Erosion Assessment

Erosion assessment may describe the “actual” risk of erosion: the Global Assessment of Soil Degradation (GLASOD) project is an example of this type of assessment applied globally.111 Similarly, Canada has published maps of the risk of wind and water erosion at the provincial level.1271 New Zealand has combined, on single maps at 1:250,000 scale, the actual and “potential” risk of erosion.1281 Potential in this case refers to a change of land use to pastoral or arable. Maps from the CORINE project show actual and potential soil erosion in southern Europe; the latter is defined as the worst possible situation that might be reached.1291 The National Soil Maps for England and Wales show similar soils grouped as “soil associations.”1301 The risk of soil erosion on each soil association has been assessed by Evans based on empirical evidence.1311

Erosion assessment may focus on changes through time in the same area. Trimble1321 has measured and modeled changes in erosion and deposition rates in the Coon Creek catchment in Wisconsin since the beginning of European settler farming in the 1850s. This study relates erosion to land use change, management, and conservation techniques, as well as the capacity of rivers to transport eroded materials. It therefore links on-site and downstream effects through time and space. Erosion rates since Neolithic clearance have been modeled on the loessic soils of the South Downs, UK. This assessment is based on changing land use, climate, soils, and farming practices. It is noteworthy that most erosion occurred during periods of intensive farming during the Bronze Age and Iron Age 4000-2000 BP (Figure 28.4).1331 Assessment of erosion and its impacts may be combined in complex scenarios of change involving population pressure, e.g., Hurni for Ethiopia.1341

Modeling constitutes an important approach to erosion assessment. This is for several reasons: many areas lack the data that allow direct estimates of erosion; modeling may be based on scenarios of future land use or climate, and therefore, estimates may be made of erosion under assumed conditions; and application of a model may be far less expensive and time-consuming than field surveys.

The spatial scale of modeling may vary from a few square meters to thousands of square kilometers. The global change and terrestrial ecosystems (GCTEs) model evaluation exercise divides models into field, catchment, and landscape, based on spatial scale.1351 Resolution of erosion models is determined by grid size, that is, km2 grid size used in CORINE for southern Europe and the 250 m grid used in the maps of seasonal erosion risk in France.

Simulated annual erosion rates and surface stone content for a “thin” soil profile (1.22 m depth to chalk) on the South Downs, southern England. (Adapted from Favis-Mortlock et all)

FIGURE 28.4 Simulated annual erosion rates and surface stone content for a “thin” soil profile (1.22 m depth to chalk) on the South Downs, southern England. (Adapted from Favis-Mortlock et all331)

Assessments are frequently expressed in map form.1361 Landscape features indicative of erosion may be mapped as a record of an actual event or series of events. In arable landscapes features may be temporary; thus, rills, ephemeral gullies, and fans are usually plowed out annually, e.g., erosional features produced as a result of a snowmelt event in May 1993 in Slovakia.1371 The ephemeral gully in Figure 28.5 was plowed out a few weeks later. In semiarid, grazed landscapes erosional features may become permanent, e.g., gullies and badlands. Assessments should distinguish features active at the present time from those that are inactive and represent erosion under former conditions.

Several methods of erosion assessment are commonly used as shown in Table 28.11381 and will be reviewed briefly.

  • 1. Small experimental plots in the laboratory or in the field are typically 22 m x 2 m in size. Experiments on plots may be conducted with natural or simulated rainfall1391 (Figures 28.6 and 28.7). Plots are assumed to represent larger parts of the landscape. Results are frequently extrapolated to fields or catchments, and are used as input to models. Extrapolation from small plots to regional, national, or continental size areas is not acceptable.1401 Plots are also unsatisfactory in that the key erosion process of gullying is not reproducible at this scale. Plots are best used to investigate small scale processes such as rill/interrill relations1411 or rill/land use relations.1421
  • 2. Sediment yield of rivers represents the net loss of soil from a catchment.1431 Erosion rates on fields may be very different owing to storage of soils between the field and the river. The highest rates of sediment yields from a large river are from the Yellow River, China, with an average silt content

TABLE 28.1 Assessment Methods: Their Merits and Limitations

Assessment Method



Experimental plots in lab or field

Control over erosion factors; ease of measurement of runoff and soil

Small size: difficult to extrapolate to larger areas

Sediment yield of rivers

Ease of measurement; may cover >100 years if reservoir sediments used

Measures total soil lost from watershed, not from specific areas

Field monitoring

Records soil lost from rills and gullies; covers large areas; inexpensive

Gives very approximate amounts of soil lost

Remote sensing

Covers large areas and may cover long time-span

Limited availability for many areas; may not be suitable for small scale features


Gives 65-year average

Many unknowns: unreliable without careful calibration; expensive and time-consuming

Stratigraphy and pedology

Deposited soil may contain artifacts; suitable for long-time-period studies

Scarcity of suitable sites; low precision of some dating methods

Expert opinion

May reveal unrecorded information

Subjectivity: difficult to compare different areas


Objectivity; ease of use even by “non-experts”; numerical output

Availability of data; complexity of process descriptions; frequent lack of validation

Sediment source

Assessment of source of eroded

Analytical time and expense; possible


material, not available by other methods

ambiguous results

Experimental plots, South Limbourg, the Netherlands (J. Boardman)

FIGURE 28.7 Experimental plots, South Limbourg, the Netherlands (J. Boardman).

of 38 kg m3.l12l Rivers of east and south Asia deliver about 67% of the sediment reaching the world’s oceans; rates are the highest in this region owing to tectonic uplift and steep slopes, high rainfall, deforestation, and population pressure resulting in intensive farming activities.!44! Alternatively, the sediment yield from river basins may be estimated by measurement of the volume of sediment trapped in a still water body such as a lake, reservoir, or pond. A date for the construction of a reservoir is often known. Sediment including bedload will be trapped. Losses in water passing through the dam may be estimated (trap efficiency). The advantage of this method is the possibility of medium-term records of >100years.l45I Sediment yield to three small dams in the Karoo, South Africa, is estimated at 115, 357 and 6541 knr2 yr1 for a period of 65years. These rates are adjusted to consider trap efficiency, and the contrasts are due to differences in land use within the catchments.!46! Similar approaches have been used in Australia!47! and in Ethiopia.!48! Estimates of soil losses for the same river basin derived from river yields and reservoir sedimentation may differ substantially probably because of the influence of basin scale on sediment yield.!49!

3. Field monitoring is defined as “field-based measurement of erosional and/or depositional forms over a significant area (e.g., >10km2) and for a period of >2years.”!5°l Regular measurements of volumes of soil loss from rills and gullies or deposited in fans are made. Notable examples are the monitoring of 17 localities in England and Wales during 1982-1986,151! a 10-year program on the South Downs, England, during 1982-1991,!52l monitoring of ephemeral gullies in Belgium during 1982-1993,153! and a long-term monitoring scheme in the Swiss midlands.!54-55!

Simple techniques of measurement and assessment of erosion suitable for Less-Developed Countries include measurement of pedestal height, depth of armor layer, plant/ tree root and fence post-exposure, tree mounds, and buildup of sediment behind barriers and in drains.!56! Many studies have used erosion pins to assess rates of erosion over time (Figure 28.8). These are particularly suited to bare degraded lands such as badlands. Rates may be related to topographic position, presence of vegetation, weathering, and rainfall/wind factors. Measurement errors should be estimated and taken into account!57! (Figure 28.9).

4. Remote sensing is used increasingly in erosion assessment. For rill and gully erosion, it may be used as a locational technique to allow detailed field measurements to be made at selected sites. Gullies and badlands may be identified on black and white air photographs of 1:20,000 scale. In the Karoo of South Africa, comparisons have been made between the extent of gullying in 1945 and 1980.I58! Also in South Africa, in a pioneering study, Talbot mapped gullying in the Swartberg resulting from land conversion to wheat farming!59! (Figure 28.10). More recently, the extent of Icelandic land degradation has been assessed using Landsat imagery.!60! The value of Google Earth images for mapping erosional and degradational features should not be overlooked. The main disadvantage is the irregular, rather random time sequencing of the images.!61!

Erosion pin site in the Karoo, South Africa (J. Boardman)

FIGURE 28.8 Erosion pin site in the Karoo, South Africa (J. Boardman).

Average rates of erosion for eroding pins at two sites in the Karoo, South Africa, plotted against rainfall amount for each measurement period (J. Boardman)

FIGURE 28.9 Average rates of erosion for eroding pins at two sites in the Karoo, South Africa, plotted against rainfall amount for each measurement period (J. Boardman).

Cesium-137, derived from weapons testing since 1954, is used as a tracer to assess amounts of erosion and deposition. Annual average rates for the last ca. 65years are estimated by measurement of the amount and distribution of cesium in comparison with an uneroded reference site.1621 Studies have been carried out in Australia, Canada, China, the Netherlands, Poland, Thailand, the United Kingdom, and the United States. However, results are controversial with issues regarding the original assumed regular distribution of cesium across the landscape via the assumed homogenous spatial distribution of rainfall, and the assumption that cesium “sticks” to fine soil particles in a predictable way. These assumptions have recently been strongly challenged, and erosion rates obtained using this technique now appear unreliable.1631 Comparison of field-measured erosion data with cesium-derived predictions shows gross over-estimation using the latter method.1641

Gullied hill country ca. 8 km northwest of Durbanville, South Africa. Gullies are on ploughed land. Photography March 1938

FIGURE 28.10 Gullied hill country ca. 8 km northwest of Durbanville, South Africa. Gullies are on ploughed land. Photography March 1938: the land had been abandoned by 1944. (Adapted from Talbot.1551)

  • 6. Total amounts of soil loss since a historical baseline such as woodland clearance or the beginning of European settlement have been assessed by comparison of existing soil profiles with uneroded ones. Evans estimates historical losses of topsoil of 150-250 mm in lowland England and Wales, much of which is now stored as alluvium and colluvium.1561 Sedimentation above marker horizons in reservoirs may be used to compute sediment yield for periods of time.1461 Pedological and stratigraphic approaches have also been used to assess the effect of past extreme events, for example, the major storms and floods in Germany in the early 14th century.1651
  • 7. Expert opinion has been used to assess erosion on a regional or global scale. GLASOD assesses erosion at a global scale (10 M).[I1 A total of 12 degradation types are recognized under the broad headings of water erosion, wind erosion, chemical deterioration, and physical deterioration. The “degree” of degradation is assessed for each mapping unit, as is its “relative extent.” A combination of degree and relative extent gives an assessment of “severity” which is expressed in “severity classes” and mapped in four colors. However, there is evidence that GLASOD is a very imperfect predictive tool. Van-Camp et al.|66) are particularly critical and suggest that the GLASOD approach should be abandoned: comparison with maps and data for Spain show that the assessment of water erosion was poor.1671 GLASOD has also been assessed by other authors.1681 Comparison of GLASOD with other approaches to assessing the area of global degraded land gives a range of under 1 billion ha to over 6 billion ha, depended on the approach adopted.1691 At a more local and detailed level, expert opinion appears more reliable: see, for example, a recent South African assessment.170!
  • 8. Assessment of erosion has frequently been based on the application of models. The Universal Soil Loss Equation (USLE) is most widely used because of its simplicity and low-level data requirements.1711 Recent developments in modeling have concentrated on the simulation of erosional processes and computerization. Models may be used to estimate average, long-term erosion rates under specified conditions (USLE) or to assess the effect of a particular rainfall/ runoff event, e.g., EUROSEM.1721 Erosion models have also been used to predict wind erosion, and to assess chemical losses in solution or attached to soil particles (for reviews, see Morgan).1751 However, problems with models are widely recognized, for example, because of lack of validation at the field scale,!741 inappropriate calibration,1751 and the quality of field data.1761 Regional mapping using the Revised Universal Soil Loss Equation (RUSLE) is shown to give unreliable results.177-781

9. Sediment source fingerprinting is used in investigating, for example, the source of sediments in a lake: are they from hilislope, topsoil, subsoil, or river bank sources?1791 The method depends on a variety of distinct sources within a catchment that can be matched with the lake sample. Ongoing geomorphological controversies such as the contribution of river bank sediment to total erosion in a catchment may be investigated using fingerprinting.1801

< Prev   CONTENTS   Source   Next >