Contents - Previous - Next


Cause and dynamics of sheet erosion

Sheet erosion is caused by the force of raindrops impacting on bare soil (Ellison 1944) and dislodging particles of earth. This force is dependent on the speed of fall (a function of the length of fall and the wind-speed) and the weight (a function of the diameter of drops). After falling for 10 metres raindrops reach 90% of their final speed, which is determined by the balance between gravity and the air resistance of the bearing surface of the drop (Laws and Parson 1943). Wind can increase the force of raindrops by 20-50% (Lal 1976), but turbulence reduces the diameter of drops to 3-5 mm. The force is often stronger under the crowns of tall trees than on cultivated plots because the drops come together on the leaf sheathes, forming larger drops (Valentin 1981). Raindrop diameter in storms of varying intensity can be observed for each region, resulting in regressions such as: energy of a storm = energy of each segment of rain falling at a given intensity multiplied by the number of millimetres fallen at this intensity. Such proportions (Figure 16) vary considerably from one region to another, and in the absence of regional data on rainfall energy, those of Wischmeier and Smith (1978) can be used.

This impact energy is dissipated in four ways:

compression of the soil under the rain's impact, following rapid moistening of the soil surface;

crushing and shearing stress: separation of aggregated particles;

projection of elementary particles in a crown formation on flat soil and transport in all directions but most effectively downhill on slopes;

noise of the impact of the drops on resistant material.

TABLE 9
Influence of season, maximum intensity in 30 minutes, and rainfall over the previous decade (R 10 days = index of soil moisture) on erosion and runoff from rainfall of similar proportions on bare soil and covered soil (cf. Roose, 1973)

Dates

Rainfall

Runoff (%)

Erosion (kg/ha)


R (mm)

R 10 days (mm)

Max. intensity 30 min

Bare soil

Panicum

Bare soil

Panicum

13.2.72

28

58

33

47

0

548

0

18.3.72

33

1

59

52

0.1

1104

0

27.3.72

32

45

23

26

0

327

0

21.5.72

34

20

28

26

0

1518

0

9.6.72

33

131

35

48

32

3833

21

11.6.72

34

164

26

44

11

2191

26

13.6.72

38

230

37

63

22

3264

31

2.7.72

32

212

43

73

0.1

6025

0.2

31.7.72

30

0

15

9

0

412

0

19.10.72

31

88

14

39

0.1

1501

0.1

23.11.72

28

18

43

71

0

1827

0

This rain energy meets opposition in the cohesion or resistance of soil matter, which may already be to some extent degraded:

by breaking up on contact of drops with dried clods;

by wetting followed by drying, creating small cracked clods;

by compression by tyres or rollers, creating small broken clods;

by dispersion of colloids, either through prolonged wetting or through salinization or the presence of exchangeable sodium.

The resistance of the soil material will depend on the presence of pebbles, the percentage of silt and fine sand (10-100 m), organic matter and clay, the presence of gypsum or limestone, iron hydroxides and free aluminium, and again the structural stability and permeability of the profile (see "Soil erodibility", page 91).

Particles are initially carried a short distance by the splash effect and then by sheet runoff. The impact of raindrops sends droplets and particles in all directions, but on slopes the distance covered uphill is less than that downhill, so that on the whole particles move downhill in jumps. Christoï's experiments (1961) at the Niangoloko IRHO station in southern Burkina Faso showed that soil particles can jump up to 50 cm into the air and travel more than 2 m at a time during heavy storms at the end of the dry season. Sheet runoff starts only after puddles have formed and water that has not infiltrated overflows from one puddle to another. As the runoff spreads over the surface, it moves slowly even on 5-10% slopes because of the roughness of the soil surface (clods, grass, leaves, roots, pebbles, etc.) which keeps the speed below 25 cm/s. Faster than that, runoff will not only carry fine particles but can also attack the soil, digging out stratified channels in which speed quickly builds up, and thus becoming linear erosion (grooves, rills and gullies). See the Hulström curves (Figure 19).

Sedimentation. As raindrops fall, particles or even aggregates (especially if large stormdrops fall on dry clods) will become detached from clods, filling in any hollows and forming sedimentation crusts which allow very little infiltration (Figure 15).

Sheet erosion observed on an erosion plot depends (Table 9) on:

the maximum intensity (I) of the rain that triggers runoff (max I in 15 minutes on steep slopes, or max I in 30 minutes on average slopes);

the energy of the rain (E C) which dislodges particles then easily carried away;

the duration of the rain and/or the soil moisture level before the rain.

Hudson (1965 and 1973) working in Zimbabwe, and Elwell and Stocking (1975) working on well-structured ferralitic soils (oxisols), have found the best relation between erosion and raindrop energy above a certain intensity threshold (I > 25 mm/h) (E = K E [kinetic energy] if I > 25 mm/h). These authors have observed that only intense rain leads to erosion. However, it is likely that any rain will have some ill effect on the soil surface, for even if not all rains produce runoff, they do foster the development of a fairly impermeable crust and accelerate runoff in future storms.

If there really is a rain intensity threshold below which runoff does not occur, it will vary according to the moisture level of the soil and the degradation of its surface before the rain starts (cf. the work of Lafforgue 1977, Raheliarisoa 1986, Casenave and Valentin 1989). Lal (1976) argues that a sudden peak intensity in 7 or 15 minutes is even better correlated with erosion than intensity for 30 minutes. This may be true in certain places (De Noni, Nouvelot and Trujillo 1984 on volcanic soil in Ecuador), but not necessarily everywhere. Roose (1973) has shown that on the sandy ferralitic soils of southern Côte d'Ivoire the longer the rain's maximum intensity lasts, the higher the regression coefficient, while Lal (1976) has shown that wind can increase the energy of raindrops - although it is difficult to take this into account, since both rain intensity and wind intensity rarely exist at the same time.

Wischmeier (1959) combined the kinetic energy of each rainstorm multiplied by the greatest amount of rain in any 30-minute period (mm/h) into a single erosivity index (EI30), which takes full account of the three conditions of rain energy, peak intensity and duration, as described above.

Inasmuch as processing a rain-gauge printout for each rainstorm is a finicky and tedious operation, and not all the necessary information on rain intensity is always available, many authors have tried to simplify the task of estimating the rainfall erosivity index.

In West Africa, Charreau (1970), Delwaulle (1973), and Galabert and Millogo (1973) found a direct relation between kinetic energy and rainfall:

R' = (a + b H) . I30

In Nigeria, Lal proposes:

R' = precipitation in cm × max I 7'.

Roose (1977a) analysed readings from 20 stations between Séfa in Senegal and Deli in Chad and between Abidjan in southern Côte d'Ivoire and Allokoto in Niger, and showed that in West Africa there is a direct relation between the mean annual aggressiveness index and the mean annual rainfall over the same period (over more than ten years).

m a R = m a R ×

0.5 + 0.05 in west Africa,
0.6 on a coastal strip 40 km wide,
0.3-0.2 in mountains in Cameroon (Roose), Rwanda, Burundi and Madagascar (Sarrailh),
0.1 in the Mediterranean area of Algeria (Arab) 1991), less than 0.01 in the oceanic temperate zone.

FIGURE 17 Sketch map showing mean annual climatic aggressiveness (Wischmeier's RUSA) in West and Central Africa: condition of erosion plots (from rainfall data collected by the Hydrology Department of ORSTOM up to 1975)

It is, however, interesting to study the extent of erosion phenomena in relation to different levels of precipitation. During the 1965 season at Adiopodoumé (maize grown on ridges following the slope) there was no runoff for showers of less than 15 mm, nor serious erosion for those of less than 30 mm. At least 30 mm were needed for runoff to occur in any given shower, and more than 90 mm to be certain of sediment transport. On the basis of the nature of the soil, but also plant cover and cropping methods, each plot therefore has its own trigger point below which there is no sign of erosion (Roose 1973). In reality precipitation levels are not totally independent of rainfall intensity - or at least intensity for 30 minutes and more.

Figure 17 is a diagram showing the distribution of the Wischmeier index, R. for West and Central Africa based simply on average isohyets adjusted on the basis of different coefficients. This was possible because there are good correlations in this part of Africa between peak intensity and annual rainfall: 1/10 frequency (Brunet-Moret 1963; 1967; Roose 1977b).

Note the curves: intensity, duration, as functions of rainfall (Figure 18).

In the United States, Wischmeier's index varies from 20 to 650 units. In Europe the index varies from 20 to 150. In the Mediterranean region, RUSA = 50-350 (Tunisia, Morocco, Algeria). In dry tropical zones, RUSA = 100-450, and in humid tropical zones, 500-1200 (Roose 1973).

It should be noted, however, that serious disparities have been observed between sheet erosion in certain regions and rainfall aggressiveness according to Wischmeier's equation. The point is that aggressive rain can take the form of storms at the onset of the rainy season as in West Africa, or summer storms as in Europe, or long showers of fine, drenching rain with little force, falling on soaked soil, at the end of winter or beginning of spring as in France or Algeria. In the latter case, erosion is caused more by runoff energy, hence taking the form of linear erosion, than by the energy of the raindrops themselves. (It may also develop into massive earth movements if the slope is steep enough.)

For watersheds of over 2000 km², Fournier showed in 1960 that sediment transport was essentially dependent on two factors: topography and aggressiveness, or what he called the "indice de continentalité" (rainfall continentality index) (c). This index is equal to the relation between the square of the rainfall in the wettest month divided by the mean annual rainfall. In this form it applies only to sediment transport in large watersheds, and cannot be applied directly to sheet erosion on plots, which depends too much on plant cover and tilling techniques. However, attempts have been made to estimate Wischmeier's index of rain aggressiveness, working from the total of Fournier's monthly indices, and good regional correlations have emerged between Wischmeier's index and Fournier's modified, monthly index (Arnoldus 1980).

FIGURE 18 Intensity/duration curves for storms of known frequency at each point in West Africa (cf. Brunet-Moret 1967)


Contents - Previous - Next