Contents - Previous - Next

Soil erodibility

The erodibility of a soil [Plate 2] as a material with a greater or lesser degree of coherence is defined by its resistance to two energy sources: the impact of raindrops on the soil surface, and the shearing action of runoff between clods in grooves or rills. The first studies on the erodibility of materials were done by Hjulström in canals (Figure 19). Hjulström's diagram shows that there are three sectors, depending on water velocity and the diameter of soil particles. Analysis of the erosion sector shows that the diameter of the particles of the most fragile matter is about 100 microns, i.e. fine sand. With finer matter, cohesion develops simply as the surfaces of the clays rub together, while coarser clumps become increasingly heavy and therefore harder to transport. This kind of trial is concerned with resistance to the erosive force of river or runoff in a wet environment.

Soil scientists have long realized that soils react at varying speeds to raindrop attack and structural degradation. A whole series of laboratory and field tests has been set up to try to define structural stability with respect to water - for example, Ellison's capsules (1944) where sifted aggregates are exposed to raindrop energy, Hénin's structural stability test (Hénin and Monnier 1956) where aggregates are submerged and sifted under water, the waterdrop test where graded clods (30 gr) are exposed to drops of water falling from a specific height (McCalla 1944) or Middleton's dispersion test (1930) which seeks to compare the content of particles naturally dispersed in water with and without dispersant.

FIGURE 19 Hjulström's diagram

This diagram provides some very important information.

1. The material most easily dislodged by runoff has a texture close to that of fine (100 m) sand. More clayey material is stickier. The coarser material has heavy particles which can only be moved at higher fluid speed. It is interesting to note that for Wischmeier the most erodible soils are those rich in loam and fine sand.

2. As long as the flow is slow (25 cm/see), it cannot erode. Measures will therefore have to be taken to spread and slow down the flow, in order to prevent linear erosion. This is the basis of the theory of dissipation of runoff energy.

3. Fine clay and loam particles are easily transported, even at low speeds, but in the case of anything coarser than fine sand, it is a short distance from erosion site to sedimentation site. This explains why ditches to channel runoff water either erode if they are too narrow or steep, or silt up with coarse material which cannot be moved. This is one of the reasons why diversion ditches are unpopular in developing countries, for such ditches and channel terraces have to be regularly cleared and maintained.

Quantin and Combeau's work (1962) on ten erosion plots at Grimari in the Central African Republic showed that a higher Hénin's instability index will also increase both erosion and the average sediment load, and that the products carried away are finer.

These authors noted that the instability index, I_S, varies according to season, plant cover and time elapsed since clearing. Tropical soils would thus be less sensitive towards the end of the dry season when plant cover decreases, and more sensitive on old cleared land - as has been tested under cotton in northern Cameroon (Boli, Bep and Roose 1991).

In order to approximate natural conditions, many authors have taken populations of aggregates from tilled horizons and exposed them to simulated rainfall in the laboratory in order to classify soils according to their resistance to erosion (Madagascar, Zimbabwe, Canada, United States, and many others). In a comparative study, Bryan (1981) showed that soils were classified differently according to type of simulator and procedures used in experiments. Recently, Le Bissonnais (1988) has clearly shown that in fact different processes of soil degradation were involved, and that these were dependent on the different programmes of simulated rainfall.

Many tests have been carried out in the field, under simulated rain. For example, Swanson and Dedrick (1965) in the United States, Dumas (1965) in Tunisia, Pontanier et al. in Cameroon (1984) and Tunisia, Lelong, Roose and Darthout (1992), Masson (1992) and Gril (1982) in France, Roose and Asseline (1978), Collinet and Valentin (1979 and 1984), Valentin and Janeau (1989) in West Africa, and Delhoume et al. (1989) in Mexico. Working on calcareous soils on 50 m² plots in Tunisia, Dumas showed that soil erodibility depends on the amount of pebbles, the amount of organic matter, and the equivalent humidity of the soil, which depends in turn on its texture (Figure 20). From this figure, it can be seen that in the case of Mediterranean calcareous soil an increase of 1% in the amount of organic matter reduces soil erodibility by a mere 5%, whereas a 10% presence of pebbles in the surface horizon will reduce its erodibility by over 15%. When the percentage of pebbles exceeds 40%, there is a decrease in the reduction in soil erodibility. In the young calcareous Mediterranean countryside, the percentage of pebbles is therefore a sign of good resistance to soil erosion.

In the United States Wischmeier and Smith have defined the standard bare reference plot as having a 9% slope, 22.2 m in length, cultivated in the direction of the slope, and having had no organic matter ploughed in for three years. On these reference plots, under both natural and simulated rainfall, Wischmeier and his colleagues calculated multiple regressions between soil erodibility and 23 different soil parameters. Simplifying, it turns out that erodibility depends essentially on the amount of organic matter in the soil, the texture of the soil, especially sand of 100-2000 microns and silt of 2-100 microns, and lastly the profile, the structure of the surface horizon and permeability (Figure 21). Several years later, Singer, Blackard and Janitsky (1978) showed that some supplementary factors have to be added in the case of Californian soil, in particular iron and free aluminium, the type of clay and the salinity of the matter. Today, if the texture of the surface horizons, their level of organic matter, iron and free aluminium are known, and the type of clay, plus some observations on the profile, an initial estimate of the soil's resistance to sheet and rill erosion can be given.

FIGURE 20 Erodibility of Tunisian soils under a rain simulator (cf. Dumas 1965)

log 100 K = 3.4623 X₁ - 0.1695 X₂ - 0.0214 X₃ - 0.0282 X1 Y² = 88%
where =

X₁ = amount of gravel in % of weight in first 5 cm
X₂ = organic matter
X₃ = equivalent moisture

Since the highest-level soil classifications do not take these parameters into account, there is no clear relation between erodibility and currently recognized soil types. However, the K erodibility index varies in the United States between 0.7 for the most fragile soils, 0.3 for brown leached soils, and 0.02 for the most resistant soils. In Africa, scientists (Roose 1980, Roose and Sarrailh 1989) have found values from 0.12 for ferralitic soils on granite, 0.2 for ferralitic soils on schist, and up to 0.4 if the ferralitic soils are covered by volcanic deposits. They found 0.2-0.3 on tropical ferruginous soils, 0.01-0.1 on vertisols, and 0.01-0.05 on soils which were gravelly even on the surface. Overall measurements using a rain simulator, even on 50 m² plots, give lower readings than long-term measurements on plots under natural rainfall, since rills develop more easily on the latter. In reality there is no one erodibility index per soil type, for the index changes over time according to the soil's moisture and roughness, plant cover, slope, and soil organic matter content.

FIGURE 21
Nomograph allowing a quick assessment of the "K" factor of soil erodibility (cf. Wischmeier, Johnson and Cross 1971)

Procedure: in examining the analysis of appropriate surface samples, enter on the left of the graph and plot the percentage of silt (0.002 to 0.1 mm), then of sand (0.10 to 2 mm), then of organic matter, structure and permeability in the direction indicated by the arrows. Interpolate between the drawn curves if necessary. The broken arrowed line indicates the procedure for a sample having 65% silt + very fine sans, 5% sand, 2.8% organic matter, 2 of structure and 4 of permeability. Erodibility factor K = 0,31.

Figure

TABLE 10
Effect of slope on runoff (KR %) and erosion t/ha/yr at Séfa, Senegal: root and tuber crops 1955-1962, tropical ferruginous soil leached in patches and concretions (cf. Roose 1967)

In conclusion, it is clear that the methodological problem of estimating soil resistance to erosion and the way this resistance develops is still awaiting solution. At present, attempts are being made to classify soils according to a variety of tests based on the different processes that may be met in different circumstances. Valentin (1979) has shown that Hénin's index of structural instability bore a good relation to soil resistance to erosion if the drops fall on dry soil, i.e. at the start of the rainy season (C = breaking of aggregates), whereas on wet soil at the end of the rainy season there are better correlations between soil loss and Atterberg's liquidity limits. De Ploey (1971) developed a similar index for brown leached soils in Europe. Infiltration capacity and the resistance of the material to gullying (the shearing force) must also be evaluated in cases where the soil is very prone to runoff (see the ORSTOM Soil Science Journal, 1989, no. 1, special issue devoted to soil erodibility).

Contents - Previous - Next