should be identified and protected/sustainably managed. A previous catchment scale study is described for this purpose. After quantifying the influence of factors in gully development (topographical variables, parent material-soils interactions and remotely sensed cover management), the identification of vegetated gully-free areas susceptible to gully development can be achieved by means of overlay analysis. Such a remote sensing/modelling approach is relatively simple, realistic and practical, and it can be applied or expanded to other areas of South Africa at a regional scale, thereby providing a tool to help with the implementation of a decision support or risk-management system for soil conservation and sustainable management (of problem soils as a geohazard). 3 Table of Contents Acknowledgements .............................................................................................................. 5 1. Introduction: Soil erosion impacts and cost implications ................................................... 6 2. Mechanisms leading to sheet-rill and gully development .................................................. 7 2.1 Factors influencing sheet and rill erosion ................................................................ 7 2.2 Factors influencing gully erosion............................................................................10 3. Review of current assessment/monitoring techniques .....................................................12 3.1 Eroded areas .........................................................................................................12 3.2 Erosion features ....................................................................................................14 3.3 Assessment of off-site impacts ..............................................................................15 3.4 Erosion controlling factors......................................................................................15 3.5 Data integration for erosion modelling ...................................................................18 4. Review of latest spatial inventories/datasets on sheet-rill and gully erosion at a national scale ...................................................................................................................................18 4.1 Gully erosion mapping ...........................................................................................18 4.2 Water erosion modelling (emphasizing sheet-rill erosion) ......................................19 5. The extent of sheet-rill and gully erosion in South Africa .................................................20 5.1 Extent of gully erosion ...........................................................................................20 5.2 Extent of sheet-rill erosion .....................................................................................22 6. Assessment of susceptible areas ....................................................................................24 6.1 Assessment of areas susceptible to gully erosion by means of GIS and remote sensing: Case study in the Eastern Cape ....................................................................24 7. Conclusion and recommendations ..................................................................................27 8. References......................................................................................................................29 List of Figures Figure 1: Surface area (ha) of provinces affected by gully erosion ......................................20 Figure 2: Gully location map of South Africa........................................................................21 Figure 3: Surface area (km2) of provinces where the mean soil loss exceeds 12 t/ha.yr (moderate to high)...............................................................................................................22 Figure 4: Water erosion prediction map of South Africa (emphasizing sheet-rill erosion).....23 Figure 5: Gully locations map of the catchment in the Eastern Cape Province ....................25 Figure 6: Areas susceptible to gully erosion in the catchment, Eastern Cape Province .......27 List of Tables Table 1: Satellites and sensor acronyms and names mentioned in text………………12 4 Acknowledgements This report necessitated close cooperation between the Agricultural Research Council Institute for Soil, Climate and Water (ARC-ISCW) and the Council of Geoscience (CG). More specifically, the author would like to thank (in alphabetical order) Mr. S. Foya, Mr. J. Hugo, Mr. D. Sebake and colleagues at the CG for initiating the project, as well as for funding this report. At the ARC-ISCW, thanks to Dr. H.J. Smith for managerial support. Editing of the final report by Dr. T.P. Fyfield is very much appreciated. 5 1. Introduction: Soil erosion impacts and cost implications Soil erosion is a major problem confronting land and water resources throughout South Africa (SA). Previous research indicates that more than 70% of South Africa (SA) is affected by varying intensities of soil erosion (Garland et al., 2000; Le Roux et al., 2008a). Although soil erosion is a natural process it is often accelerated by human activities, for example by the clearing of vegetation or overgrazing (Snyman, 1999). Erosion is a process of detachment and transportation of soil materials by wind or water (Morgan, 1995). Although 25% of SA is highly susceptible to wind erosion (Hoffman and Todd, 2000), water is the dominant agent causing erosion in SA and therefore the focus of this report. Water erosion occurs mostly through rain-splash, in unconcentrated flow as sheet erosion, as well as in concentrated flow as rill and/or gully erosion. Sheet and rill erosion are usually not as conspicuous as gully erosion; however, several authors agree that it is the dominant cause for loss of fertile topsoil throughout SA. Soil erosion not only involves the loss of fertile topsoil and reduction of soil productivity, but is also coupled with serious off-site impacts related to increased mobilization of sediment and delivery to rivers. Flügel et al. (2003) states that eroded soil material leads to sedimentation/siltation of reservoirs, as well as an increase in pollution due to suspended sediment concentrations in streams which affects water use and ecosystem health. For example, due to siltation, the storage capacity of the Welbedacht Dam near Dewetsdorp in the Free State reduced rapidly from the original 115 million m3 to approximately 16 million m3 within twenty years since completion in 1973 (DWA, 2011). With a catchment area of 15 245 km2, its purpose was to supply water to the city of Bloemfontein via the 115 km long Caledon-Bloemfontein pipeline. Due to the high sediment concentration in the water, the transfer from Welbedacht Dam has to be purified at the purification plant downstream. According to the latest State of Environment Report of SA (Hoffman and Ashwell, 2001; cited in Gibson et al., 2006), the off-site purification of silted dam water caused by soil erosion costs an estimated R2 billion per annum. It is therefore imperative to prevent these negative impacts and to remediate affected areas. However, before prevention of soil erosion or remediation can be undertaken, the spatial extent of the problem should be established and continually monitored by means of remote sensing. A brief overview of the mechanisms leading to sheet-rill and gully development follows. 6 2. Mechanisms leading to sheet-rill and gully development Soil erosion processes are complex and influenced by several soil erosion factors that respond differently, even in the same area. In addition, different processes and interactions are likely to emerge as dominant when crossing scale boundaries (Kirkby et al., 1996; Wilson and Gallant, 2000). Nevertheless, soil erosion by water mainly depends on the combined and interactive effects of five erosion factors, namely rainfall erosivity, soil erodibility, slope steepness and slope length, crop management, and support practice. These factors are also known as the Universal Soil Loss Equation (USLE) factors developed by Wischmeier and Smith (1978), emphasizing the sheet and rill aspects of the erosion cycle. A brief overview of these factors is discussed below, followed by a description of factors which specifically promote gullying. 2.1 Factors influencing sheet and rill erosion Foremost, sheet erosion involves the detachment and transport of soil particles by raindrop impact (rainsplash erosion) and transport by shallow overland flow (Lal and Elliot, 1994), whereas rill erosion is a process in which numerous small channels of several centimetres up to about 30 cm are formed (Bergsma et al., 1996). Causal actors are discussed individually, despite the interdependency that exists between them. Climate erosivity Erosivity is the ability of rainfall and runoff to cause soil detachment and transport. The ability of rain to cause detachment and transport is partly the result of raindrop impact, and partly due to the runoff that rainfall generates (Lal and Elliot, 1994). As a result, rainfall erosivity is based on the kinetic energy of the rain, which is a function of the rainfall intensity and duration, and the mass, diameter and velocity of the raindrops (Morgan, 1995). Rainfall erosivity indices, such as the widely used EI30 index, were developed to correlate rainfall with potential soil loss (Wischmeier and Smith, 1978). EI30 is a product of the total kinetic energy (E) of the storm multiplied by its maximum 30-minute intensity (I30). I30 is calculated as twice the greatest amount of rain falling in any 30 consecutive minutes. Some areas in SA are occasionally characterized by high intensity rainstorms. For instance, the erosivity values for the northern and eastern parts of SA are four to eight times higher than 7 the southern and western parts of the country (Smithen and Schulze, 1982). Short-lived intense storms, where the infiltration capacity of the soil is exceeded, are usually responsible for the bulk of seasonal soil loss (e.g. Rydgren, 1996). Soil erodibility Soil erodibility (K) accounts for the resistance of soil to the impacting forces of rainfall and runoff. Therefore, soil erodibility is a measure of the susceptibility of a given soil to particle detachment and transport (Wischmeier and Smith, 1978). The physical, as well as chemical, soil properties and their interactions that affect K-values are many and varied. However, K depends primarily on the structural stability of the soil and on its ability to absorb rainfall (i.e. its infiltration capacity). These properties, in turn, vary according to a number of attributes, including soil texture, organic matter content, carbon content, salinity and pH. Several studies demonstrate that dispersibility is a fundamental soil property to be considered in erodibility analysis (e.g. Rienks et al., 2000; Valentin et al., 2005). Dispersion processes supply fine particles to overland flow for transportation, as well as shift dispersed particles into pores, decreasing infiltration and accelerating runoff. Several authors state the importance of parent material in terms of soil erodibility (e.g. Dardis et al., 1988; Laker, 2004; Le Roux and Sumner, 2011). For example, soils from the Tarkastad and Molteno Formations in northern parts of the Eastern Cape Province are associated with duplex soils (Land Type Survey Staff, 1972-2008) that are highly erodible with widespread gully erosion evident (Le Roux et al., 2010). Furthermore, in the Eastern Cape a distinction can be made between erodible soils developed on the silts and mudstones of the Beaufort Series which give rise to fine textured dispersive soils and the dolerites which give rise to well structured clay soils with a lower erodibility (Garland et al., 2000). Likewise, the quartzitic Table Mountain sandstones weather more slowly, producing coarse textured soils with very little silt and clay, thus accounting for the clear waters and widespread distribution of sand bed rivers in the lowlands of the Western Cape. Botha et al. (1994) cite the importance of the Quaternary Masotchini Formation, a thick layer of transported colluvial sediments as host of most of KwaZulu-Natal’s gully systems, whereas in situ soils are far less erodible. Soil depth is also important since deep soils typically have a higher water-holding capacity, and thus are able to absorb larger rainfall amounts before overland flow is generated (Van Zyl, 2004). Soil erodibility also depends on topographic position and slope steepness. 8 Slope gradient and length The effects of topography on erosion are partitioned in the effects of slope steepness and slope length. Ground slope is important when considering the overall transport of soil particles (Wang et al., 2000). As the slope steepens, the proportion of downslope movement increases as a result of respective increases in velocity and volume of surface runoff. Gentle slopes usually erode less, because there is more surface ponding and slower overland flow which protect the surface against the impact of rain. Runoff and erosion also tend to increase with increasing slope length. Slope length, together with the slope form, determines the severity of erosion through the type of erosion that will dominate, and the probability for deposition (Liu et al., 2000). Soil creep upslope (due to rainsplash erosion) gives way first to interrill erosion (due to broad overland flow) and then to rill erosion further downslope (due to concentrated flow). Therefore, the presence or absence of rill erosion and the density of rills are strongly influenced by slope length (Loch, 1996). According to Van Zyl (2004), the effect of topography also often operates at an essentially local level, erosion being initiated at specific locations on the slope, or in association with minor topographic variations. Renard et al. (1994) state that surface runoff usually concentrates in less than 400 feet (91 m), which is a practical slope-length limit in many situations. Long steep slopes, a common feature in the KwaZulu-Natal Drakensberg, render the land extremely susceptible to erosion once the vegetation cover is degraded (Schulze, 1979). Areas of pronounced relief yield most suspended sediment, including large tracts of the Drakensberg, the former Transkei and Waterberg Plateau (Rooseboom et al., 1992). Vegetation cover and land management Of all hazard factors the cover management code and land use is the most important soil erosion factor which can rapidly change as a result of human activities (Wischmeier and Smith, 1978). Not only does it represent conditions that can be managed to reduce erosion, but it also represents the actual soil erosion risk, which relates to the current risk of erosion under present vegetation and land use conditions. The cover management factor is mainly a function of the canopy cover and residual effect (McPhee and Smithen, 1984). However, the canopy is not always protective against erosion. The effect of the canopy cover on soil loss depends not only on its density but also on its height. If the canopy height is too high (>3 m), the fall velocity of the drip is higher than unintercepted rain. According to McPhee and Smithen (1984), surface cover including mulch and gravel is more effective than equivalent percentages of canopy cover. Surface cover, such as the rocks and residue 9 composition, is considered as one of the most sensitive factors controlling erosion (Renard et al., 1994; Evans, 2000). Various studies across SA report that good vegetation cover protects the soil, even on very steep slopes and under highly erosive rainfall (e.g. Smith et al., 2000; Le Roux et al., 2008a). Boardman et al. (2003) postulate that overgrazing in the first half of the 20th century in the Karoo (Sneeuberg) was the driving factor for gully development. 2.2 Factors influencing gully erosion Gully erosion is a process where surface (or subsurface) water concentrates in narrow flow paths and removes the soil resulting in incised channels that are too large to be destroyed by normal tillage operations (Kirkby and Bracken, 2009). Once initiated, individual gullies can expand into a network of active gullies that contribute significantly to soil loss in a catchment (e.g. Martinez-Casasnovas et al., 2003). Several factors contribute to gully development including topographical variables, parent material-soil associations and land use-cover interactions (Valentin et al., 2005). Rainfall The influence of rainfall erosivity on gullying is not widely researched. Although gullies can be long-term features of the landscape, most research agrees that high magnitude events are required for the initiation of gullies (e.g. Rydgren, 1996; Flügel et al., 2003). Lithological and pedological factors Gully erosion largely depends on soil and lithological characteristics. At the regional scale, several authors note that the inherent erodibility of the parent material is the overriding erosion risk factor (e.g. Watson and Ramokgopa, 1997; Laker, 2004; Le Roux and Sumner, 2011). Several gullies in KwaZulu-Natal developed on transported colluvial and alluvial sediments (Botha et al., 1994). Laker (2004) indicates that in South Africa various mudstones are susceptible to gully erosion mainly due to highly erodible duplex soils derived therefrom. The most prominent feature of duplex soils is a permeable horizon overlying an impermeable one. As a result, water infiltrates and saturates the top layer above the impermeable one where it moves along as subsurface flow causing tunnel erosion (Beckedahl, 1998). In addition, several studies agree that soils prone to tunnel erosion are usually dispersive and easily lose aggregation as a result of high sodium 10 absorption (e.g. Rienks et al., 2000; Valentin et al., 2005). The tunnel network is exposed as gullies where their roofs collapsed. Topographical factors In contradiction to interrill and rill erosion, gully erosion cannot be correlated positively with slope steepness. Several authors agree that gullies in SA are mainly located on gentle slopes with gradients less than 10˚ (Flügel et al., 2003; Kakembo et al., 2009). Tamene et al. (2006) found in Ethiopia that gully erosion is less severe on steep slopes, probably due to steep areas being less accessible and less exposed to human and livestock disturbances. Another possible reason is provided by Poesen et al. (2003), explaining that the so-called critical drainage area needed for gully initiation decreases as slope steepens. Likewise, Kakembo et al. (2009) observed that gullying in several catchments of the Eastern Cape Province predominantly occurs on gentle slopes where the critical drainage area or upslope contributing area is large. Areas with large contributing areas have high flow accumulation (number of upslope cells that flow into each cell) used to identify drainage areas and flow paths vulnerable to gully erosion (Desmet et al., 1999). In addition, areas with large contributing areas are associated with zones of saturation with high surface soil water along drainage paths where slope is low. These saturated areas favour gully formation since the surface soils lose their strength as they become wet. Several studies in South Africa state that gully development is specially favoured in certain terrain units, namely footslopes and valley floors (e.g. Descroix et al., 2008; Kakembo et al., 2009). Gully development is favoured in footslopes and valley floors since they represent areas where overland flow is concentrated into preferred pathways of flow (Beckedahl and Dardis, 1988), especially concave hollows adjacent to drainage lines where soils are deep, as opposed to upland convex hillslope sections (Kakembo et al., 2009). Although soil depth is not a topographical factor per se, it is highly correlated with terrain units usually increasing downslope or towards the lower hillslope elements (Land Type Survey Staff, 1972–2008). Moreover, gully development also depends on the availability of deep soils (e.g. Descroix et al., 2008; Kakembo et al., 2009). Therefore, relatively large fractions of deep soils are affected by gully erosion, especially where footslopes and valleys are subject to overgrazing. Land use and vegetation cover Several studies identify the reduction in vegetation cover as the main driving factor of gully erosion (e.g. Tamene et al., 2006; Descroix et al., 2008). As indicated by examples 11 worldwide (e.g. Boardman and Foster, 2008; Gutiérrez et al., 2009), gully erosion is often triggered and/or accelerated by inappropriate land use. Le Roux et al. (2010) observed that a relatively large portion of cultivated and grassland areas are affected by gully erosion due to livestock disturbance, including overgrazing and trampling along cattle tracks. Several studies mention the importance of paths and tracks in creating gullies (e.g. Garland et al., 2000; Boardman et al., 2003, Hochschild et al., 2003). Clearly, a major problem with gully erosion is the temporal and spatial scale of reporting and the spatial extent to which the phenomenon occurs. 3. Review of current assessment/monitoring techniques The intention of this report is to review methodologies for regional water erosion assessment using satellite remote sensing. The current remote sensing review adopts a similar structure provided by Vrieling (2006), addressing mainly erosion detection of eroded areas and features, followed by the assessment of off-site impacts and erosion controlling factors, and finally data integration for erosion modelling. See Table 1 for satellite and sensor acronyms and names mentioned in text. Table 1: Satellites and sensor acronyms and names mentioned in text Acronym Name ASTER IRS - LISS Landsat MSS Landsat TM JERS-1 SAR SPOT HRV Advanced Spaceborne Thermal Emission and Reflection Radiometer Indian Remote Sensing Satellites - Linear Imaging and Self-Scanning Sensor Landsat Multispectral Scanner Landsat Thematic Mapper Japanese Earth Resources Satellite Synthetic Aperture Radar (Syste`me Pour l’Observation de la Terre High Resolution Visible) 3.1 Eroded areas Until recently, satellite data have been limited to detecting large eroded areas, instead of detecting individual erosion features due to the heterogeneity of the object itself as well as the environment (King et al., 2005; Vrieling, 2006). For example, until recently only large areas suffering extensive gully erosion have been mapped with visual interpretation techniques on optical image composites of different sensors (e.g. Dwivedi et al., 1997a). 12 Some studies separated erosion classes based on a combination of data such as vegetation cover derived from visual interpretation (e.g. Dwivedi and Ramana, 2003) or vegetation and topographic characteristics (e.g. Yuliang and Yun, 2002). An alternative for visual interpretation is direct correlation between erosion and spectral reflectance values, sometimes permitting the detection of erosion and its intensity (Vrieling, 2006). For example, Price (1993) found good correlation between reflectance values of single Landsat TM bands, especially band 4 (near infrared; NIR), and erosion rates for pinyon-juniper woodlands. Pickup and Nelson (1984) successfully distinguished eroding, stable, and depositional areas for arid rangelands in Australia using the data space defined by the 4/6 and 5/6 band ratios of Landsat MSS imagery (corresponding to green/NIR and red/NIR respectively). In addition to abovementioned interpretation techniques is the automatic extraction of eroded areas from imagery (Vrieling, 2006). For example, Servenay and Prat (2003) applied an unsupervised classification algorithm to multispectral SPOT HRV data to distinguish four stages of erosion. Bocco and Valenzuela (1988) separated several erosion classes, whereas Floras and Sgouras (1999) determined one erosion class, using the maximum likelihood classifier after principal component analysis of Landsat TM imagery. Usually, the higher resolution SPOT data performs better in classifying eroded areas, whereas the larger number of spectral bands of Landsat TM resulted in a better classification of land cover and land use (Dwivedi et al., 1997b). Metternicht and Zinck (1998) achieved highest classification accuracy using a combination of images (Landsat TM and JERS-1; SAR data). Change detection methods can also supply direct information on erosion occurrence (Vrieling, 2006). A great variety of methods for change detection from satellite imagery exists (Coppin et al., 2004). These include albedo and spectral image differencing, spectral change vector analysis, repeat-pass SAR interferometry, digital elevation model (DEM) extraction and slight deformation measurements (Rosen et al., 2000; Dhakal et al., 2002). In this context, Vrieling (2006) states that derived interferometric coherence imagery has most potential for erosion detection, even for the assessment of erosion volumes (Smith et al., 2000). However, Vrieling (2006) stressed the need to integrate coherence imagery with additional spatial data, like optical imagery, due to multiple causes of temporal decorrelation. 13 Furthermore, spatial resolutions such as offered by Landsat do not allow gully growth analysis with sequential imagery (Bocco and Valenzuela, 1993). 3.2 Erosion features Until recently, it was difficult to detect erosion features with satellite data due to coarse scale resolution (Hochschild et al., 2003). Most remote sensing techniques have been based on the use of airphoto interpretation and photogrammetry to map erosion features such as gullies (Martinez-Casasnovas, 2003). With the development in sensor technology, spaceborne data with improved spectral, spatial and temporal resolution is now available. New high-resolution satellite imagery such as SPOT 5, IKONOS and Quickbird are very promising for measurement of erosion features, e.g. individual gullies. Le Roux and Sumner (2011) based gully erosion mapping in a large catchment in the Eastern Cape Province on analysis of SPOT 5 imagery from various acquisition dates in 2008. A gully erosion map was created for the catchment by means of manual vectorization at a scale of 1:10 000. SPOT 5 satellite imagery was utilized because the panchromatic sharpened images at 2.5 m resolution provides high resolution air photo-like quality for gully mapping (Taruvinga, 2008) and was acquired from government agencies for the whole country. In order to speed up the processing of data and to exclude subjectivity of manual interpretation, the study first considered different techniques of classification, as well as object-based modelling (i.e. eCognition software) (e.g. Lück, 2002). However, latter mentioned techniques could not express individual gullies with acquired accuracy due to their spectral complexity, especially over such a large area. The spectral reflectance between gullies varies significantly and depends on vegetation cover inside gullies, as well as several soil properties such as the soil organic matter and soil moisture contents. Other regional studies that utilized classification techniques in semi-arid regions of SA confirm this trend, i.e. could not rapidly, nor accurately, define individual gullies from bare soil (e.g. Le Roux et al., 2008b). Taruvinga (2008) demonstrated that classification techniques (i.e. support vector machine applied on SPOT 5 imagery) mainly identifies large (>3.5 ha) prominent gullies (continuous), omitting small intermittent gullies (discontinuous). 14 3.3 Assessment of off-site impacts Most studies that have applied satellite imagery to assess the off-site impacts of erosion focus on the sedimentation of reservoirs and lakes (Vrieling, 2006). For example, sedimentation volumes (water-spread areas at varying depths) have been estimated for Indian reservoirs using simple classification algorithms on multi-temporal IRS LISS-2 and LISS-3 imagery (Jain et al., 2002). Suspended sediment concentration can be assessed using relationships between in situ determined suspended sediment concentration and spectral reflectance (spectral bands between 500 and 800 nm within VNIR range) (e.g. Nellis et al., 1998). However, these relationships are site-specific due to the influence of different sediments, chlorophyll and algae on water reflection (Liu et al., 2003). 3.4 Erosion controlling factors Most remote sensing studies of soil erosion concentrated on the assessment of erosion controlling factors, especially soil erodibility, vegetation attributes and to a lesser extent conservation practices and topography (King et al., 2005; Vrieling, 2006). These inputs are usually integrated for erosion modelling. Soil erodibility Different soils can be classified and mapped with satellite remote sensing on the basis of factors such as soil properties, climate, vegetation, topography and lithology (McBratney et al., 2003; Vrieling, 2006). Soil classification, especially visual interpretation of optical satellite imagery, has been used to assess spatial differences in soil erodibility (e.g. Reusing et al., 2000). For example, Wang et al. (2003) assigned soil erodibility values determined in the field to a soil classification. Erodibility values were then extrapolated to the whole sampling region using geostatistical methods, while Landsat TM band 7 was used to reproduce the spatial variability of the erodibility. Singh et al. (2004) found significant relationships between soil colours defined by the Munsell system and optical satellite imagery. Dwivedi (2001) points out that the soil’s spectral reflectance is influenced by the topsoil characteristics such as soil texture, moisture content, iron oxides and minerals. This, however, can be a limitation for determining one particular property, but useful to classify the surface state such as surface crusting and the uncovering of subsoil (Mathieu et al., 1997; Escel et al., 2004). An increase in radiation relates to a decreasing surface state, whereas a decreasing albedo indicates improved conditions (Ritchie, 2000). A difficulty in measuring 15 topsoil reflectance with satellite data is the influence of vegetation cover, which greatly limits satellite-based soil studies for temperate and humid areas, unless agricultural practices leave the soil bare periodically (Vrieling, 2006). A commonly used technique to separate the soil from the vegetation signal in (semi-)arid environments is linear spectral unmixing (Smith et al., 1990). So called pure components or end-members are those elements that represent the spectral variability of the landscape, e.g. soil and rock outcrops (De Jong et al., 1999). Spectral unmixing requires reference spectra (or end-members) that represent spectrally pure elements. Therefore, this approach is difficult to apply at a regional scale due to spectral variability of different soils and, especially, erosion features. Spectral unmixing allows erosion assessment if the regional soil types are known, as well as their respective climax and degradation forms in combination with their spectral characteristics (Haboudane et al., 2002). Finally, SAR systems can assess soil properties including surface roughness, texture and to a lesser extent soil moisture. For more detail about SAR systems readers should refer to Walker et al. (2004). Topography Nearly all regional erosion models require DEM input for the assessment of slope characteristics. DEMs can be obtained from contour lines, stereo aerial photography, and from satellite data, such as stereo optical imagery provided by SPOT and ASTER (Toutin and Cheng, 2003) or SAR imagery (Toutin and Gray, 2000). Vegetation attributes Until now, most remote sensing studies deal with the measurement of inputs for erosion models, especially for vegetation cover. The reason is remote sensing data have the advantage to account for seasonal vegetation dynamics. Land use classification is often used to map vegetation types to assign USLE C-values reported in literature (Vrieling, 2006). Land use classification has been performed with either optical satellite systems through visual interpretation of image composites (e.g. Khan et al., 2001), or automated classification approaches. Classification accuracy can be further improved by direct linear regression between image bands or ratios and C-values determined in the field (Wang et al., 2002). A specific class of spectral band ratios is known as vegetation indices. Vegetation indices often exploit the fact that green vegetation has high reflectance in the NIR and low 16 reflectance in the red part of the spectrum (Vrieling, 2006). The Normalized Difference Vegetation Index (NDVI) is the most common index which has been used directly as an indication of the protective cover of vegetation (Thiam, 2003) or was related to vegetation cover with regression analysis (Symeonakis and Drake, 2004). The NDVI is sometimes inaccurate due to the effect of soil reflectance and the sensitivity to the vitality of the vegetation. To account for soil reflectance, several soil adjusted vegetation indices (e.g. Transformed Soil Adjusted Vegetation Index; TSAVI) have been developed. Although soil adjusted indices perform better for the assessment of low vegetation covers than NDVI (Flügel et al., 2003; Hochschild et al., 2003), they have difficulty in accounting for spatially variable soil types. One of the most important vegetation vitality effects occur during vegetation senescence when vegetation indices usually decrease even when the cover remains the same. Since senescent vegetation offers the same protection to the soil as vigorous vegetation, it is important to also detect dry vegetation or crop residues (French et al., 2000). Conservation practices Few studies assessed conservation practices with satellite remote sensing, mainly due to the coarse resolution. Tillage practice is the most frequently detected conservation practice by satellite remote sensing because it affects the surface roughness and amount of crop residues (Vrieling, 2006). For tillage detection, remote sensing techniques include classification and visual interpretation (DeGloria et al., 1986), and logistic regression techniques (Van Deventer et al., 1997). Although the authors mentioned above used different band ratios, both included TM band 5 (shortwave infrared; SWIR), which has been related to crop residue. Image timing is very important since tillage operations are performed during a specific time of the year. Cloud cover sometimes restricts acquisitions of optical imagery at the time of tillage, especially in temperate regions (Vrieling, 2006). Therefore, SAR data has been used to assess tillage because the radar return is dependent on surface roughness. McNairn et al. (1998) stress that multi-temporal SAR imagery is required for effective class separation. Moran et al. (2002) stress that integration of SAR and optical imagery provides more information on tillage than separate analysis of both data sources. 17 3.5 Data integration for erosion modelling From the above discussion it is apparent that remote sensing data assist erosion mapping through direct erosion detection or through the use of erosion controlling factors. Erosion models integrate erosion controlling factors, whereas satellite imagery has the potential to provide regional spatial data for several input parameters of erosion models. Until recently, however, most published studies merely use optical satellite data to assess the vegetation component (Vrieling, 2006). The only model which was developed with the intention to be used with satellite data is the Soil Erosion Model for Mediterranean regions, SEMMED (De Jong et al., 1999). It is based on the Morgan et al. (1984) model, but modifications were made to enable the input from satellite imagery to assess the rainfall interception of vegetation at different moments. King et al. (2005) state that the nature of models is changing, with a strong development of physical models which explicitly incorporate spatial generalization. 4. Review of latest spatial inventories/datasets on sheet-rill and gully erosion at a national scale Due to the increasing threat to land and water resources of SA by soil erosion, several spatial inventories have been compiled. Two most recent studies are listed below. 4.1 Gully erosion mapping The most recent spatial inventory includes a gully erosion map for SA created by visual interpretation and vectorization from SPOT 5 satellite imagery (Le Roux et al., 2010; Mararakanye and Le Roux, 2011). SPOT 5 satellite imagery was utilized because the panchromatic sharpened images at 2.5 m resolution provide high resolution air photo-like quality for gully mapping (Taruvinga, 2008) and were acquired from government agencies for the whole of South Africa. As a result, the study successfully mapped gully erosion features and established the spatial extent of the problem at a national level (displayed as gully erosion maps in a GIS; see results in Section 5.1). The study resolved to map gully erosion for the whole country by means of manual vectorization at a scale of 1:10 000. Although the technique is time-consuming, automated mapping techniques could not express individual gullies with the required accuracy due to their spectral complexity over 18 such a large area. The proportion of the provinces and each of their local municipalities that is affected by gully erosion was determined by means of zonal functions in the Spatial Analyst extension of ArcGIS 9.3. The next recommended step is to utilize the gully maps to estimate the degree of erosion in conjunction with collateral classification techniques and the quantification of the influence of factors in gully development. This can mainly be achieved by quantifying the influence of different gully contributing factors by correlating existing gully maps with other available spatial datasets in a GIS (e.g. Geology, Land Types, DEMs, Land Cover etc.). Quantifying the influence of factors in gully development will facilitate the selection of suitable alternatives for preventing or reducing soil loss, especially at municipal and provincial levels. It is worth mentioning here that sheet and rill erosion features could not be interpreted from SPOT 5 imagery with the required accuracy. The reason is that it is not possible to discern where the boundaries of such eroded areas exactly are. Sheet and rill eroded areas are usually ambiguous, especially as observed from SPOT 5 imagery (changing gradually from eroded to bare to vegetated areas). It is postulated that modelling approaches are required in order to express/map areas subjected to sheet and rill erosion at a regional scale. 4.2 Water erosion modelling (emphasizing sheet-rill erosion) Prior to above-mentioned study, a simplification of the Universal Soil Loss Equation (USLE) was applied to the whole country (Le Roux et al., 2008a). The model primarily estimates rainfall erosion and combines sufficient simplicity for application on a national scale with incorporation of the main factors causing soil erosion; emphasizing the sheet and rill aspects of the erosion cycle. Indicators of erosion susceptibility of the physical environment, including climate erosivity, soil erodibility and topography were improved by feeding current available data into advanced algorithms. Actual soil-erosion risk, which relates to the current risk of erosion under contemporary vegetation and land use conditions, was accounted for by regression equations between vegetation cover and MODIS-derived spectral index. Subsequently, a significant update was provided on previous assessments of erosion by inclusion of improved or new national datasets which were not available until recently. 19 5. The extent of sheet-rill and gully erosion in South Africa Results from the studies mentioned above confirm that erosion is a major soil degradation problem in SA, confronting both land and water resource management throughout the country. The results support the general perception that most highly degraded areas have combinations of sheet, rill and gully erosion. 5.1 Extent of gully erosion The results of Le Roux et al. (2010) and Mararakanye and Le Roux (2011) show that all provinces are affected by gully erosion (see Figures 1 and 2). The Northern Cape (160 885 ha) and Eastern Cape (151 759 ha) are the most severely affected, followed by KwaZuluNatal (87 522 ha), Free State (64 674 ha), Limpopo (58 669 ha), Western Cape (25 403 ha), 160,000 140,000 120,000 100,000 80,000 60,000 40,000 20,000 0 G au te ng No rth W es M t pu m al an W ga es te rn Ca pe Li m po po Fr ee St Kw at e aZ ul uNa Ea ta st l er n Ca No pe rth er n C ap e Area affected by gully erosion (ha) Mpumalanga (17 420 ha), North West (10 782 ha) and Gauteng (110 ha). Figure 1: Surface area (ha) of provinces affected by gully erosion 20 Figure 2: Gully location map of South Africa 21 5.2 Extent of sheet-rill erosion The USLE-based results (emphasizing sheet-rill erosion) of Le Roux et al. (2008a) illustrate that areas with high erosion risk occur mostly in the eastern parts of SA (see Figures 3 and 4). Areas are classified as having a moderate to high erosion risk when the average annual soil loss rate exceeds 12 t/ha.yr (Bergsma et al., 1996). In this context, the Eastern Cape (6 188 581 km2) is the most severely affected province, followed by the Free State (5 503 875 km2), Northern Cape (5 407 138 km2), Limpopo (4 422 688 km2), Mpumalanga (2 578 363 km2), Kwazulu Natal (2 138 038 km2), Western Cape (2 065 319 km2), North West (1 485 263 km2), and Gauteng (775 688 km2) provinces. In quantitative terms, the average predicted soil loss rate for SA is 12.6 t/ha.yr. It should be stressed that these results give a broad overview of the general pattern of the relative differences, rather than providing accurate absolute erosion rates. It is also noteworthy that differences between sediment yield and soil loss can be very high (Garland et al., 2000). Research findings of Scott and Schulze (1991) suggest that soil loss within a catchment can be up to five times greater than sediment yield due to the reduction of the total eroded volume by deposition within the catchment. Consequently, a soil-erosion figure of 12.6 t/ha.yr could correspond with a 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 au te ng No rth W W es es t te rn Ca Kw pe aZ ul uNa M ta pu l m al an ga Li m po No po rth er n C ap e Fr ee St Ea at e st er n Ca pe 0 G Area where mean soil loss>12 t/ha.yr (sq km) sediment yield of 2.5 t/ha.yr. Figure 3: Surface area (km2) of provinces where the mean soil loss exceeds 12 t/ha.yr (moderate to high) 22 Figure 4: Water erosion prediction map of South Africa (emphasizing sheet-rill erosion) 23 Compared to Australia, the average predicted soil loss rate for SA is three times as much as that estimated (4.1 t/ha.yr) by Lu et al. (2003). SA has a higher soil loss rate than Australia presumably due to extensive cultivation and overgrazing. A total of 62% of the country is currently under commercial and subsistence farming (National Land Cover, 2000). The areas predicted to be greatly affected by soil loss when compared to the National Land Cover appear to be the degraded unimproved grasslands. Unimproved grasslands are associated with subsistence agriculture where overgrazing of livestock has been excessive. This problem is exacerbated by the inherent erodibility of the parent materials in SA that give rise to erodible soils (Laker, 2004). In terms of management intervention it is especially important to highlight areas that are intrinsically susceptible to erosion (before being extrinsically triggered or accelerated by land use and human-induced reduction of the vegetation cover). 6. Assessment of susceptible areas The assessment of areas that are susceptible to erosion is important in policy terms because it indicates those areas which are inherently susceptible to erosion (potential risk), but which are presently protected by vegetation. For example, due to limited financial resources it will not be feasible to rehabilitate large gullies with large and expensive structures at a catchment scale. In addition to rehabilitating existing gullies, currently vegetated gully-free areas susceptible to gully development should be identified and protected/sustainably managed. Le Roux and Sumner (2011) selected a specific catchment located in the Eastern Cape Province for this purpose (coded as tertiary catchment 35 by the Department of Water Affairs). 6.1 Assessment of areas susceptible to gully erosion by means of GIS and remote sensing: Case study in the Eastern Cape The catchment lies between 30º 46' 58'' and 31º 28' 55'' south and 27º 55' 56'' and 29º 13' 47'' east in the Eastern Cape Province of South Africa, north of Mthatha. The catchment was chosen for its high erosion risk on high potential agricultural land (Le Roux et al., 2008a; b). Soils from the Tarkastad and Molteno Formations in the central part of the catchment are associated with duplex soils (Land Type Survey Staff, 1972-2008) that are 24 highly erodible with widespread gully erosion evident (Le Roux et al. 2008b). Gully erosion mapping was based on analysis of SPOT 5 imagery from various acquisition dates in 2008. SPOT 5 satellite imagery was utilized because the panchromatic sharpened images at 2.5 m resolution provides high resolution air photo-like quality for gully mapping (Taruvinga, 2008) and was acquired from government agencies for the whole of SA (as mentioned above in Section 4.1). The study resolved to map gully erosion for the whole catchment by means of manual vectorization at a scale of 1:10 000. Although the technique is timeconsuming, automated mapping techniques could not express individual gullies with the required accuracy due to their spectral complexity over such a large area. Subsequently, 4 253 gullies were mapped in the catchment, in total affecting an area of approximately 5 273 ha (1.1% of the catchment) (see Figure 5). Figure 5: Gully locations map of the catchment in the Eastern Cape Province In order to determine the influence of factors in gully development, the study correlated the above-mentioned gully map with other spatial datasets in a GIS including DEM-derived topographical variables (i.e. slope, contributing area, a topographical wetness index, terrain units), parent material-soils interactions (i.e. geology maps, land types, soil erodibility maps) 25 and remotely sensed cover management (i.e. land use and vegetation cover). Since not all gully factors can be taken into account at a regional scale, the study considered incorporation of the most important factors for which regional data already existed, or that could be readily derived by means of remote sensing techniques or in a GIS for the whole catchment. Descriptions of the gully contributing factors, methods of derivation and data sources are detailed in Le Roux and Sumner (2011). Each gully factor layer was categorized into 5 expert-based rankings or classes that, according to observations, uniquely influence gully development. Due to the spatially thematic configuration of the gully factor layers it was decided to determine the proportion that each of the 5 classes are affected by gully erosion (by means of zonal functions in the Spatial Analyst extension of ArcGIS 9.3). The study of Le Roux and Sumner (2011) confirms that factors leading to the development of gullies in the catchment are consistent with other studies. Factors leading to the development of gullies are gentle slopes in zones of saturation along drainage paths with a large contributing area, erodible duplex soils derived from mudstones, and poor vegetation cover due to overgrazing. When integrated with drainage networks, gullies expand from valley floors and footslopes onto concave midslopes where the soils are deep. A combination of overgrazing and susceptible mudstones proves to be key factors that consistently determine the development of gullies in the catchment. After quantifying the influence of factors in gully development, the identification of vegetated gully-free areas susceptible to gully development was achieved by means of overlay analysis. Figure 6 illustrates areas that are intrinsically susceptible to gully erosion, yet are vegetated and gully-free (estimated at approximately 7 260 ha). Appropriate strategies need to be designed for these susceptible areas in order to protect the current vegetation cover. This approach proved to be relatively simple, realistic and practical, and it can be applied or expanded to other areas of SA at a regional scale; thereby providing a tool to help with the implementation of plans for soil conservation and sustainable management (Kheir et al., 2007). 26 Figure 6: Areas susceptible to gully erosion in the catchment, Eastern Cape Province 7. Conclusion and recommendations In this report, a brief description of soil erosion impacts and cost implications is followed by a review of mechanisms leading to sheet-rill and gully development including rainfall erosivity, topographical variables, parent material-soil associations and land use-cover interactions. The ability to map erosion features based on satellite imagery is illustrated similar to Vrieling (2006), addressing mainly erosion detection of eroded areas and features, followed by the assessment of off-site impacts and erosion controlling factors, and finally data integration for erosion modelling. The report further describes two most recent spatial inventories, namely a gully location map of South Africa created by visual interpretation and vectorization from SPOT 5 satellite imagery, and a water erosion prediction map of South Africa created by simplification of the Universal Soil Loss Equation (emphasizing sheet-rill erosion). Results from the latter mentioned spatial inventories confirm that erosion is a major soil degradation problem in South Africa. Finally, the report emphasizes the importance of spatially 27 modelling areas that are currently erosion free, but under threat and in need of area-specific management, e.g. protection of the current vegetation cover. The assessment of areas that are susceptible to erosion is important in policy terms because it indicates those areas which are inherently susceptible to erosion (potential risk), but which are presently protected by vegetation. The catchment scale study of Le Roux and Sumner (2011) is described for this purpose. After quantifying the influence of factors in gully development (topographical variables, parent material-soils interactions and remotely sensed cover management), the study identified vegetated gully-free areas susceptible to gully development (estimated at approximately 7 260 ha) by means of overlay analysis. It is recommended that the above-mentioned remote sensing/modelling approach be applied or expanded to other areas of South Africa at a regional scale. With a catchment area of 15 245 km2, the Welbedacht Dam catchment (near Dewetsdorp in the Free State) will constitute a good case study area. Due to siltation, the storage capacity of the Welbedacht Dam near Dewetsdorp in the Free State reduced rapidly from the original 115 million m3 to approximately 16 million m3 within twenty years since completion in 1973 (DWA, 2011). Therefore, the causal factors need to be assessed by means of remote sensing and modelling techniques similar to the study of Le Roux and Sumner (2011). This will allow identification of areas within the Welbedacht Dam catchment and other similar areas that are intrinsically susceptible to erosion (before being extrinsically triggered or accelerated by land use and human-induced reduction of the vegetation cover). Further refinement will be possible given additional research, including investigation of the effect of land use history and vegetation conditions prior to erosion development, using a combination of different optical and multi-temporal data, assessment of sediment yields and distinction between active and passive erosion features. These tools will assist the implementation of a decision support or risk-management system for soil conservation and sustainable management (of problem soils as a geohazard). 28 8. References Beckedahl HR. 1998. Aspects of subsurface erosion phenomena in eastern southern Africa. Peterman’s Geographische Mitteilungen (Supplement) 245: 213. Beckedahl HR, Dardis GF. 1988. The role of artificial drainage in the development of soil pipes and gullies, some examples from Transkei, southern Africa. In: Dardis GF, Moon BP (eds.). Geomorphological studies in southern Africa. Balkema: Rotterdam. pg 229-245. Bergsma E, Charman P, Gibbons F, Hurni H, Moldenhauer WC, Panichapong S. 1996. Terminology for soil erosion and conservation. ISSS, ITC, ISRIC. Boardman J, Foster I. 2008. Badland and gully erosion in the Karoo, South Africa. Journal of Soil and Water Conservation 63(4): 121A-125A. Boardman J, Parsons AJ, Holland R, Holmes PJ and Washington R. 2003. Development of badlands and gullies in the Sneeuberg, Great Karoo, South Africa. Catena 50: 165-184. Bocco G, Valenzuela CR. 1988. Integration of GIS and image processing in soil erosion studies using ILWIS. ITC Journal 4: 309– 319. Bocco G, Valenzuela CR. 1993. Integrating satellite remote sensing and Geographic Information Systems technologies in gully erosion research. Remote Sensing Reviews 7: 233–240. Botha G, Wintle AG, Vogel JC. 1994. Episodic late Quaternary palaeogully erosion in northern Kwazulu-Natal, South Africa. Catena 23: 327-340. Coppin P, Jonckheere I, Nackaerts K, Muys B, Lambin E. 2004. Digital change detection methods in ecosystem monitoring: a review. International Journal of Remote Sensing 25(9): 1565– 1596. Dardis GF, Beckedahl, HR, Bowyer-Bower TAS, Hanvey PM. 1988. Soil erosion forms in southern Africa. In: Dardis GF, Moon BP (eds.). Geomorphological studies in southern Africa, Balkema: Rotterdam. pg 187-213. DeGloria SD, Wall SL, Benson AS, Whiting ML. 1986. Monitoring conservation tillage practices using Landsat multispectral data. Journal of Soil and Water Conservation 41(3): 187– 190. De Jong SM, Paracchini ML, Bertolo F, Folving S, Megier J, De Roo APJ. 1999. Regional assessment of soil erosion using the distributed model SEMMED and remotely sensed data. Catena 37(3–4): 291– 308. Descroix L, González Barrios JL, Viramontes D, Poulenard J, Anaya E, Esteves M, Estrada J. 2008. Gully and sheet erosion on subtropical mountain slopes: Their respective roles and the scale effect. Catena 72: 325–339. Desmet PJJ, Poesen J, Govers G, Vandaele K. 1999. Importance of slope gradient and contributing area for optimal prediction of the initiation and trajectory of ephemeral gullies. Catena 37: 377–392. Dhakal AS, Amada T, Aniya M, Sharma RR. 2002. Detection of areas associated with flood and erosion caused by a heavy rainfall using multitemporal Landsat TM data. Photogrammetric Engineering and Remote Sensing 68(3): 233–239. DWA. 2011. Dams in South Africa: Welbedacht Dam: Free State. Department of Water Affairs: South Africa. http://www.ewisa.co.za/misc/damfswelbedacht. Accessed on 07 February 2011. 29 Dwivedi RS. 2001. Soil resources mapping: a remote sensing perspective. Remote Sensing Reviews 20: 89–122. Dwivedi RS, Kumar AB, Tewari KN. 1997b. The utility of multisensor data for mapping eroded lands. International Journal of Remote Sensing 18(11): 2303–2318. Dwivedi RS, Ramana KV. 2003. The delineation of reclamative groups of ravines in the IndoGangetic alluvial plains using IRS-1D LISS-III data. International Journal of Remote Sensing 24(22): 4347– 4355. Dwivedi RS, Sankar TR, Venkaratnam L, Karale RL, Gawande SP, Rao KVS, Senchaudhary S, Bhaumik KR, Mukharjee KK. 1997a. The inventory and monitoring of eroded lands using remote sensing data. International Journal of Remote Sensing 18(1): 107– 119. Escel G, Levy GJ, Singer MJ. 2004. Spectral reflectance properties of crusted soils under solar illumination. Soil Science Society of America Journal 68(6): 1982–1991. Evans R. 2000. Some soil factors influencing accelerated water erosion of arable land. Progress in Physical Geography 20: 205-215. Floras SA, Sgouras ID. 1999. Use of geoinformation techniques in identifying and mapping areas of erosion in a hilly landscape of central Greece. International Journal of Applied Earth Observation and Geoinformation 1(1): 68–77. Flügel W, Märker M, Moretti S, Rodolfi G, Sidrochuk A. 2003. Integrating geographical information systems, remote sensing, ground truthing and modeling approaches for regional erosion classification of semi-arid catchments in South Africa. Hydrological Processes 17: 929-942. French AN, Schmugge TJ, Kustas WP. 2000. Discrimination of senescent vegetation using thermal emissivity contrast. Remote Sensing of the Environment 74: 249-254. Garland GG, Hoffman MT, Todd S. 2000. Soil degradation. In: Hoffman MT, Todd S, Ntshona Z, Turner S (eds.). A National Review of Land Degradation in South Africa. South African National Biodiversity Institute: Pretoria, South Africa. pp. 69-107. http://www.nbi.ac.za/landdeg. Gibson DJD, Paterson DG, Newby TS, Laker MC, Hoffman MT. 2006. Chapter 4: Land. In: Lickindorf E and Clark R (eds.). South Africa Environment Outlook. A Report on the State of the Environment. Department of Environmental Affairs and Tourism: Pretoria, South Africa. pp. 87-106. Gutiérrez AG, Schnabel S, Felicísimo AM. 2009. Modelling the occurrence of gullies in rangelands of southwest Spain. Earth Surface Processes & Landforms 34: 1894–1902. Haboudane D, Bonn F, Royer A, Sommer S, Mehl W. 2002. Land degradation and erosion risk mapping by fusion of spectrally-based information and digital geomorphometric attributes. International Journal of Remote Sensing 23(18): 3795–3820. Hochschild V, Märker M, Rodolfi G, Staudenrausch H. 2003. Delineation of erosion classes in semiarid southern African grasslands using vegetation indices from optical remote sensing data. Hydrological Processes 17: 917-928. Hoffman MT, Todd S. 2000. A national review of land degradation in South Africa: The influence of biophysical and socio-economic factors. Journal of South African Studies 26(4) 743-758. Jain SK, Singh P, Seth SM. 2002. Assessment of sedimentation in Bhakra Reservoir in the western Himalayan region using remotely sensed data. Hydrological Sciences Journal 47(2): 203– 212. 30 Kakembo V, Xanga WW, Rowntree K. 2009. Topographic thresholds in gully development on the hillslopes of communal areas in Ngqushwa Local Municipality, Eastern Cape, South Africa. Geomorphology 110: 188–194. Khan MA, Gupta VP, Moharana PC. 2001. Watershed prioritization using remote sensing and geographical information system: a case study from Guhiya, India. Journal of Arid Environments 49(3): 465–475. Kheir RB, Wilson J, Deng Y. 2007. Use of terrain variables for mapping gully erosion susceptibility in Lebanon. Earth Surface Processes & Landforms 32: 1770–1782. King C, Baghdadi N, Lemote V, Cerdan O. 2005. The application of remote-sensing data to monitoring and modelling of soil erosion. Catena 62: 79-93. Kirkby MJ, Bracken LJ. 2009. Gully processes and gully dynamics. Earth Surface Processes & Landforms 34: 1841–1851. Kirkby MJ, Imeson AC, Bergkamp G, Cammeraat LH. 1996. Scaling up processes and models from the field plot to the watershed and regional areas. Journal of Soil and Water Conservation 51:391396. Laker MC. 2004. Advances in soil erosion, soil conservation, land suitability evaluation and land use planning research in South Africa. South African Journal of Plant and Soil 21: 345-368. Lal R, Elliot W. 1994. Erodibility and erosivity. In: Lal R (ed.). Soil Erosion Research Methods. St. Lucie Press: Delray Beach, Florida. pg 181-207. Land Type Survey Staff. 1972–2008. Land Types of South Africa: Digital Map (1:250 000 scale) and Soil Inventory Databases. ARC-Institute for Soil, Climate and Water: Pretoria, South Africa. Le Roux JJ, Mashimbye ZE, Weepener HL, Newby TS, Pretorius DJ. 2008b. Erosion status of priority tertiary catchment areas identified by the Soil Protection Strategy of the Department of Agriculture. ISCW Report No. GW/A/2008/17. ARC-Institute for Soil, Climate and Water: Pretoria, South Africa. Le Roux JJ, Morgenthal TL, Malherbe J, Sumner PD, Pretorius DJ. 2008a. Water erosion prediction at a national scale for South Africa. Water SA 34(3): 305-314. Le Roux JJ, Nkambule VT, Mararakanye N, Pretorius DJ. 2010. Provincial mapping of gully erosion at the field scale using high resolution satellite imagery (SPOT 5). ISCW Report No. GW A/2009/04. ARC-Institute for Soil, Climate and Water: Pretoria, South Africa. Le Roux JJ, Sumner PD. 2011. Factors controlling gully development: Comparing continuous and discontinuous gullies. Land Degradation and Development, In press. Liu BY, Nearing MA, Shi PJ, Jia ZW. 2000. Slope length on soil loss for steep slopes. Soil Science Society of America Journal 64: 1759-1763. Liu Y, Islam MA, Gao J. 2003. Quantification of shallow water quality parameters by means of remote sensing. Progress in Physical Geography 27(1): 24–43. Loch RJ. 1996. Using rill/interrill comparisons to infer likely responses of erosion to slope length: implications for land management. Australian Journal of Soil Research 34: 489-502. Lu H, Prosser IP, Moran CJ, Gallant JC, Priestley G, Stevenson JG. 2003. Predicting sheetwash and rill erosion over the Australian continent. Australian Journal of Soil Research 41: 1037-1062. 31 Lück W. 2002. Evaluating the potential for the automation of optical satellite image rectification and classification, MSc thesis. University of Stellenbosch, Department of Geography and Environmental Sciences: Stellenbosch, South Africa. Mararakanye N, Le Roux JJ. 2011. Manual Digitizing of Gully Erosion in South Africa Using High Resolution SPOT 5 Satellite Imagery at 1: 10 000 Scale, Unpublished report. Department of Agriculture, Forestry and Fisheries: Pretoria, South Africa. Martinez-Casasnovas JA. 2003. A spatial information technology approach for the mapping and quantification of gully erosion, Catena 50: 293-308. Martinez-Casasnovas JA, Anton-Fernandez C, Ramos MC. 2003. Sediment production in large gullies of the Mediterranean area (NE Spain) from high-resolution digital elevation models and geographical information systems analysis. Earth Surface Processes and Landforms 28: 443–456. Mathieu R, King C, Bissonnais YL. 1997. Contribution of multitemporal SPOT data to the mapping of a soil erosion index: the case of the loamy plateaux of northern France. Soil Technology 10(2): 99– 110. McBratney AB, Santos MLM, Minasny B. 2003. On digital soil mapping. Geoderma 117(1–2): 3 – 52. McNairn H, Wood D, Gwyn QHJ, Brown RJ, Charbonneau F. 1998. Mapping tillage and crop residue management practices with RADARSAT. Canadian Journal of Remote Sensing 24(1): 28–35. McPhee PJ, Smithen AA. 1984. Application of the USLE in the Republic of South Africa. Agricultural Engineering in South Africa 18: 5-13. Metternicht GI, Zinck JA. 1998. Evaluating the information content of JERS-1 SAR and Landsat TM data for discrimination of soil erosion features. ISPRS Journal of Photogrammetry and Remote Sensing 53(3): 143– 153. Moran MS, Hymer DC, Qi J, Kerr Y. 2002. Comparison of ERS-2 SAR and Landsat TM imagery for monitoring agricultural crop and soil conditions. Remote Sensing of Environment 79(2– 3): 243– 252. Morgan RPC. 1995. Soil Erosion and Conservation (2nd edn.). Longman Group Limited: London, UK. Morgan RPC, Morgan DDV, Finney JJ. 1984. A predictive model for the assessment of erosion risk. Journal of Agricultural Engineering Research 30(3): 245– 253. National Land Cover. 2000. National Land Cover 2000, Unpublished. ARC-Institute for Soil, Climate and Water; Council for Scientific and Industrial research: Pretoria, South Africa. Nellis MD, Harrington JA Jr., Wu J. 1998. Remote sensing of temporal and spatial variations in pool size, suspended sediment, turbidity, and Secchi depth in Tuttle Creek Reservoir, Kansas: 1993. Geomorphology 21(3– 4): 281–293. Pickup G, Nelson J. 1984. Use of Landsat radiance parameters to distinguish soil erosion, stability, and deposition in arid Central Australia. Remote Sensing of Environment 16(3): 195– 209. Poesen J, Nachtergaele J, Verstraeten G, Valentin C. 2003. Gully erosion and environmental change: importance and research needs. Catena 50: 91– 133. Price KP. 1993. Detection of soil erosion within pinyon-juniper woodlands using Thematic Mapper (TM) data. Remote Sensing of Environment 45(3): 233–248. 32 Renard KG, Foster GR, Weesies GA, McCool DK, Yoder DC. 1994. RUSLE users guide. Predicting soil erosion by water: a guide to conservation planning with the Revised Universal Soil Loss Equation, USDA Agriculture Handbook No. 703. USDA: Washington, USA. Reusing M, Schneider T, Ammer U. 2000. Modelling soil loss rates in the Ethiopian Highlands by integration of high resolution MOMS-02/D2-stereo-data in a GIS. International Journal of Remote Sensing 21(9): 1885– 1896. Rienks SM, Botha GA, Hughes JC. 2000. Some physical and chemical properties of sediments exposed in a gully (donga) in northern KwaZulu-Natal, South Africa and their relationship to the erodibility of the colluvial layers. Catena 39: 11–31. Ritchie JC. 2000. Soil erosion. In: Schultz GA. Engman ET (eds.). Remote sensing in hydrology and water management. Springer: Berlin, Germany. Rooseboom A, Verster E, Zietsman HL, Lotriet HH. 1992. The development of the new Sediment Yield Map of South Africa, WRC report no 297/2/92. Water Research Commission: Pretoria, South Africa. Rosen PA, Hensley S, Joughin IR, Li FK, Madsen SN, Rodriguez E, Goldstein RM. 2000. Synthetic aperture radar interferometry. Proceedings of the IEEE 88(3): 333– 382. Rydgren B. 1996. Soil erosion; its measurement, effects and prediction. Case study from the southern Lesotho lowlands. Zeitschrift für Geomorphology 40: 429-445. Schulze RE. 1979. Soil loss in the key area of the Drakensberg – A regional application of the Soil Loss Estimation Model for Southern Africa (SLEMSA): Hydrology and Water Resources of the Drakensberg. Natal and Regional Planning Commision: Pietermaritzburg, South Africa, pg 149-167. Scott DF, Schulze R. 1991. The hydrological effects of a wildfire in a eucalypt afforested catchment. South African Forestry Journal 160: 67-74. Servenay A, Prat C. 2003. Erosion extension of indurated volcanic soils of Mexico by aerial photographs and remote sensing analysis. Geoderma 117(3– 4): 367– 375. Singh D, Herlin I, Berroir JP, Silva EF, Meirelles MS. 2004. An approach to correlate NDVI with soil colour for erosion process using NOAA/AVHRR data. Advances in Space Research 33(3): 328–332. Smith MO, Ustin SL, Adams JB, Gillespie AR. 1990. Vegetation in deserts: I. A regional measure of abundance from multispectral images, Remote Sensing of Environment 31(1): 1– 26. Smith HJ, Van Zyl AJ, Claasens AS, Schoeman JL, Laker MC. 2000. Soil loss modelling in the Lesotho Highlands Water Project catchment areas. South African Geographical Journal 82: 64-69. Smithen AA, Schulze RE. 1982. The spatial distribution in Southern Africa of Rainfall Erosivity for use in the Universal Soil Loss Equation. Water SA 8(2):74-78. Snyman HA. 1999. Soil erosion and conservation. In: Tainton NM (ed.). Veld Management in South Africa. University of Natal Press: Scottsville, South Africa. pp. 355-380. Symeonakis E, Drake N. 2004. Monitoring desertification and land degradation over sub-Saharan Africa. International Journal of Remote Sensing 25(3): 573–592. Tamene L, Park SJ, Dikau R, Vlek PLG. 2006. Analysis of factors determining sediment yield variability in the highlands of northern Ethiopia. Geomorphology 76: 76– 91. 33 Taruvinga K. 2008. Gully mapping using remote sensing: Case study in KwaZulu-Natal, South Africa, MSc thesis (Geography). Waterloo University: Waterloo, Canada. Thiam AK. 2003. The causes and spatial pattern of land degradation risk in southern Mauritania using multitemporal AVHRR-NDVI imagery and field data. Land Degradation & Development 14(1): 133– 142. Toutin T, Cheng P. 2003. Comparison of automated digital elevation model extraction results using along-track ASTER and across-track SPOT stereo images. Optical Engineering 41(9): 2102–2106. Toutin T, Gray L. 2000. State-of-the-art of elevation extraction from satellite SAR data. ISPRS Journal of Photogrammetry and Remote Sensing 55(1): 13– 33. Valentin C, Poesen J, Li Y. 2005. Gully erosion: Impacts, factors and control. Catena 63: 132–153. Van Deventer AP, Ward AD, Gowda PH, Lyon JG. 1997. Using Thematic Mapper data to identify contrasting soil plains and tillage practices. Photogrammetric Engineering and Remote Sensing 63(1): 87– 93. Van Zyl AJ. 2004. Modelling non-point source pollution in agriculture from field to catchment scale: A scoping study –sediments and metal pollution and related activities, WRC report no 2004/08. Water Research Commission by Sigma Beta: Pretoria, South Africa. Vrieling, A. 2006. Satellite remote sensing for water erosion assessment: A review. Catena 65: 2-18. Walker JP, Houser PR, Willgoose GR. 2004. Active microwave remote sensing for soil moisture measurement: a field evaluation using ERS-2. Hydrological Processes 18(11): 1975– 1997. Wang G, Gertner G, Anderson AB. 2000. Spatial prediction and uncertainty analysis of topographic factors for the Revised Universal Soil Loss Equation (RUSLE). Journal of Soil and Water Conservation 55: 374-384. Wang G, Gertner G, Fang S, Anderson AB. 2003. Mapping multiple variables for predicting soil loss by geostatistical methods with TM images and a slope map. Photogrammetric Engineering and Remote Sensing 69(8): 889–898. Wang G, Wente S, Gertner GZ, Anderson A. 2002. Improvement in mapping vegetation cover factor for the universal soil loss equation by geostatistical methods with Landsat Thematic Mapper images. International Journal of Remote Sensing 23(18): 3649– 3667. Watson HK, Ramokgopa R. 1997. Factors influencing the distribution of gully erosion in KwaZuluNatal’s Mfolozi catchment – land reform implications. South African Geographical Journal 79(1): 2734. Wilson JP, Gallant JC. 2000. Terrain analysis, principles and applications. John Wiley & Sons, Inc: New York, USA. Wischmeier WH, Smith DD. 1978. Predicting rainfall erosion losses, a guide to conservation planning. USDA Agricultural Handbook No 537. USDA: USA Yuliang Q, Yun Q. 2002. Fast soil erosion investigation and dynamic analysis in the Loess Plateau of China by using information composite technique. Advances in Space Research 29(1): 85– 88. 34