- Email: [email protected]
Soil mineralogy effects on seal formation, runoff and soil loss
Applied Clay Science 24 (2004) 209 – 224 www.elsevier.com/locate/clay
Soil mineralogy effects on seal formation, runoff and soil loss M. Lado, M. Ben-Hur * Institute of Soil, Water and Environmental Sciences, the Volcani Center, Agricultural Research Organization, P.O. Box 6, Bet Dagan 50250, Israel Received 13 January 2003; received in revised form 13 March 2003; accepted 24 March 2003
Abstract Interaction between clay particles is one of the main factors responsible for soil aggregation. Therefore, soil mineralogy, which has substantial effects on clay dispersion, may also influence aggregate stability, seal formation, runoff and soil loss. In spite of this, the effects of soil mineralogy on these phenomena have received much less attention in the literature than those of other soil properties. This paper reviews the effects of soil mineralogy on aggregate stability, seal formation and micromorphology, and on the associated infiltration rate (IR), runoff and soil loss. The paper is focused mainly on the authors’ previous studies. In order to determine these effects, various soils with different mineralogy were collected from Israel, South Africa and Kenya. In these soils, the aggregate stability was determined under fast wetting conditions, and the IR and the interrill soil loss were determined using rainfall simulator of a rotating disk type. The micromorphology of the crust, which was developed at the soil surface by the rainstorm, was determined by scanning electron microscope. Clay mineralogy was found to be a dominant factor in controlling aggregate stability, seal formation, soil IR and interrill soil loss. The phyllosilicate soils, which were review in this paper, were divided into two main groups: (i) stable soils with final IR>8.0 mm h 1 and (ii) unstable soils with final IR < 4.5 mm h 1. These two soil groups differ in their mineralogy. Kaolinitic and illitic soils that do not contain smectite were stable soils and less susceptible to seal formation. In contrast, kaolinitic and illitic soils that contain some smectite and smectitic soils were unstable. Examination of the susceptibility of 21 phyllosilicate soils to interrill erosion indicated that these soils could be divided into three groups. Soil loss was higher for unstable soils than for stable soils, but the soil loss of the smectitic soils was significantly higher than that of the unstable soils, which contained kaolinite or illite as the dominant clay. D 2003 Elsevier B.V. All rights reserved. Keywords: Aggregate stability; Clay dispersion; Infiltration rate; Soil loss; Soil mineralogy; Surface seal
1. Introduction Runoff and soil erosion are widespread land degradation problems in many parts of the world, because of their contributions to water and soil fertility losses, on the one hand, and their enhancement of flooding
* Corresponding author. E-mail address: [email protected] (M. Ben-Hur). 0169-1317/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.clay.2003.03.002
and surface water pollution risks, on the other hand. Runoff occurs when the rainfall intensity exceeds the soil infiltration rate (IR) and the soil surface water holding capacity, therefore, a decrease in the soil IR could increase the runoff. The main factor that decreases the soil IR under rainfall conditions in arid and semi-arid regions is seal formation at the soil surface (Morin et al., 1981; BenHur and Letey, 1989). A surface seal is thin and is characterized by greater density and lower saturated
210
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
hydraulic conductivity than the underlying soil (McIntyre, 1958; Gal et al., 1984). McIntyre (1958) found that a seal consists of two distinct parts: an upper skin seal attributed to compaction by raindrop impact and a ‘‘washed in’’ zone with decreased porosity, attributed to the accumulation of dispersed clay particles. Likewise, Agassi et al. (1981) suggested that the seal formation is a result of two complementary mechanisms: (i) a physical disintegration of surface soil aggregates caused mainly by the impact energy of the raindrops and leading to formation of the upper skin layer; and (ii) the physicochemical dispersion of soil clay particles, which migrate into the soil with the infiltrating water and clog the pores immediately beneath the surface to form the ‘‘washed in’’ zone. Soil erosion processes can be divided into two components: rill and interrill erosion. Runoff from a soil surface may concentrate into small, erodible channels known as rills. In rill erosion, soil loss is due mainly to detachment of soil particles by flowing water, whereas in interrill erosion, soil particle detachment is caused essentially by raindrop impact, and the particles are transported by raindrop splash and runoff flow (Watson and Laflen, 1986). The present paper deals with interrill erosion only. The tendency of a soil to form a seal, and the resulting amounts of runoff and soil loss depend on soil properties. The soil properties that have received the most attention in this connection are: texture (e.g., Ben-Hur et al., 1985; Romkens et al., 1995; Sharma et al., 1995), organic matter content (e.g., Ekwe, 1991; Fullen, 1991; Guerra, 1994; Le Bissonnais and Arrouays, 1997) and soil sodicity (e.g., Shainberg and Letey, 1984; Ben-Hur et al., 1998). However, soil mineralogy has substantial effects on aggregate stability and clay dispersion (Singer, 1994) and, therefore, may also have effects on seal formation, runoff and soil loss. In spite of this, the effects of soil mineralogy on these phenomena have received much less attention in the literature than the other properties and many studies (e.g., Bradford and Foster, 1996; Meyer, 1981; Watson and Laflen, 1986; Liebenow et al., 1990; Kinnel and Cummings, 1993; Sharma et al., 1995) addressed the runoff and soil loss from various soils under rainfall conditions, without regard to the effects of soil mineralogy on these phenomena. On the basis of mineralogy, soils can be divided into two main groups (Neaman et al., 2000): (i)
phyllosilicate soils that consist mainly of phyllosilicate minerals, of which kaolinite, illite and montmorillonite are the most abundant clays that occur in this mineral group (Allen and Hajek, 1989); (ii) nonphyllosilicate soils that consist of clay-sized minerals, such as quartz and feldspar. Kaolinite is a 1:1 phyllosilicate, in which the basic layer consists of a tetrahedral and an octahedral sheets. Adjacent layers are joined by electrostatic forces between the oxygen ions of the tetrahedral sheet and the hydroxyls of the octahedral sheet in the adjacent layer, to form a tactoid of blocky shape having a typical size of < 2 mm (Dixon, 1989). The bonds between the layers in the tactoid hold them close together, so that water molecules and ions cannot penetrate between the layers. Isomorphic substitution of Si4 + with Al3 + in the tetrahedral sheet, and of Al3 + with Mg2 + or Fe2 + in the octahedral sheet results in a negative charge on the planar surface (van Olphen, 1977). However, because there are few isomorphic substitutions in kaolinite, the electrical charge at the planar surface is weak. There may be an electrical charge at the broken edges of the clay layers; it depends on the pH of the surrounding solution and is positive at neutral pH. Illite and smectite are 2:1 phyllosilicate minerals, in which the basic layer consists of two tetrahedral sheets with one octahedral sheet between them (Schulze, 1989). The main difference between illite and smectite is in the isomorphic substitutions. In illite, most of the substitutions take place in the tetrahedral sheet, where the Si4 + is substituted by Al3 +, leading to a negative charge on the planar surface. This negative charge is balanced mainly by K+ that occupies the interlayer space between two adjacent layers in the tactoid. In this case in general, water molecules and electrolyte cannot penetrate into the interlayer space, and only the external surface of the tactoid is available for interaction with the surrounding solution and other clay particles (Schulze, 1989). In the smectite group, of which montmorillonite is one of the most common clays (Allen and Hajek, 1989), the isomorphic substitutions occur in both the octahedral and the tetrahedral sheets. In this case, water molecules and exchangeable cations can penetrate between adjacent layers. As in the case of kaolinite, there is an electrical charge at the broken edges of the smectite and illite layers, and it depends
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
on the pH of the surrounding solution. The differing morphologies of smectite, illite and kaolinite lead to differing packing of their tactoids, so that they form clay aggregates that differ in their stability (Fig. 1).
211
These clay aggregates are the essential components of soil structure. This paper reviews the effects of soil mineralogy on seal formation and its micromorphology, and the
Fig. 1. (A) Schematic structure of clay aggregates (after Oades and Waters, 1991). (B) Scanning electron micrographs of clay tactoids (after Morin, 1993).
212
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
associated infiltration, runoff and soil loss, and it is focused mainly on the authors’ previous studies. In order to determine these effects, various soils with different mineralogy were collected from Israel, South Africa and Kenya. The general properties of these soils are presented in Table 1, when their ESP values were < 2.5%. These soils were subjected to fastwetting aggregate stability test proposed by Le Bissonnais (1996) and to simulated rainstorm. For the aggregate stability, 5 g of the oven-dry aggregates 3 –4 mm in size were immersed in deionized water for 10 mi and then sieved through a 50-Am sieve in ethanol. The >50-Am fraction was oven-dried and then gently sieved by hand on a column of sieves of mesh sizes 2, 1, 0.5, 0.25 and 0.1 mm.
For the rainfall simulator study, disturbed soil samples ( < 4 mm) were packed in 0.5 by 0.3 m perforated trays in a 0.02-m-thick layer. The trays were placed over a 0.08-m-thick layer of coarse sand in a rainfall simulator of the rotating disk type (Morin et al., 1967). The soil was subjected to a 80-mm rainstorm of deionized water. The rainfall properties were: 1.9 mm, median diameter of the drop; 6.02 m s 1, raindrop median velocity; 18.1 J mm 1 m 2, kinetic energy; and 42 mm h 1, rainfall intensity. During the rainstorm, the volume of rainfall percolating through the soil was measured and the IR was calculated. Likewise, the total volume of the runoff from the entire rainstorm was collected and the particle size distribution of the eroded
Table 1 Mineralogy, mechanical composition, organic matter (OM), CEC and pH for the various studied soils (after Wakindiki and Ben-Hur, 2002; Stern et al., 1991) Soil
Mechanical composition
Location Tunyai Neve Ya’ar Netanya Molo Njoro Irene Collondale Glen Rouxville Orangia Aliwal North Riviersonderend Roodeplaat Cullinam Stutterheim Stanger Dundee Cedara East London Geneva Bapsfontein Potchefstroom Queenstown Silverdale
Kenya Israel Israel Kenya Kenya South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa
Taxonomy
Sand Silt Clay (g kg 1) (g kg 1) (g kg 1)
Eutrustox Chromoxerert Rhodoxeralf Haplustalf Haplustepts Paleudalf Lithic-Dystrchrept Rhodustalf Haplustalf Natrustalf Haplustalf Haplargids Rhodudalf Rhodudalf Haplustalf Plinthudalf Paleudalf Orthox Lithic-Eutrchrepts Natuargid Tepic-Paleustalf Xerochrepts Argiudolls Hapludolls
121 108 900 312 319 32 27 29 13 16 17 23 30 22 27 20 24 39 28 15 21 19 19 29
239 262 0 304 340 26 22 12 12 8 7 32 14 11 18 4 19 42 51 13 13 11 49 47
640 630 100 384 341 42 51 59 75 76 76 45 56 67 55 76 57 19 21 72 66 70 32 24
OM CEC pH (g kg 1) (cmolc kg 1) 34.0 24.3 9.0 40.1 42.0 2.9 2.4 0.7 2.2 1.9 1.9 2.9 1.6 2.2 3.8 2.4 3.4 4.0 3.6 0.9 0.9 2.4 3.6 4.1
17.0 56.0 15.0 24.6 25.6 9.1 10.8 13.5 6.6 5.2 5.8 10.6 12.7 7.1 13.9 10.3 12.0 21.2 17.3 5.5 4.7 4.8 10.6 12.9
N.D.b N.D. N.D. N.D. N.D. 6.9 4.3 6.3 6.1 6.5 6.7 5.7 6.0 6.7 7.1 5.7 6.1 4.9 5.3 6.0 5.2 5.2 5.4 5.1
Soil mineralogya K M S Others 5 3 3 0 0 5 5 4 1 4 1 4 5 5 4 3 5 1 3 5 5 5 5 5
0 3 2 1 1 2 0 5 3 0 5 5 1 2 5 2 2 0 0 5 1 1 3 2
0 5 5 0 0 1 2 1 3 1 1 1 2 1 0 0 0 0 0 0 0 0 0 0
G (1)
Q (5) F (3) Q (5) F (3)
Is [V-Cl] V Is [V-K]
V V
a K = kaolinite, M = mica (mainly illite), S = smectite (mainly montmorillonite), G = gibbsite, Q = quartz, F = feldspar, Cl = clorite, V = vermiculite, Is = interstratified. The numbers for each soil represent the relative amount of the minerals: 0 indicating undetectable amount and 5 indicating the highest amount. b N.D. = not determined.
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
sediment in the runoff was determined by a wetsieving procedure (Gabriels and Moldenhauer, 1978).
2. Clay dispersion and aggregate stability One important characteristic of clays that controls the aggregate stability is their capacity to disperse. Dispersion of clays occurs when the attractive forces between the clay particles are not strong enough to hold them together under wetting conditions (van Olphen, 1977). Dispersion occurs when the electrolyte concentration of the solution decreases below the flocculation value (van Olphen, 1977). Kaolinite flocculates at pH < 7 even in distilled water (Schofield and Samson, 1954). In contrast, the flocculation values of clay systems with 0.05, 0.1 and 0.2 Na/Ca ratios in the exchangeable complex are 3, 4 and 7 meq/l, respectively, for montmorillonite, and 6, 10 and 18 meq/l, respectively, for illite (Shainberg and Letey, 1984). These results suggest that illite is more dispersive than montmorillonite. Singer (1994) reviewed the effects of clay mineralogy on soil dispersivity and found that smectitic soils were the most dispersive and kaolinitic soils the least dispersive. The dispersivity of illitic soils was generally intermediate, but in some cases might exceed that of smectitic soils. The differences in the dispersive behaviour of the clays are attributed to the differences in their morphology and electric charge. The low flocculation value of kaolinite is related to the relatively large interactions (Fig. 1) between the positive charge at the edge of some clay tactoids and the negative charge at the planar surface of the other tactoids (edge to face interaction) (Frenkel et al., 1978). Illite particles consist of platelets stacked together to form tactoids 10 nm in thickness. Electron micrographs reveal that illite tactoids have irregular surfaces and that their planar surfaces are terraced (Quirk, 1978). Thus, when illite particles come together, the unavoidable mismatch of terraces leads to poor contact between the edge and planar surfaces, resulting in high dispersivity. Smectite units are held together in tactoids in faceto-face and/or edge-to-face orientation. In the former orientation, the interactions are maintained by weak Van der Waals’ forces, and by polyvalent metal ions that form bridges between the negatively charged clay platelets. In contrast, in the latter orientation, the
213
interactions are through electrostatic forces (van Olphen, 1977). However, the penetration of water between the clay platelets, and the weakness of the edge-to-face interactions in smectite enhance its dispersivity. Aggregate stability is related to the strength of the interaction between the primary soil particles—sand, silt and clay—in the aggregate. This aggregate stability could be expressed as the mean weight diameter (MWD) of soil particle fractions obtained after soil wetting, by means of Eq. (1): MWD ¼
n X
ð1Þ
x¯ i wi
i¼1
where wi is the weight fraction of aggregates in the size class i with a diameter x¯i. The higher the MWD, the greater was the aggregate stability of the tested soil (Le Bissonnais, 1996). The mean weight diameter values of some soils, as determined by the fast-wetting method (Le Bissonnais, 1996), are presented in Table 2. On the basis of the mineralogy, the studied soils were divided into two main groups: (i) phyllosilicate soils, which include the Tunyai, Neve Ya’ar and Netanya soils; and (ii) non-phyllosilicate soils, which include the Molo and Njoro soils. The first group of soils comprises kaolinitic (Tunyai) and montmorillonitic (Neve Ya’ar and Netanya) soils. The kaolinitic Tunyai soil had the greatest aggregate stability, the montmorillonitic Neve Ya’ar and Netanya soils the lowest, and the non-phyllosilicate Molo and Njoro soils intermediate levels of aggregate Table 2 Mean weight diameter and total soil loss values for various studied soils (after Wakindiki and Ben-Hur, 2002) Soil
Mean weight diameter
Location
Mineralogy
In fast- In runoff wetting (mm) test (mm)
Tunyai Neve Ya’ar Netanya Molo Njoro
Kaolinitic Montmorillonitic Montmorillonitic Non-phyllosilicate Non-phyllosilicate
2.80aa 0.25b 0.31b 0.84c 0.87c
0.12a 0.03b 0.20c 0.18c 0.21c
Total soil loss (kg m 2)
0.33a 1.24b 1.14b 0.75c 0.80c
a Different letters in a column indicate significant differences among soils, P V 0.05.
214
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
stability (Table 2). Although the soil textures and the organic matter contents of the kaolinitic soil (Tunyai) and the montmorillonitic soil (Neve Ya’ar) were fairly similar (Table 1), their MWD values after fast wetting were significantly different. On the other hand, despite the significant differences between the soil textures and organic matter contents of the two montmorillonitic soils (Neve Ya’ar and Netanya), their MWD values were similar (Table 2). These results suggest that the mineralogy of the studied soils had the dominant effect on soil aggregate stability. The higher dispersivity of montmorillonite than of kaolinite gives montmorillonitic soils lower aggregate stability than that of the kaolinitic soil (Table 2). In contrast, the non-phyllosilicate soils do not contain secondary clay minerals. The surface charges of quartz and feldspar, which are the dominant minerals in these soils, are close to zero; therefore, these minerals cannot act as a cement to hold the particles in the aggregate together. Hence, it is likely that the relatively higher organic matter contents in these soils (Table 1) raised their aggregate stability (Le Bissonnais, 1996) above those of the montmorillonitic soils, but not as high as that of the kaolinitic soil (Table 2). Six et al. (2000) determined the aggregate stability of two soil groups: (i) soils characterized by a clay mineralogy dominated by kaolinite and vermiculite; and (ii) soils in which the clay fraction was dominated by chlorite and/or illite. The aggregate stability of the soil from the first group was higher than of the soils from the second group. Six et al. (2000) concluded that, in addition to the stabilizing effect of the kaolinite in the soil from the first group, interactions between the kaolinite and Fe and Al oxides increased the aggregate stability of this soil. The adsorption of Fe and Al oxides on the negatively charged sites on the planar surfaces of the kaolinite reduced the cation exchange capacity (CEC) and increased the positive charge property of this clay. Therefore, this interaction between Fe and Al oxides and kaolinite was synergetic and increased the aggregation potential of kaolinite.
3. Seal formation, infiltration and runoff The effects of soil mineralogy on IR and runoff for five different soils are presented in Fig. 2. In this
figure, the IR is presented as a function of cumulative rainfall. For all the soils, the IR decreased with increasing cumulative rainfall until a final IR was reached (Fig. 2). The photographs of the surfaces of these soils after the simulated rainstorm (Fig. 3) indicate that most of the aggregates at the soil surface were broken down and that a seal was formed. In the montmorillonitic soils, all the aggregates at the soil surface had been dispersed and a smooth seal was developed (Fig. 3). In contrast, in the other soils, the soil surface was partly covered with aggregates, with more of them on the kaolinitic soil surface than on that of the non-phyllosilicate soil (Fig. 3). These photographs indicate that the decrease of the IR during the rainstorm (Fig. 2) was mainly a result of seal formation at the soil surface. On the basis of the final IR values and the total runoff obtained during the entire rainstorm (Fig. 2), the five studied soils were separated into two groups: in the first, comprising only the kaolinitic Tunyai soil, the IR decreased gradually and the final IR was high (20.5 mm h 1); in the second group, comprising the rest of the soils, the IR decreased sharply and the final IR was low ( V 9.3 mm h 1) and significantly different from that of the Tunyai soil. The final IR and the total runoff values shown in Fig. 2 are not consistent with the MWD values obtained in the fast-wetting test (Table 2), which divided the soils into three groups, according to their aggregate stability. This indicates that the aggregate stability of the soil is not the only factor that affects seal formation and IR. The rearrangement of particles in the seal and its thickness should affect the IR of the soil under sealing conditions. The crust micromorphology of the kaolinitic Tunyai, the montmorillonitic Neve Ya’ar and the non-phyllosilicate Molo soils are shown in Fig. 4 and that of the montmorillonitic Netanya soil in Fig. 5. These scanning electron microscope observations reveal that the crust on the kaolinitic soil (Tunyai) was made up of a thin layer ( f 0.1 mm) of slightly compacted aggregates f 0.1 mm in size. In contrast, beneath this 0.1-mm uppermost layer, large aggregates, >0.3 mm in size, were observed. The structure of the layer immediately under the uppermost crust layer (Fig. 4A) was similar to that of the bulk soil at f 5 mm depth, which was not affected by the rainfall (Fig. 4B). This suggests that the soil layer immedi-
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
215
Fig. 2. Infiltration rate as a function of cumulative rainfall for various studied. Bars indicate standard deviation and different letters preceding the final infiltration rate (FIR) and total runoff values indicate significant ( P < 0.05) difference between the soils for each parameter (after Wakindiki and Ben-Hur, 2002).
ately below the crust layer (Fig. 4A) was not disturbed by raindrop impact, and no washed-in layer is apparent on the SEM micrograph of this layer. In the montmorillonitic Neve Ya’ar soil, an uppermost layer f 0.2 mm in thickness (‘‘skin’’ layer) was observed (Fig. 4A), which was packed with aggregates f 0.02 mm in size. Under this layer, a very dense and compacted layer f 0.5 mm thick was formed; this was a washed-in zone, in which dispersed clay particles, transferred from the upper layer with the infiltrated water, accumulated (Fig. 4B). The high clay content of the Neve Ya’ar soil and the high
dispersivity of montmorillonite allowed the high accumulation of the clay particles in this washed-in zone. In the montmorillonitic Netanya soil, the structure of the upper 3-mm layer—the crust layer—was different from that of the bulk soil (Fig. 5). In the crust layer, partly naked sand particles, f 0.2 mm in size, with fine material between them, were observed. In contrast, in the bulk soil, the sand particles were coated with fine materials, and there was almost no fine material in the pores between the sand grains. The crust layer showed two zones: an uppermost 0.5-mm
216
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
Fig. 3. Photographs of the crust surfaces of the various soils (after Wakindiki and Ben-Hur, 2002).
layer with higher density, which was underlain by a 2.5-mm layer with lower density. The high density of the 0.5-mm uppermost layer of the crust could result from infilling of pores with fine material and compaction by the raindrop impacts. It could be that the accumulated relatively fine material between 0.5 and 3 mm depth was the washed-in layer. However, because of the low clay content in the Netanya soil, the amount of dispersed material was small, therefore, there was only limited accumulation of fine materials in the washed-in layer. In the non-phyllosilicate Molo soil, an upper layer, f 0.2 mm in thickness, comprising well-compacted particles f 0.01 mm in size, was observed (Fig. 4A). This upper layer was the crust and its structure was significantly different from that of the bulk soil at the f 5-mm depth (Fig. 4B). In the lower soil layer (Fig. 4B), relatively large aggregates, measuring f 0.3 mm, with small micro-aggregates between them were observed. The low final IR values ( V 9.3 mm h 1) of the kaolinitic, montmorillonitic and non-phyllosilicate soils (Fig. 2) can be explained in terms of the dense structures and relatively large thickness (>0.2 mm) of the crusts that developed on these soils (Figs. 4 and 5). These crusts comprised either only small particles ( f 0.01 mm), or an uppermost layer of particles measuring f 0.02 mm with a highly developed washed-in zone beneath it, or relatively large particles ( f 0.2 mm) with fine material between them. In contrast, in the kaolinitic soil, characterized by a high final IR (20.5 mm h 1) (Fig. 2), the crust was thin ( f 0.1 mm) and it contained relatively large particles ( f 0.1 mm in size) (Fig. 4A). The effects of soil mineralogy on seal formation and IR were also studied by Romkens et al. (1995) in a laboratory rainfall simulator. In this study, the presence of highly expansive smectite clay in selected loess soils caused a rapid reduction of IR despite the high organic C content and the coarse texture of these soils. In this case, the runoff was >50 mm after 82 mm of rain. In contrast, in a soil in which the clay fraction was dominated by vermiculite, mica and kaolinite, no runoff was observed after 82 mm of simulated rainfall. In order to generalize the effects of soil mineralogy on seal formation and IR, Stern et al. (1991) determined the IR of 22 phyllosilicate soils: 19 from South
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
217
Fig. 4. Scanning electron micrographs of (A) the crusts and the underlying soil and (B) the bulk soil at 0.5 cm depth, which was not affected by rainfall, for the Tunyai, Neve Ya’ar and Molo soils, and the washed-in zone of the Neve Ya’ar soil that developed after the application of 80 mm of simulated rain. Scales of the micrographs differ (after Wakindiki and Ben-Hur, 2002).
Africa and three from Israel (Fig. 6). The dominant clay type in these soils was kaolinite, illite or smectite. In Fig. 6, the soils from South Africa are arranged according in descending order of their final IR, with the soil having the highest final IR (19.1 mm h 1) at the beginning and the one with the lowest (2.4 mm h 1) at the end. The three numbers below each column in the Fig. 6 represent the relative amounts of the minerals, kaolinite, illite and smectite, respectively, in the soil; 0 indicating undetectable amounts and 5 indicating the highest amount. It is evident from Fig. 6 that the soils from South Africa can be divided into two main groups: stable soils, with final IR values >8.0 mm h 1; and unstable soils, with final IRs < 4.5 mm h 1. Likewise, the final IRs of the unstable soils were similar to those of the smectitic soils from Israel. Examination of the prop-
erties of the studied soils (Table 1) indicated that the differences in texture, ESP, organic matter and pH of the various soils cannot explain the differences in the final IR values, between the stable and the unstable soil groups (Stern et al., 1991). In contrast, according to the mineralogy of the soils from South Africa, it is evident that kaolinite was the dominant clay in most of the soils and illite in the remaining ones. However, it was noted that only the unstable soils contained smectite, whereas the stable soils did not. Thus, it was suggested that kaolinitic and illitic soils that do not contain any smectite, or in which the smectite level is below the detection threshold, are stable soils, which are less susceptible to seal formation. Conversely, only soils which contain enough smectite (>5%) to be clearly detected by X-rays are unstable and are as susceptible to seal formation as the smectitic soils.
218
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
Fig. 5. Scanning electron micrographs of the crusts and underlying bulk soil for the Netanya soil (after Wakindiki and Ben-Hur, 2002).
Small amounts of smectite may promote kaolinite dispersion by the following mechanism. In kaolinite, the attraction between the positive charges on the edges of the particles and the negative charges on the planar surfaces of other particles was regarded as the cause of the flocculation of this clay, which occurred even in
absence of salt (Schofield and Samson, 1954). However, a rapid and large increase in the flocculation value of kaolinite was observed with increasing amounts of smectite (Schofield and Samson, 1954; Arora and Coleman, 1979); this significant effect was attributed to the deposition of smectite platelets on the positively charged edges of the kaolinite particles, thus preventing the edge-to-face flocculation that occurs in pure kaolinite. A similar phenomenon was reported by Frenkel et al. (1978) who found that the hydraulic conductivity of kaolinitic soils from North Carolina was not affected by 20% Na on the exchange complex, but when the soil was mixed with 2% montmorillonite it became very susceptible to dispersion. Thus, the presence of smectite (even in small quantities) in the soils increased the clay dispersion significantly. This mechanism explains the greater decrease in IR in the soils, which contained smectite than in those that did not (Fig. 6). The role of illite in seal formation is not completely clear from Fig. 6 and should be more studied. It is known, however, that illite is as dispersive as smectite (Oster et al., 1980). However, it is evident from Fig. 6 that the dispersive effect of illite on kaolinite was less pronounced than the effect of smectite. The inefficiency of illite in dispersing kaolinite may be attributed to the terraced shape of the illite particles (Quirk, 1978), which probably led to poor contact between the kaolinite edges and the illite planar surfaces, resulting in inefficient screening of the kaolinite edges by the illite and, consequently, less dispersivity. For the soils that contain smectite (unstable soils), the illite and kaolinite had no consistent effect on the final IR (Fig. 6). In contrast, for the stable soils, the four soils that had the highest final IRs are dominated by kaolinite. Thus, in the absence of smectite, the stabilizing effect of kaolinite in reducing seal formation is more pronounced. As noted above, seal formation is a result of two complementary mechanisms: (i) a physical disintegration of soil aggregates and (ii) physicochemical dispersion of soil clay. The clay dispersion at the soil surface during a rainstorm could occur when rainwater (distilled water) reduces the electrolyte concentration in the soil to below the flocculation value (Agassi et al., 1981). Soil dispersivity under rainfall can be expressed as a dispersion index (DI), which is calculated by dividing the percentage of clay in the runoff sediments by that in the original soil. A DI value of 1
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
219
Fig. 6. Final IRs for stable, unstable and smectitic soils. The numbers below the columns represent the minerals kaolinite, illite and smectite, respectively. Bars represent standard deviation (after Stern et al., 1991).
indicates that no clay dispersion occurred at the soil surface during the rainstorm. In contrast, when the clay fraction at the soil surface is dispersed, the clay percentage in the sediments should be higher than that in the original soil, because the clay particles are more easily transported by the overland flow than the bigger silt and sand particles. In this case, the DI is greater than 1: the higher the DI the more dispersive the soil. The effects of soil dispersion on seal formation and IR, for soils containing various clay types can be seen in Fig. 7. In this figure are presented the DI values of three representative soils from the stable soil group and three representative soils from the unstable group (Fig. 6), that were subjected to 80 mm of distilled
water in a rainfall simulator. The DI values of the Dundee, Cedara and E. London soils (stable soils) were f 1, which indicates that in these soils, there was insignificant clay dispersion at the surface during the rainstorm (Fig. 7). In contrast, the DI values of the Irene, A. North and Roodeplaat soils (unstable soils) ranged from 1.2 to 1.8, indicating that high clay dispersion had occurred at the surfaces of these soils during the seal formation. The final IR values of the Irene, A. North and Roodeplaat soils ranged from 2.4 to 4.5 mm h 1 and those of the Dundee, Cedara and E. London soils from 10.8 to 11.4 mm h 1 (Fig. 6). These results indicate that the physicochemical dispersion of the clay at the soil surface during a
Fig. 7. Dispersion indexes for stable and unstable soils. Bars represent standard deviation (after Stern et al., 1991).
220
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
rainstorm had a significant effect on the seal formation and the IR and that this effect is associated with the soil clay mineralogy. Adding phosphogypsum to the soil surface could prevent soil dispersion during a rainstorm and, consequently, could diminish seal formation. Warrington et al. (1989) found that when phosphogypsum was spread over sandy loam soil exposed to a rainstorm, it dissolved and by doing so, increased the electrolyte concentration in the soil surface, prevented clay dispersion, tripled the soil permeability, and decreased the runoff and the soil loss. The effects of the application of phosphogypsum at 5 Mg ha 1 on the IR of the representative soils from the two soil groups that were subjected to a simulated rainstorm of 80 mm are presented in Fig. 9. The IR for all the soils, both untreated and treated, decreased as the of cumulative rainfall increased, mainly because of seal formation. In the untreated soils, the impact energy of the raindrops and chemical dispersion of soil clay particles caused a seal to form, and the IRs dropped sharply to low final values (2.4 –11.4 mm h 1). In contrast, in the soils treated with phosphogypsum, the phosphogypsum
dissolution increased the electrolyte concentration in the soil solution at the soil surface, preventing the clay dispersion. In this treatment, the seal formation was due mainly to raindrop impact (Agassi et al., 1985) and relatively high final IRs (10.6 – 24.2 mm h 1) were obtained. It can be seen, however, that the final IR values of the treated stable soils were higher than those of the treated unstable soils: the final IR values of the treated stable soils ranged from 20.6 to 24.2 mm h 1 and those of the treated unstable soils from 10.6 to 12.9 mm h 1 (Fig. 8). These results indicate that the soil mineralogy also affects the mechanical stability of the soils: the physical disintegration of aggregates in soils that contain smectite was greater than that in the soils that do not contain smectite.
4. Soil loss During the rainstorm, raindrop impacts break and disperse aggregates at the soil surface and small soil particles are released into the rain-impacted flow at the soil surface. These particles can be transported
Fig. 8. Infiltration curves of untreated and phosphogypsum-treated soils. Bars represent standard deviation (after Stern et al., 1991).
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
by surface runoff and splash and this process constitutes the interrill erosion (Foster, 1982). Because the soil detachment is affected by the soil dispersivity and the aggregate stability, soil erosion should also be influenced by the soil mineralogy. Total soil losses for five different soils during the entire simulated rainstorm are presented in Table 2. The soil loss was significantly lowest in the kaolinitic soil (Tunyai), highest in the montmorillonitic soils (Neve Ya’ar and Netanya) and intermediate in the non-phyllosilicate soils (Molo and Njoro). However, these soil losses were not consistent with the IR values of these soils (Fig. 2), since the soils were divided into only two groups according to their IRs. Interrill erosion involves two major processes: (i) detachment of soil material from the soil surface by raindrop impact and (ii) transport of the resulting sediment by runoff flow. Soil detachment depends mainly on rainfall characteristics and the aggregate stability of the soil. Because all the five soils (Table 2) were subjected to rainfall with the same characteristics, in these soils the soil detachment was depended mainly on their aggregate stability. In contrast, the transport of the sediments depends on the runoff characteristics and the particle-size distribution of the impacted soil surface. Since all the five soils were subjected to rainfall onto a 9% slope, the main runoff characteristic that determined the runoff transport capacity was the runoff rate. Under these conditions, the soil losses from these soils can be explained as follows. The soil loss from the kaolinitic soil was the lowest (Table 2) because the aggregate stability of this soil was the highest (high MWD value in the fast-wetting test, Table 2), which decreased soil detachment, and because the runoff rate obtained during the rainstorm in this soil was the lowest (Fig. 2), which decreased the runoff transport capacity. The montmorillonitic soils, with the lowest aggregate stability (Table 2) and the highest runoff rates (Fig. 2), showed the highest soil losses (Table 2). In the case of the non-phyllosilicate soils, their intermediate aggregate stability (Table 2) and their high runoff rates (Fig. 2) contributed to the intermediate soil losses found in these soils (Table 2). The particle-size distributions of the sediments in the runoff and their MWD values for Tunyai, Neve Ya’ar, Netanya, Molo and Njoro soil are presented in Fig. 9 and Table 2, respectively. The particle-size
221
Fig. 9. Particle size distributions of the sediments in the runoff from the various soils. Bars indicate standard deviation (after Wakindiki and Ben-Hur, 2002).
distribution in the runoff was controlled mainly by that at the impacted soil surface and by the runoff transport capacity during the rainstorm. The high clay
222
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
content in the montmorillonitic Neve Ya’ar soil and the high dispersivity of montmorillonite probably led to the large proportion of very small particles at the impacted soil surface. This led to the large proportion ( f 80%) of clay-size particles (Fig. 9) and the very small (0.03 mm) MWD in the runoff (Table 2) from this soil. In contrast, in the montmorillonitic Netanya soil, which contained 90% sand and 10% clay (Table 1), the sand particles apparently were the most abundant at the impacted soil surface. In this case, the sand particles measuring 0.1 – 0.25 mm in size formed the largest proportion ( f 50%) of the sediments (Fig. 9) and the MWD in the runoff was large (0.2 mm) (Table 2). The clay contents in the kaolinitic Tunyai soil and the montmorillonitic Neve Ya’ar soil were similar, but the MWD in the runoff from the former soil was significantly larger than that from the latter (Table 2). The low dispersivity of the kaolinite in Tunyai soil probably led to high proportions of relatively large particles at the impacted soil surface and in the runoff (Fig. 9) and to the large MWD (0.12 mm) of the sediments (Table 2). The clay contents of the nonphyllosilicate soils were relatively high (Table 1). Likewise, the particles at the crust in these soils were very small ( f 0.01 mm) (Fig. 4A), indicating that the aggregates at the surfaces of these soils were broken down into small particles. However, the MWDs in the runoff from these soils were relatively large, similar to that from the Netanya soil (Table 2). This suggests that, in these soils, some of the aggregates at the soil
surface were not broken down and/or were broken down to relatively large particles that were eroded with the runoff during the rainstorm. Mermut et al. (1997) found also that the soil loss from a soil dominated by smectite was high. The splash and wash erosion in 80 mm of rainstorm were 23 and 2.1 Mg ha 1, respectively, in a loamy soil in which smectite, mica and vermiculite were the dominant clays, and 7.3 and 0.91 Mg ha 1 respectively, in a silt loam soil in which vermiculite, mica and kaolinite were dominant (Mermut et al., 1997). In order to generalize the effects of soil mineralogy on interrill soil loss, Stern et al. (1991) determined the interrill soil loss of 21 phyllosilicate soils that were subjected to 80 mm of distilled water in rainfall simulator (Fig. 10). The results indicated that the phyllosilicate soils could be divided into three groups according their soil mineralogy (Fig. 10). Examination of the susceptibility of the studied phyllosilicate soils to erosion indicated that the soil loss rates (soil loss per mm of rainfall) of all of the unstable soils were significantly ( P < 0.05) higher than those of all of the stable soils (Fig. 10). However, the soil loss rates of the smectitic soils from Israel were significantly higher than those of the unstable soils. These results show that the effect of soil mineralogy on the soil loss (Fig. 10) is different from its effect on final IR (Fig. 6). With regard to soil loss, the phyllosilicate soils fell into three groups according to their mineralogy, whereas with regard to final IR, they fell into two groups. These
Fig. 10. Soil loss rates of stable, unstable and smectitic soils. The numbers below the columns represent the minerals kaolinite, illite and smectite, respectively. Bars represent standard deviation (after Stern et al., 1991).
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
differences were probably because the soil mineralogy affects the size and density of the detached particles, which, in turn, increased the soil loss in the smectitic soils even when the runoff rates on these soils and the unstable soil were similar (Fig. 6).
5. Summary and conclusions Clay mineralogy was found to be a dominant factor in controlling aggregate stability. In phyllosilicate soils, the higher dispersivity of montmorillonite than of kaolinite gives montmorillonitic soils lower aggregate stability than kaolinitic soils. In contrast, the nonphyllosilicate soils contained minerals, such as quartz and feldspar, with surface charges close to zero, which therefore could not act as a cement to hold aggregate particles together. Therefore, in these soils, the aggregate stability was controlled mainly by other agents such as organic matter. The main factor that decreases the soil IR under rainfall conditions in arid and semi-arid regions is seal formation at the soil surface. According to the IR, the phyllosilicate soils could be divided into two main groups: stable soils with final IR >8.0 mm h 1 and unstable soils with final IR < 4.5 mm h 1. These soil groups differ in their mineralogy. Kaolinitic and illitic soils which do not contain smectite were stable soils and less susceptible to seal formation. In contrast, kaolinitic and illitic soils that contain some smectite and smectitic soils were unstable. The soil mineralogy was found to affect the physicochemical dispersion of the clay and the physical disintegration of the aggregates, which were greater in soils that contain smectite than in those that do not. Electron micrographs of the resulting seals revealed that those in smectitic soils were thick (>0.2 mm) and included a highly developed washed-in zone. In contrast, in kaolinitic soils that did not contain smectite the seal was thin ( f 0.1 mm) and contained relatively large particles ( f 0.1 mm in size). Examination of the susceptibility of 21 phyllosilicate soils to interrill erosion indicated that these soils could be divided into three groups. Soil loss was higher for unstable soils than for stable soils, but the soil loss of the smectitic soils was significantly higher than that of the unstable soils, which contained kaolinite or illite as the dominant clay. These results
223
show that soil mineralogy has differing effects on soil loss and on final IR, according to which, the soils were divided into two groups. These differences probably arose because the soil mineralogy affects the size and density of the detached particles, and this led to increased soil loss in smectitic soils, even when those soils and the other unstable soils showed similar runoff rates.
Acknowledgements Contribution from the Agricultural Research Organization, the Volcani Center, no. 625/02, 2002 series.
References Agassi, M., Shainberg, I., Morin, J., 1981. Effect of electrolyte concentration and soil sodicity on the infiltration rate and crust formation. Soil Sci. Soc. Am. J. 45, 848 – 851. Agassi, M., Morin, J., Shainberg, I., 1985. Effect of raindrop impact energy and water salinity on infiltration rates of sodic soils. Soil Sci. Soc. Am. J. 49, 186 – 190. Allen, B.L., Hajek, B.F., 1989. Mineral occurrence in soil environments. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Sci. Soc. Am., Madison, pp. 199 – 239. Arora, H.S., Coleman, N.T., 1979. The influence of electrolyte concentrations on flocculation of clay suspensions. Soil Sci. 127, 134 – 139. Ben-Hur, M., Letey, J., 1989. Effect of polysaccharides, clay dispersion and impact energy on water infiltration. Soil Sci. Soc. Am. J. 53, 233 – 238. Ben-Hur, M., Shainberg, I., Bakker, D., Keren, R., 1985. Effect of soil texture and CaCO3 content on water infiltration in crusted soils as related to water salinity. Irrig. Sci. 6, 281 – 284. Ben-Hur, M., Agassi, M., Keren, R., Zhang, J., 1998. Compaction, aging and raindrop impact effect on hydraulic properties of saline and sodic vertisols. Soil Sci. Soc. Am. J. 62, 1377 – 1383. Bradford, J.M., Foster, G.R., 1996. Interrill soil erosion and slope steepness factors. Soil Sci. Soc. Am. J. 60, 909 – 915. Dixon, J.B., 1989. Kaolin and serpentine group minerals. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Sci. Soc. Am., Madison, pp. 467 – 525. Ekwe, E.I., 1991. The effects of soil organic matter content, rainfall duration and aggregates size on soil detachment. Soil Technol. 4, 197 – 207. Foster, G.R., 1982. Modeling the erosion process. In: Haan, C.T., et al. (Eds.), Hydrologic Modeling of Small Watersheds. ASAE Monogr., vol. 5. ASAE, St. Joseph, MI, pp. 297 – 380. Frenkel, H., Goertzen, J.O., Rhoades, J.D., 1978. Effects of clay type and content, exchangeable sodium percentage, and electro-
224
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
lyte concentration on clay dispersion and soil hydraulic conductivity. Soil Sci. Soc. Am. J. 42, 32 – 39. Fullen, M.A., 1991. Soil organic matter and erosion processes on arable loamy sand soils in the West Midlands of England. Soil Technol. 4, 19 – 31. Gabriels, D., Moldenhauer, W.C., 1978. Size distribution of eroded material from simulated rainfall: effect over a range of soil texture. Soil Sci. Soc. Am. J. 42, 954 – 958. Gal, M., Arcon, L., Shainberg, I., Keren, R., 1984. The effect of exchangeable Na and phosphogypsum on the structure of soil crust—SEM observation. Soil Sci. Soc. Am. J. 48, 872 – 878. Guerra, A., 1994. The effect of organic matter content on soil erosion in simulated rainfall experiments in W. Sussex, UK. Soil Use Manag. 10, 60 – 64. Kinnel, P.I.A., Cummings, D., 1993. Soil/slope gradient interactions in erosion by rain-impacted flow. Trans ASAE 36, 382 – 387. Le Bissonnais, Y., 1996. Aggregate stability and assessment of soil crusting and erodibility: 1. Theory and methodology. Eur. J. Soil Sci. 47, 425 – 437. Le Bissonnais, Y., Arrouays, D., 1997. Aggregate stability and assessment of soil crustability and erodibility: II. Application to humic loamy soils with various organic carbon contents. Eur. J. Soil Sci. 48, 39 – 48. Liebenow, A.M., Elliot, W.J., Laflen, J.M., Kohl, K.D., 1990. Interrill erodibility: collection and analysis of data from crop lands. Trans ASAE 24, 1882 – 1888. McIntyre, D.S., 1958. Permeability measurements of soil crust formed by raindrop impact. Soil Sci. 85, 158 – 189. Mermut, A.R., Luk, S.H., Ro¨mkens, M.J.M., Poesen, J.W.A., 1997. Soil loss by splash and wash during rainfall from two loess soils. Geoderma 75, 203 – 214. Meyer, L.D., 1981. How rain intensity affects interrill erosion. Trans ASAE 24, 1472 – 1475. Morin, J., 1993. Soil crusting and sealing. Soil Tillage in Africa: Needs and Challenges. FAO Soils Bull., vol. 69. FAO Information Division, Rome, Chapter 5. Morin, J., Goldberg, S., Seginer, I., 1967. A rainfall simulator with a rotating disk. Trans ASAE 10, 74 – 79. Morin, J., Benyamini, Y., Michaeli, A., 1981. The dynamics of soil crusting by rainfall impact and the water movement in the soil profile. J. Hydrol. 52, 321 – 335. Neaman, A., Singer, A., Stahr, K., 2000. Dispersion and migration of particles in two palygorskite containing soils of the Jordan Valley. J. Plant Nutr. Soil Sci. 163, 537 – 547.
Oades, J.M., Waters, A.G., 1991. Aggregate hierarchy in soils. Aust. J. Soil Res. 29, 815 – 820. Oster, J.D., Shainberg, I., Wood, J.D., 1980. Flocculation value and gel structure of Na/Ca montmorillonite and illite suspension. Soil Sci. Soc. Am. J. 44, 955 – 959. Quirk, J.P., 1978. Some physico-chemical aspects of soil structural stability: a review. In: Emerson, W.W., Bond, R.D., Dexter, A.R. (Eds.), Modification of Soil Structure. Wiley, New York, pp. 3 – 16. Romkens, M.J.M., Luk, S.H., Poesen, J.W.A., Mermut, A.R., 1995. Rain infiltration into loess soils from different geographic regions. Catena 25, 21 – 32. Schofield, R.K., Samson, H.R., 1954. Flocculation of kaolinite due to the attraction of oppositely charged crystal faces. Faraday Discuss. Chem. Soc. 18, 135 – 145. Schulze, D.G., 1989. An introduction to soil mineralogy. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Sci. Soc. Am., Madison, pp. 1 – 34. Shainberg, I., Letey, J., 1984. Response of soils to sodic and saline conditions. Hilgardia 52, 1 – 57. Sharma, P.P., Gupta, S.C., Foster, G.R., 1995. Raindrop-induced soil detachment and sediment transport from interrill areas. Soil Sci. Soc. Am. J. 59, 727 – 734. Singer, A., 1994. Clay mineralogy as affecting dispersivity and crust formation in aridisols. In: Etchevers, J.D. (Ed.), Transactions of the 15th World Congress of Soil Science. Acapulco, Mexico. Int. Soc. Soil Sci. and Mexican Soc. Soil Sci., Acapulco, Mexico, vol. 8a. Six, J., Elliott, E.T., Paustian, K., 2000. Soil structure and soil organic matter: II. A normalized stability index and the effect of mineralogy. Soil Sci. Soc. Am. J. 64, 1042 – 1049. Stern, R., Ben-Hur, M., Shainberg, I., 1991. Clay mineralogy effect on rain infiltration, seal formation and soil losses. Soil Sci. 152, 455 – 462. van Olphen, H., 1977. An Introduction to Clay Colloid Chemistry, 2nd ed. Interscience Publications, New York. Wakindiki, I.I.C., Ben-Hur, M., 2002. Soil mineralogy and texture effects on crust micromorphology, infiltration and erosion. Soil Sci. Soc. Am. J. 66, 897 – 905. Warrington, D., Shainberg, I., Agassi, M., Morin, J., 1989. Slope and phosphogypsum effect on runoff and erosion. Soil Sci. Soc. Am. J. 53, 1201 – 1205. Watson, D.A., Laflen, J.M., 1986. Soil strength, slope, and rainfall intensity effects on interrill erosion. Trans ASAE 29, 98 – 102.
Soil mineralogy effects on seal formation, runoff and soil loss M. Lado, M. Ben-Hur * Institute of Soil, Water and Environmental Sciences, the Volcani Center, Agricultural Research Organization, P.O. Box 6, Bet Dagan 50250, Israel Received 13 January 2003; received in revised form 13 March 2003; accepted 24 March 2003
Abstract Interaction between clay particles is one of the main factors responsible for soil aggregation. Therefore, soil mineralogy, which has substantial effects on clay dispersion, may also influence aggregate stability, seal formation, runoff and soil loss. In spite of this, the effects of soil mineralogy on these phenomena have received much less attention in the literature than those of other soil properties. This paper reviews the effects of soil mineralogy on aggregate stability, seal formation and micromorphology, and on the associated infiltration rate (IR), runoff and soil loss. The paper is focused mainly on the authors’ previous studies. In order to determine these effects, various soils with different mineralogy were collected from Israel, South Africa and Kenya. In these soils, the aggregate stability was determined under fast wetting conditions, and the IR and the interrill soil loss were determined using rainfall simulator of a rotating disk type. The micromorphology of the crust, which was developed at the soil surface by the rainstorm, was determined by scanning electron microscope. Clay mineralogy was found to be a dominant factor in controlling aggregate stability, seal formation, soil IR and interrill soil loss. The phyllosilicate soils, which were review in this paper, were divided into two main groups: (i) stable soils with final IR>8.0 mm h 1 and (ii) unstable soils with final IR < 4.5 mm h 1. These two soil groups differ in their mineralogy. Kaolinitic and illitic soils that do not contain smectite were stable soils and less susceptible to seal formation. In contrast, kaolinitic and illitic soils that contain some smectite and smectitic soils were unstable. Examination of the susceptibility of 21 phyllosilicate soils to interrill erosion indicated that these soils could be divided into three groups. Soil loss was higher for unstable soils than for stable soils, but the soil loss of the smectitic soils was significantly higher than that of the unstable soils, which contained kaolinite or illite as the dominant clay. D 2003 Elsevier B.V. All rights reserved. Keywords: Aggregate stability; Clay dispersion; Infiltration rate; Soil loss; Soil mineralogy; Surface seal
1. Introduction Runoff and soil erosion are widespread land degradation problems in many parts of the world, because of their contributions to water and soil fertility losses, on the one hand, and their enhancement of flooding
* Corresponding author. E-mail address: [email protected] (M. Ben-Hur). 0169-1317/$ - see front matter D 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.clay.2003.03.002
and surface water pollution risks, on the other hand. Runoff occurs when the rainfall intensity exceeds the soil infiltration rate (IR) and the soil surface water holding capacity, therefore, a decrease in the soil IR could increase the runoff. The main factor that decreases the soil IR under rainfall conditions in arid and semi-arid regions is seal formation at the soil surface (Morin et al., 1981; BenHur and Letey, 1989). A surface seal is thin and is characterized by greater density and lower saturated
210
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
hydraulic conductivity than the underlying soil (McIntyre, 1958; Gal et al., 1984). McIntyre (1958) found that a seal consists of two distinct parts: an upper skin seal attributed to compaction by raindrop impact and a ‘‘washed in’’ zone with decreased porosity, attributed to the accumulation of dispersed clay particles. Likewise, Agassi et al. (1981) suggested that the seal formation is a result of two complementary mechanisms: (i) a physical disintegration of surface soil aggregates caused mainly by the impact energy of the raindrops and leading to formation of the upper skin layer; and (ii) the physicochemical dispersion of soil clay particles, which migrate into the soil with the infiltrating water and clog the pores immediately beneath the surface to form the ‘‘washed in’’ zone. Soil erosion processes can be divided into two components: rill and interrill erosion. Runoff from a soil surface may concentrate into small, erodible channels known as rills. In rill erosion, soil loss is due mainly to detachment of soil particles by flowing water, whereas in interrill erosion, soil particle detachment is caused essentially by raindrop impact, and the particles are transported by raindrop splash and runoff flow (Watson and Laflen, 1986). The present paper deals with interrill erosion only. The tendency of a soil to form a seal, and the resulting amounts of runoff and soil loss depend on soil properties. The soil properties that have received the most attention in this connection are: texture (e.g., Ben-Hur et al., 1985; Romkens et al., 1995; Sharma et al., 1995), organic matter content (e.g., Ekwe, 1991; Fullen, 1991; Guerra, 1994; Le Bissonnais and Arrouays, 1997) and soil sodicity (e.g., Shainberg and Letey, 1984; Ben-Hur et al., 1998). However, soil mineralogy has substantial effects on aggregate stability and clay dispersion (Singer, 1994) and, therefore, may also have effects on seal formation, runoff and soil loss. In spite of this, the effects of soil mineralogy on these phenomena have received much less attention in the literature than the other properties and many studies (e.g., Bradford and Foster, 1996; Meyer, 1981; Watson and Laflen, 1986; Liebenow et al., 1990; Kinnel and Cummings, 1993; Sharma et al., 1995) addressed the runoff and soil loss from various soils under rainfall conditions, without regard to the effects of soil mineralogy on these phenomena. On the basis of mineralogy, soils can be divided into two main groups (Neaman et al., 2000): (i)
phyllosilicate soils that consist mainly of phyllosilicate minerals, of which kaolinite, illite and montmorillonite are the most abundant clays that occur in this mineral group (Allen and Hajek, 1989); (ii) nonphyllosilicate soils that consist of clay-sized minerals, such as quartz and feldspar. Kaolinite is a 1:1 phyllosilicate, in which the basic layer consists of a tetrahedral and an octahedral sheets. Adjacent layers are joined by electrostatic forces between the oxygen ions of the tetrahedral sheet and the hydroxyls of the octahedral sheet in the adjacent layer, to form a tactoid of blocky shape having a typical size of < 2 mm (Dixon, 1989). The bonds between the layers in the tactoid hold them close together, so that water molecules and ions cannot penetrate between the layers. Isomorphic substitution of Si4 + with Al3 + in the tetrahedral sheet, and of Al3 + with Mg2 + or Fe2 + in the octahedral sheet results in a negative charge on the planar surface (van Olphen, 1977). However, because there are few isomorphic substitutions in kaolinite, the electrical charge at the planar surface is weak. There may be an electrical charge at the broken edges of the clay layers; it depends on the pH of the surrounding solution and is positive at neutral pH. Illite and smectite are 2:1 phyllosilicate minerals, in which the basic layer consists of two tetrahedral sheets with one octahedral sheet between them (Schulze, 1989). The main difference between illite and smectite is in the isomorphic substitutions. In illite, most of the substitutions take place in the tetrahedral sheet, where the Si4 + is substituted by Al3 +, leading to a negative charge on the planar surface. This negative charge is balanced mainly by K+ that occupies the interlayer space between two adjacent layers in the tactoid. In this case in general, water molecules and electrolyte cannot penetrate into the interlayer space, and only the external surface of the tactoid is available for interaction with the surrounding solution and other clay particles (Schulze, 1989). In the smectite group, of which montmorillonite is one of the most common clays (Allen and Hajek, 1989), the isomorphic substitutions occur in both the octahedral and the tetrahedral sheets. In this case, water molecules and exchangeable cations can penetrate between adjacent layers. As in the case of kaolinite, there is an electrical charge at the broken edges of the smectite and illite layers, and it depends
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
on the pH of the surrounding solution. The differing morphologies of smectite, illite and kaolinite lead to differing packing of their tactoids, so that they form clay aggregates that differ in their stability (Fig. 1).
211
These clay aggregates are the essential components of soil structure. This paper reviews the effects of soil mineralogy on seal formation and its micromorphology, and the
Fig. 1. (A) Schematic structure of clay aggregates (after Oades and Waters, 1991). (B) Scanning electron micrographs of clay tactoids (after Morin, 1993).
212
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
associated infiltration, runoff and soil loss, and it is focused mainly on the authors’ previous studies. In order to determine these effects, various soils with different mineralogy were collected from Israel, South Africa and Kenya. The general properties of these soils are presented in Table 1, when their ESP values were < 2.5%. These soils were subjected to fastwetting aggregate stability test proposed by Le Bissonnais (1996) and to simulated rainstorm. For the aggregate stability, 5 g of the oven-dry aggregates 3 –4 mm in size were immersed in deionized water for 10 mi and then sieved through a 50-Am sieve in ethanol. The >50-Am fraction was oven-dried and then gently sieved by hand on a column of sieves of mesh sizes 2, 1, 0.5, 0.25 and 0.1 mm.
For the rainfall simulator study, disturbed soil samples ( < 4 mm) were packed in 0.5 by 0.3 m perforated trays in a 0.02-m-thick layer. The trays were placed over a 0.08-m-thick layer of coarse sand in a rainfall simulator of the rotating disk type (Morin et al., 1967). The soil was subjected to a 80-mm rainstorm of deionized water. The rainfall properties were: 1.9 mm, median diameter of the drop; 6.02 m s 1, raindrop median velocity; 18.1 J mm 1 m 2, kinetic energy; and 42 mm h 1, rainfall intensity. During the rainstorm, the volume of rainfall percolating through the soil was measured and the IR was calculated. Likewise, the total volume of the runoff from the entire rainstorm was collected and the particle size distribution of the eroded
Table 1 Mineralogy, mechanical composition, organic matter (OM), CEC and pH for the various studied soils (after Wakindiki and Ben-Hur, 2002; Stern et al., 1991) Soil
Mechanical composition
Location Tunyai Neve Ya’ar Netanya Molo Njoro Irene Collondale Glen Rouxville Orangia Aliwal North Riviersonderend Roodeplaat Cullinam Stutterheim Stanger Dundee Cedara East London Geneva Bapsfontein Potchefstroom Queenstown Silverdale
Kenya Israel Israel Kenya Kenya South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa South Africa
Taxonomy
Sand Silt Clay (g kg 1) (g kg 1) (g kg 1)
Eutrustox Chromoxerert Rhodoxeralf Haplustalf Haplustepts Paleudalf Lithic-Dystrchrept Rhodustalf Haplustalf Natrustalf Haplustalf Haplargids Rhodudalf Rhodudalf Haplustalf Plinthudalf Paleudalf Orthox Lithic-Eutrchrepts Natuargid Tepic-Paleustalf Xerochrepts Argiudolls Hapludolls
121 108 900 312 319 32 27 29 13 16 17 23 30 22 27 20 24 39 28 15 21 19 19 29
239 262 0 304 340 26 22 12 12 8 7 32 14 11 18 4 19 42 51 13 13 11 49 47
640 630 100 384 341 42 51 59 75 76 76 45 56 67 55 76 57 19 21 72 66 70 32 24
OM CEC pH (g kg 1) (cmolc kg 1) 34.0 24.3 9.0 40.1 42.0 2.9 2.4 0.7 2.2 1.9 1.9 2.9 1.6 2.2 3.8 2.4 3.4 4.0 3.6 0.9 0.9 2.4 3.6 4.1
17.0 56.0 15.0 24.6 25.6 9.1 10.8 13.5 6.6 5.2 5.8 10.6 12.7 7.1 13.9 10.3 12.0 21.2 17.3 5.5 4.7 4.8 10.6 12.9
N.D.b N.D. N.D. N.D. N.D. 6.9 4.3 6.3 6.1 6.5 6.7 5.7 6.0 6.7 7.1 5.7 6.1 4.9 5.3 6.0 5.2 5.2 5.4 5.1
Soil mineralogya K M S Others 5 3 3 0 0 5 5 4 1 4 1 4 5 5 4 3 5 1 3 5 5 5 5 5
0 3 2 1 1 2 0 5 3 0 5 5 1 2 5 2 2 0 0 5 1 1 3 2
0 5 5 0 0 1 2 1 3 1 1 1 2 1 0 0 0 0 0 0 0 0 0 0
G (1)
Q (5) F (3) Q (5) F (3)
Is [V-Cl] V Is [V-K]
V V
a K = kaolinite, M = mica (mainly illite), S = smectite (mainly montmorillonite), G = gibbsite, Q = quartz, F = feldspar, Cl = clorite, V = vermiculite, Is = interstratified. The numbers for each soil represent the relative amount of the minerals: 0 indicating undetectable amount and 5 indicating the highest amount. b N.D. = not determined.
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
sediment in the runoff was determined by a wetsieving procedure (Gabriels and Moldenhauer, 1978).
2. Clay dispersion and aggregate stability One important characteristic of clays that controls the aggregate stability is their capacity to disperse. Dispersion of clays occurs when the attractive forces between the clay particles are not strong enough to hold them together under wetting conditions (van Olphen, 1977). Dispersion occurs when the electrolyte concentration of the solution decreases below the flocculation value (van Olphen, 1977). Kaolinite flocculates at pH < 7 even in distilled water (Schofield and Samson, 1954). In contrast, the flocculation values of clay systems with 0.05, 0.1 and 0.2 Na/Ca ratios in the exchangeable complex are 3, 4 and 7 meq/l, respectively, for montmorillonite, and 6, 10 and 18 meq/l, respectively, for illite (Shainberg and Letey, 1984). These results suggest that illite is more dispersive than montmorillonite. Singer (1994) reviewed the effects of clay mineralogy on soil dispersivity and found that smectitic soils were the most dispersive and kaolinitic soils the least dispersive. The dispersivity of illitic soils was generally intermediate, but in some cases might exceed that of smectitic soils. The differences in the dispersive behaviour of the clays are attributed to the differences in their morphology and electric charge. The low flocculation value of kaolinite is related to the relatively large interactions (Fig. 1) between the positive charge at the edge of some clay tactoids and the negative charge at the planar surface of the other tactoids (edge to face interaction) (Frenkel et al., 1978). Illite particles consist of platelets stacked together to form tactoids 10 nm in thickness. Electron micrographs reveal that illite tactoids have irregular surfaces and that their planar surfaces are terraced (Quirk, 1978). Thus, when illite particles come together, the unavoidable mismatch of terraces leads to poor contact between the edge and planar surfaces, resulting in high dispersivity. Smectite units are held together in tactoids in faceto-face and/or edge-to-face orientation. In the former orientation, the interactions are maintained by weak Van der Waals’ forces, and by polyvalent metal ions that form bridges between the negatively charged clay platelets. In contrast, in the latter orientation, the
213
interactions are through electrostatic forces (van Olphen, 1977). However, the penetration of water between the clay platelets, and the weakness of the edge-to-face interactions in smectite enhance its dispersivity. Aggregate stability is related to the strength of the interaction between the primary soil particles—sand, silt and clay—in the aggregate. This aggregate stability could be expressed as the mean weight diameter (MWD) of soil particle fractions obtained after soil wetting, by means of Eq. (1): MWD ¼
n X
ð1Þ
x¯ i wi
i¼1
where wi is the weight fraction of aggregates in the size class i with a diameter x¯i. The higher the MWD, the greater was the aggregate stability of the tested soil (Le Bissonnais, 1996). The mean weight diameter values of some soils, as determined by the fast-wetting method (Le Bissonnais, 1996), are presented in Table 2. On the basis of the mineralogy, the studied soils were divided into two main groups: (i) phyllosilicate soils, which include the Tunyai, Neve Ya’ar and Netanya soils; and (ii) non-phyllosilicate soils, which include the Molo and Njoro soils. The first group of soils comprises kaolinitic (Tunyai) and montmorillonitic (Neve Ya’ar and Netanya) soils. The kaolinitic Tunyai soil had the greatest aggregate stability, the montmorillonitic Neve Ya’ar and Netanya soils the lowest, and the non-phyllosilicate Molo and Njoro soils intermediate levels of aggregate Table 2 Mean weight diameter and total soil loss values for various studied soils (after Wakindiki and Ben-Hur, 2002) Soil
Mean weight diameter
Location
Mineralogy
In fast- In runoff wetting (mm) test (mm)
Tunyai Neve Ya’ar Netanya Molo Njoro
Kaolinitic Montmorillonitic Montmorillonitic Non-phyllosilicate Non-phyllosilicate
2.80aa 0.25b 0.31b 0.84c 0.87c
0.12a 0.03b 0.20c 0.18c 0.21c
Total soil loss (kg m 2)
0.33a 1.24b 1.14b 0.75c 0.80c
a Different letters in a column indicate significant differences among soils, P V 0.05.
214
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
stability (Table 2). Although the soil textures and the organic matter contents of the kaolinitic soil (Tunyai) and the montmorillonitic soil (Neve Ya’ar) were fairly similar (Table 1), their MWD values after fast wetting were significantly different. On the other hand, despite the significant differences between the soil textures and organic matter contents of the two montmorillonitic soils (Neve Ya’ar and Netanya), their MWD values were similar (Table 2). These results suggest that the mineralogy of the studied soils had the dominant effect on soil aggregate stability. The higher dispersivity of montmorillonite than of kaolinite gives montmorillonitic soils lower aggregate stability than that of the kaolinitic soil (Table 2). In contrast, the non-phyllosilicate soils do not contain secondary clay minerals. The surface charges of quartz and feldspar, which are the dominant minerals in these soils, are close to zero; therefore, these minerals cannot act as a cement to hold the particles in the aggregate together. Hence, it is likely that the relatively higher organic matter contents in these soils (Table 1) raised their aggregate stability (Le Bissonnais, 1996) above those of the montmorillonitic soils, but not as high as that of the kaolinitic soil (Table 2). Six et al. (2000) determined the aggregate stability of two soil groups: (i) soils characterized by a clay mineralogy dominated by kaolinite and vermiculite; and (ii) soils in which the clay fraction was dominated by chlorite and/or illite. The aggregate stability of the soil from the first group was higher than of the soils from the second group. Six et al. (2000) concluded that, in addition to the stabilizing effect of the kaolinite in the soil from the first group, interactions between the kaolinite and Fe and Al oxides increased the aggregate stability of this soil. The adsorption of Fe and Al oxides on the negatively charged sites on the planar surfaces of the kaolinite reduced the cation exchange capacity (CEC) and increased the positive charge property of this clay. Therefore, this interaction between Fe and Al oxides and kaolinite was synergetic and increased the aggregation potential of kaolinite.
3. Seal formation, infiltration and runoff The effects of soil mineralogy on IR and runoff for five different soils are presented in Fig. 2. In this
figure, the IR is presented as a function of cumulative rainfall. For all the soils, the IR decreased with increasing cumulative rainfall until a final IR was reached (Fig. 2). The photographs of the surfaces of these soils after the simulated rainstorm (Fig. 3) indicate that most of the aggregates at the soil surface were broken down and that a seal was formed. In the montmorillonitic soils, all the aggregates at the soil surface had been dispersed and a smooth seal was developed (Fig. 3). In contrast, in the other soils, the soil surface was partly covered with aggregates, with more of them on the kaolinitic soil surface than on that of the non-phyllosilicate soil (Fig. 3). These photographs indicate that the decrease of the IR during the rainstorm (Fig. 2) was mainly a result of seal formation at the soil surface. On the basis of the final IR values and the total runoff obtained during the entire rainstorm (Fig. 2), the five studied soils were separated into two groups: in the first, comprising only the kaolinitic Tunyai soil, the IR decreased gradually and the final IR was high (20.5 mm h 1); in the second group, comprising the rest of the soils, the IR decreased sharply and the final IR was low ( V 9.3 mm h 1) and significantly different from that of the Tunyai soil. The final IR and the total runoff values shown in Fig. 2 are not consistent with the MWD values obtained in the fast-wetting test (Table 2), which divided the soils into three groups, according to their aggregate stability. This indicates that the aggregate stability of the soil is not the only factor that affects seal formation and IR. The rearrangement of particles in the seal and its thickness should affect the IR of the soil under sealing conditions. The crust micromorphology of the kaolinitic Tunyai, the montmorillonitic Neve Ya’ar and the non-phyllosilicate Molo soils are shown in Fig. 4 and that of the montmorillonitic Netanya soil in Fig. 5. These scanning electron microscope observations reveal that the crust on the kaolinitic soil (Tunyai) was made up of a thin layer ( f 0.1 mm) of slightly compacted aggregates f 0.1 mm in size. In contrast, beneath this 0.1-mm uppermost layer, large aggregates, >0.3 mm in size, were observed. The structure of the layer immediately under the uppermost crust layer (Fig. 4A) was similar to that of the bulk soil at f 5 mm depth, which was not affected by the rainfall (Fig. 4B). This suggests that the soil layer immedi-
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
215
Fig. 2. Infiltration rate as a function of cumulative rainfall for various studied. Bars indicate standard deviation and different letters preceding the final infiltration rate (FIR) and total runoff values indicate significant ( P < 0.05) difference between the soils for each parameter (after Wakindiki and Ben-Hur, 2002).
ately below the crust layer (Fig. 4A) was not disturbed by raindrop impact, and no washed-in layer is apparent on the SEM micrograph of this layer. In the montmorillonitic Neve Ya’ar soil, an uppermost layer f 0.2 mm in thickness (‘‘skin’’ layer) was observed (Fig. 4A), which was packed with aggregates f 0.02 mm in size. Under this layer, a very dense and compacted layer f 0.5 mm thick was formed; this was a washed-in zone, in which dispersed clay particles, transferred from the upper layer with the infiltrated water, accumulated (Fig. 4B). The high clay content of the Neve Ya’ar soil and the high
dispersivity of montmorillonite allowed the high accumulation of the clay particles in this washed-in zone. In the montmorillonitic Netanya soil, the structure of the upper 3-mm layer—the crust layer—was different from that of the bulk soil (Fig. 5). In the crust layer, partly naked sand particles, f 0.2 mm in size, with fine material between them, were observed. In contrast, in the bulk soil, the sand particles were coated with fine materials, and there was almost no fine material in the pores between the sand grains. The crust layer showed two zones: an uppermost 0.5-mm
216
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
Fig. 3. Photographs of the crust surfaces of the various soils (after Wakindiki and Ben-Hur, 2002).
layer with higher density, which was underlain by a 2.5-mm layer with lower density. The high density of the 0.5-mm uppermost layer of the crust could result from infilling of pores with fine material and compaction by the raindrop impacts. It could be that the accumulated relatively fine material between 0.5 and 3 mm depth was the washed-in layer. However, because of the low clay content in the Netanya soil, the amount of dispersed material was small, therefore, there was only limited accumulation of fine materials in the washed-in layer. In the non-phyllosilicate Molo soil, an upper layer, f 0.2 mm in thickness, comprising well-compacted particles f 0.01 mm in size, was observed (Fig. 4A). This upper layer was the crust and its structure was significantly different from that of the bulk soil at the f 5-mm depth (Fig. 4B). In the lower soil layer (Fig. 4B), relatively large aggregates, measuring f 0.3 mm, with small micro-aggregates between them were observed. The low final IR values ( V 9.3 mm h 1) of the kaolinitic, montmorillonitic and non-phyllosilicate soils (Fig. 2) can be explained in terms of the dense structures and relatively large thickness (>0.2 mm) of the crusts that developed on these soils (Figs. 4 and 5). These crusts comprised either only small particles ( f 0.01 mm), or an uppermost layer of particles measuring f 0.02 mm with a highly developed washed-in zone beneath it, or relatively large particles ( f 0.2 mm) with fine material between them. In contrast, in the kaolinitic soil, characterized by a high final IR (20.5 mm h 1) (Fig. 2), the crust was thin ( f 0.1 mm) and it contained relatively large particles ( f 0.1 mm in size) (Fig. 4A). The effects of soil mineralogy on seal formation and IR were also studied by Romkens et al. (1995) in a laboratory rainfall simulator. In this study, the presence of highly expansive smectite clay in selected loess soils caused a rapid reduction of IR despite the high organic C content and the coarse texture of these soils. In this case, the runoff was >50 mm after 82 mm of rain. In contrast, in a soil in which the clay fraction was dominated by vermiculite, mica and kaolinite, no runoff was observed after 82 mm of simulated rainfall. In order to generalize the effects of soil mineralogy on seal formation and IR, Stern et al. (1991) determined the IR of 22 phyllosilicate soils: 19 from South
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
217
Fig. 4. Scanning electron micrographs of (A) the crusts and the underlying soil and (B) the bulk soil at 0.5 cm depth, which was not affected by rainfall, for the Tunyai, Neve Ya’ar and Molo soils, and the washed-in zone of the Neve Ya’ar soil that developed after the application of 80 mm of simulated rain. Scales of the micrographs differ (after Wakindiki and Ben-Hur, 2002).
Africa and three from Israel (Fig. 6). The dominant clay type in these soils was kaolinite, illite or smectite. In Fig. 6, the soils from South Africa are arranged according in descending order of their final IR, with the soil having the highest final IR (19.1 mm h 1) at the beginning and the one with the lowest (2.4 mm h 1) at the end. The three numbers below each column in the Fig. 6 represent the relative amounts of the minerals, kaolinite, illite and smectite, respectively, in the soil; 0 indicating undetectable amounts and 5 indicating the highest amount. It is evident from Fig. 6 that the soils from South Africa can be divided into two main groups: stable soils, with final IR values >8.0 mm h 1; and unstable soils, with final IRs < 4.5 mm h 1. Likewise, the final IRs of the unstable soils were similar to those of the smectitic soils from Israel. Examination of the prop-
erties of the studied soils (Table 1) indicated that the differences in texture, ESP, organic matter and pH of the various soils cannot explain the differences in the final IR values, between the stable and the unstable soil groups (Stern et al., 1991). In contrast, according to the mineralogy of the soils from South Africa, it is evident that kaolinite was the dominant clay in most of the soils and illite in the remaining ones. However, it was noted that only the unstable soils contained smectite, whereas the stable soils did not. Thus, it was suggested that kaolinitic and illitic soils that do not contain any smectite, or in which the smectite level is below the detection threshold, are stable soils, which are less susceptible to seal formation. Conversely, only soils which contain enough smectite (>5%) to be clearly detected by X-rays are unstable and are as susceptible to seal formation as the smectitic soils.
218
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
Fig. 5. Scanning electron micrographs of the crusts and underlying bulk soil for the Netanya soil (after Wakindiki and Ben-Hur, 2002).
Small amounts of smectite may promote kaolinite dispersion by the following mechanism. In kaolinite, the attraction between the positive charges on the edges of the particles and the negative charges on the planar surfaces of other particles was regarded as the cause of the flocculation of this clay, which occurred even in
absence of salt (Schofield and Samson, 1954). However, a rapid and large increase in the flocculation value of kaolinite was observed with increasing amounts of smectite (Schofield and Samson, 1954; Arora and Coleman, 1979); this significant effect was attributed to the deposition of smectite platelets on the positively charged edges of the kaolinite particles, thus preventing the edge-to-face flocculation that occurs in pure kaolinite. A similar phenomenon was reported by Frenkel et al. (1978) who found that the hydraulic conductivity of kaolinitic soils from North Carolina was not affected by 20% Na on the exchange complex, but when the soil was mixed with 2% montmorillonite it became very susceptible to dispersion. Thus, the presence of smectite (even in small quantities) in the soils increased the clay dispersion significantly. This mechanism explains the greater decrease in IR in the soils, which contained smectite than in those that did not (Fig. 6). The role of illite in seal formation is not completely clear from Fig. 6 and should be more studied. It is known, however, that illite is as dispersive as smectite (Oster et al., 1980). However, it is evident from Fig. 6 that the dispersive effect of illite on kaolinite was less pronounced than the effect of smectite. The inefficiency of illite in dispersing kaolinite may be attributed to the terraced shape of the illite particles (Quirk, 1978), which probably led to poor contact between the kaolinite edges and the illite planar surfaces, resulting in inefficient screening of the kaolinite edges by the illite and, consequently, less dispersivity. For the soils that contain smectite (unstable soils), the illite and kaolinite had no consistent effect on the final IR (Fig. 6). In contrast, for the stable soils, the four soils that had the highest final IRs are dominated by kaolinite. Thus, in the absence of smectite, the stabilizing effect of kaolinite in reducing seal formation is more pronounced. As noted above, seal formation is a result of two complementary mechanisms: (i) a physical disintegration of soil aggregates and (ii) physicochemical dispersion of soil clay. The clay dispersion at the soil surface during a rainstorm could occur when rainwater (distilled water) reduces the electrolyte concentration in the soil to below the flocculation value (Agassi et al., 1981). Soil dispersivity under rainfall can be expressed as a dispersion index (DI), which is calculated by dividing the percentage of clay in the runoff sediments by that in the original soil. A DI value of 1
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
219
Fig. 6. Final IRs for stable, unstable and smectitic soils. The numbers below the columns represent the minerals kaolinite, illite and smectite, respectively. Bars represent standard deviation (after Stern et al., 1991).
indicates that no clay dispersion occurred at the soil surface during the rainstorm. In contrast, when the clay fraction at the soil surface is dispersed, the clay percentage in the sediments should be higher than that in the original soil, because the clay particles are more easily transported by the overland flow than the bigger silt and sand particles. In this case, the DI is greater than 1: the higher the DI the more dispersive the soil. The effects of soil dispersion on seal formation and IR, for soils containing various clay types can be seen in Fig. 7. In this figure are presented the DI values of three representative soils from the stable soil group and three representative soils from the unstable group (Fig. 6), that were subjected to 80 mm of distilled
water in a rainfall simulator. The DI values of the Dundee, Cedara and E. London soils (stable soils) were f 1, which indicates that in these soils, there was insignificant clay dispersion at the surface during the rainstorm (Fig. 7). In contrast, the DI values of the Irene, A. North and Roodeplaat soils (unstable soils) ranged from 1.2 to 1.8, indicating that high clay dispersion had occurred at the surfaces of these soils during the seal formation. The final IR values of the Irene, A. North and Roodeplaat soils ranged from 2.4 to 4.5 mm h 1 and those of the Dundee, Cedara and E. London soils from 10.8 to 11.4 mm h 1 (Fig. 6). These results indicate that the physicochemical dispersion of the clay at the soil surface during a
Fig. 7. Dispersion indexes for stable and unstable soils. Bars represent standard deviation (after Stern et al., 1991).
220
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
rainstorm had a significant effect on the seal formation and the IR and that this effect is associated with the soil clay mineralogy. Adding phosphogypsum to the soil surface could prevent soil dispersion during a rainstorm and, consequently, could diminish seal formation. Warrington et al. (1989) found that when phosphogypsum was spread over sandy loam soil exposed to a rainstorm, it dissolved and by doing so, increased the electrolyte concentration in the soil surface, prevented clay dispersion, tripled the soil permeability, and decreased the runoff and the soil loss. The effects of the application of phosphogypsum at 5 Mg ha 1 on the IR of the representative soils from the two soil groups that were subjected to a simulated rainstorm of 80 mm are presented in Fig. 9. The IR for all the soils, both untreated and treated, decreased as the of cumulative rainfall increased, mainly because of seal formation. In the untreated soils, the impact energy of the raindrops and chemical dispersion of soil clay particles caused a seal to form, and the IRs dropped sharply to low final values (2.4 –11.4 mm h 1). In contrast, in the soils treated with phosphogypsum, the phosphogypsum
dissolution increased the electrolyte concentration in the soil solution at the soil surface, preventing the clay dispersion. In this treatment, the seal formation was due mainly to raindrop impact (Agassi et al., 1985) and relatively high final IRs (10.6 – 24.2 mm h 1) were obtained. It can be seen, however, that the final IR values of the treated stable soils were higher than those of the treated unstable soils: the final IR values of the treated stable soils ranged from 20.6 to 24.2 mm h 1 and those of the treated unstable soils from 10.6 to 12.9 mm h 1 (Fig. 8). These results indicate that the soil mineralogy also affects the mechanical stability of the soils: the physical disintegration of aggregates in soils that contain smectite was greater than that in the soils that do not contain smectite.
4. Soil loss During the rainstorm, raindrop impacts break and disperse aggregates at the soil surface and small soil particles are released into the rain-impacted flow at the soil surface. These particles can be transported
Fig. 8. Infiltration curves of untreated and phosphogypsum-treated soils. Bars represent standard deviation (after Stern et al., 1991).
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
by surface runoff and splash and this process constitutes the interrill erosion (Foster, 1982). Because the soil detachment is affected by the soil dispersivity and the aggregate stability, soil erosion should also be influenced by the soil mineralogy. Total soil losses for five different soils during the entire simulated rainstorm are presented in Table 2. The soil loss was significantly lowest in the kaolinitic soil (Tunyai), highest in the montmorillonitic soils (Neve Ya’ar and Netanya) and intermediate in the non-phyllosilicate soils (Molo and Njoro). However, these soil losses were not consistent with the IR values of these soils (Fig. 2), since the soils were divided into only two groups according to their IRs. Interrill erosion involves two major processes: (i) detachment of soil material from the soil surface by raindrop impact and (ii) transport of the resulting sediment by runoff flow. Soil detachment depends mainly on rainfall characteristics and the aggregate stability of the soil. Because all the five soils (Table 2) were subjected to rainfall with the same characteristics, in these soils the soil detachment was depended mainly on their aggregate stability. In contrast, the transport of the sediments depends on the runoff characteristics and the particle-size distribution of the impacted soil surface. Since all the five soils were subjected to rainfall onto a 9% slope, the main runoff characteristic that determined the runoff transport capacity was the runoff rate. Under these conditions, the soil losses from these soils can be explained as follows. The soil loss from the kaolinitic soil was the lowest (Table 2) because the aggregate stability of this soil was the highest (high MWD value in the fast-wetting test, Table 2), which decreased soil detachment, and because the runoff rate obtained during the rainstorm in this soil was the lowest (Fig. 2), which decreased the runoff transport capacity. The montmorillonitic soils, with the lowest aggregate stability (Table 2) and the highest runoff rates (Fig. 2), showed the highest soil losses (Table 2). In the case of the non-phyllosilicate soils, their intermediate aggregate stability (Table 2) and their high runoff rates (Fig. 2) contributed to the intermediate soil losses found in these soils (Table 2). The particle-size distributions of the sediments in the runoff and their MWD values for Tunyai, Neve Ya’ar, Netanya, Molo and Njoro soil are presented in Fig. 9 and Table 2, respectively. The particle-size
221
Fig. 9. Particle size distributions of the sediments in the runoff from the various soils. Bars indicate standard deviation (after Wakindiki and Ben-Hur, 2002).
distribution in the runoff was controlled mainly by that at the impacted soil surface and by the runoff transport capacity during the rainstorm. The high clay
222
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
content in the montmorillonitic Neve Ya’ar soil and the high dispersivity of montmorillonite probably led to the large proportion of very small particles at the impacted soil surface. This led to the large proportion ( f 80%) of clay-size particles (Fig. 9) and the very small (0.03 mm) MWD in the runoff (Table 2) from this soil. In contrast, in the montmorillonitic Netanya soil, which contained 90% sand and 10% clay (Table 1), the sand particles apparently were the most abundant at the impacted soil surface. In this case, the sand particles measuring 0.1 – 0.25 mm in size formed the largest proportion ( f 50%) of the sediments (Fig. 9) and the MWD in the runoff was large (0.2 mm) (Table 2). The clay contents in the kaolinitic Tunyai soil and the montmorillonitic Neve Ya’ar soil were similar, but the MWD in the runoff from the former soil was significantly larger than that from the latter (Table 2). The low dispersivity of the kaolinite in Tunyai soil probably led to high proportions of relatively large particles at the impacted soil surface and in the runoff (Fig. 9) and to the large MWD (0.12 mm) of the sediments (Table 2). The clay contents of the nonphyllosilicate soils were relatively high (Table 1). Likewise, the particles at the crust in these soils were very small ( f 0.01 mm) (Fig. 4A), indicating that the aggregates at the surfaces of these soils were broken down into small particles. However, the MWDs in the runoff from these soils were relatively large, similar to that from the Netanya soil (Table 2). This suggests that, in these soils, some of the aggregates at the soil
surface were not broken down and/or were broken down to relatively large particles that were eroded with the runoff during the rainstorm. Mermut et al. (1997) found also that the soil loss from a soil dominated by smectite was high. The splash and wash erosion in 80 mm of rainstorm were 23 and 2.1 Mg ha 1, respectively, in a loamy soil in which smectite, mica and vermiculite were the dominant clays, and 7.3 and 0.91 Mg ha 1 respectively, in a silt loam soil in which vermiculite, mica and kaolinite were dominant (Mermut et al., 1997). In order to generalize the effects of soil mineralogy on interrill soil loss, Stern et al. (1991) determined the interrill soil loss of 21 phyllosilicate soils that were subjected to 80 mm of distilled water in rainfall simulator (Fig. 10). The results indicated that the phyllosilicate soils could be divided into three groups according their soil mineralogy (Fig. 10). Examination of the susceptibility of the studied phyllosilicate soils to erosion indicated that the soil loss rates (soil loss per mm of rainfall) of all of the unstable soils were significantly ( P < 0.05) higher than those of all of the stable soils (Fig. 10). However, the soil loss rates of the smectitic soils from Israel were significantly higher than those of the unstable soils. These results show that the effect of soil mineralogy on the soil loss (Fig. 10) is different from its effect on final IR (Fig. 6). With regard to soil loss, the phyllosilicate soils fell into three groups according to their mineralogy, whereas with regard to final IR, they fell into two groups. These
Fig. 10. Soil loss rates of stable, unstable and smectitic soils. The numbers below the columns represent the minerals kaolinite, illite and smectite, respectively. Bars represent standard deviation (after Stern et al., 1991).
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
differences were probably because the soil mineralogy affects the size and density of the detached particles, which, in turn, increased the soil loss in the smectitic soils even when the runoff rates on these soils and the unstable soil were similar (Fig. 6).
5. Summary and conclusions Clay mineralogy was found to be a dominant factor in controlling aggregate stability. In phyllosilicate soils, the higher dispersivity of montmorillonite than of kaolinite gives montmorillonitic soils lower aggregate stability than kaolinitic soils. In contrast, the nonphyllosilicate soils contained minerals, such as quartz and feldspar, with surface charges close to zero, which therefore could not act as a cement to hold aggregate particles together. Therefore, in these soils, the aggregate stability was controlled mainly by other agents such as organic matter. The main factor that decreases the soil IR under rainfall conditions in arid and semi-arid regions is seal formation at the soil surface. According to the IR, the phyllosilicate soils could be divided into two main groups: stable soils with final IR >8.0 mm h 1 and unstable soils with final IR < 4.5 mm h 1. These soil groups differ in their mineralogy. Kaolinitic and illitic soils which do not contain smectite were stable soils and less susceptible to seal formation. In contrast, kaolinitic and illitic soils that contain some smectite and smectitic soils were unstable. The soil mineralogy was found to affect the physicochemical dispersion of the clay and the physical disintegration of the aggregates, which were greater in soils that contain smectite than in those that do not. Electron micrographs of the resulting seals revealed that those in smectitic soils were thick (>0.2 mm) and included a highly developed washed-in zone. In contrast, in kaolinitic soils that did not contain smectite the seal was thin ( f 0.1 mm) and contained relatively large particles ( f 0.1 mm in size). Examination of the susceptibility of 21 phyllosilicate soils to interrill erosion indicated that these soils could be divided into three groups. Soil loss was higher for unstable soils than for stable soils, but the soil loss of the smectitic soils was significantly higher than that of the unstable soils, which contained kaolinite or illite as the dominant clay. These results
223
show that soil mineralogy has differing effects on soil loss and on final IR, according to which, the soils were divided into two groups. These differences probably arose because the soil mineralogy affects the size and density of the detached particles, and this led to increased soil loss in smectitic soils, even when those soils and the other unstable soils showed similar runoff rates.
Acknowledgements Contribution from the Agricultural Research Organization, the Volcani Center, no. 625/02, 2002 series.
References Agassi, M., Shainberg, I., Morin, J., 1981. Effect of electrolyte concentration and soil sodicity on the infiltration rate and crust formation. Soil Sci. Soc. Am. J. 45, 848 – 851. Agassi, M., Morin, J., Shainberg, I., 1985. Effect of raindrop impact energy and water salinity on infiltration rates of sodic soils. Soil Sci. Soc. Am. J. 49, 186 – 190. Allen, B.L., Hajek, B.F., 1989. Mineral occurrence in soil environments. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Sci. Soc. Am., Madison, pp. 199 – 239. Arora, H.S., Coleman, N.T., 1979. The influence of electrolyte concentrations on flocculation of clay suspensions. Soil Sci. 127, 134 – 139. Ben-Hur, M., Letey, J., 1989. Effect of polysaccharides, clay dispersion and impact energy on water infiltration. Soil Sci. Soc. Am. J. 53, 233 – 238. Ben-Hur, M., Shainberg, I., Bakker, D., Keren, R., 1985. Effect of soil texture and CaCO3 content on water infiltration in crusted soils as related to water salinity. Irrig. Sci. 6, 281 – 284. Ben-Hur, M., Agassi, M., Keren, R., Zhang, J., 1998. Compaction, aging and raindrop impact effect on hydraulic properties of saline and sodic vertisols. Soil Sci. Soc. Am. J. 62, 1377 – 1383. Bradford, J.M., Foster, G.R., 1996. Interrill soil erosion and slope steepness factors. Soil Sci. Soc. Am. J. 60, 909 – 915. Dixon, J.B., 1989. Kaolin and serpentine group minerals. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Sci. Soc. Am., Madison, pp. 467 – 525. Ekwe, E.I., 1991. The effects of soil organic matter content, rainfall duration and aggregates size on soil detachment. Soil Technol. 4, 197 – 207. Foster, G.R., 1982. Modeling the erosion process. In: Haan, C.T., et al. (Eds.), Hydrologic Modeling of Small Watersheds. ASAE Monogr., vol. 5. ASAE, St. Joseph, MI, pp. 297 – 380. Frenkel, H., Goertzen, J.O., Rhoades, J.D., 1978. Effects of clay type and content, exchangeable sodium percentage, and electro-
224
M. Lado, M. Ben-Hur / Applied Clay Science 24 (2004) 209–224
lyte concentration on clay dispersion and soil hydraulic conductivity. Soil Sci. Soc. Am. J. 42, 32 – 39. Fullen, M.A., 1991. Soil organic matter and erosion processes on arable loamy sand soils in the West Midlands of England. Soil Technol. 4, 19 – 31. Gabriels, D., Moldenhauer, W.C., 1978. Size distribution of eroded material from simulated rainfall: effect over a range of soil texture. Soil Sci. Soc. Am. J. 42, 954 – 958. Gal, M., Arcon, L., Shainberg, I., Keren, R., 1984. The effect of exchangeable Na and phosphogypsum on the structure of soil crust—SEM observation. Soil Sci. Soc. Am. J. 48, 872 – 878. Guerra, A., 1994. The effect of organic matter content on soil erosion in simulated rainfall experiments in W. Sussex, UK. Soil Use Manag. 10, 60 – 64. Kinnel, P.I.A., Cummings, D., 1993. Soil/slope gradient interactions in erosion by rain-impacted flow. Trans ASAE 36, 382 – 387. Le Bissonnais, Y., 1996. Aggregate stability and assessment of soil crusting and erodibility: 1. Theory and methodology. Eur. J. Soil Sci. 47, 425 – 437. Le Bissonnais, Y., Arrouays, D., 1997. Aggregate stability and assessment of soil crustability and erodibility: II. Application to humic loamy soils with various organic carbon contents. Eur. J. Soil Sci. 48, 39 – 48. Liebenow, A.M., Elliot, W.J., Laflen, J.M., Kohl, K.D., 1990. Interrill erodibility: collection and analysis of data from crop lands. Trans ASAE 24, 1882 – 1888. McIntyre, D.S., 1958. Permeability measurements of soil crust formed by raindrop impact. Soil Sci. 85, 158 – 189. Mermut, A.R., Luk, S.H., Ro¨mkens, M.J.M., Poesen, J.W.A., 1997. Soil loss by splash and wash during rainfall from two loess soils. Geoderma 75, 203 – 214. Meyer, L.D., 1981. How rain intensity affects interrill erosion. Trans ASAE 24, 1472 – 1475. Morin, J., 1993. Soil crusting and sealing. Soil Tillage in Africa: Needs and Challenges. FAO Soils Bull., vol. 69. FAO Information Division, Rome, Chapter 5. Morin, J., Goldberg, S., Seginer, I., 1967. A rainfall simulator with a rotating disk. Trans ASAE 10, 74 – 79. Morin, J., Benyamini, Y., Michaeli, A., 1981. The dynamics of soil crusting by rainfall impact and the water movement in the soil profile. J. Hydrol. 52, 321 – 335. Neaman, A., Singer, A., Stahr, K., 2000. Dispersion and migration of particles in two palygorskite containing soils of the Jordan Valley. J. Plant Nutr. Soil Sci. 163, 537 – 547.
Oades, J.M., Waters, A.G., 1991. Aggregate hierarchy in soils. Aust. J. Soil Res. 29, 815 – 820. Oster, J.D., Shainberg, I., Wood, J.D., 1980. Flocculation value and gel structure of Na/Ca montmorillonite and illite suspension. Soil Sci. Soc. Am. J. 44, 955 – 959. Quirk, J.P., 1978. Some physico-chemical aspects of soil structural stability: a review. In: Emerson, W.W., Bond, R.D., Dexter, A.R. (Eds.), Modification of Soil Structure. Wiley, New York, pp. 3 – 16. Romkens, M.J.M., Luk, S.H., Poesen, J.W.A., Mermut, A.R., 1995. Rain infiltration into loess soils from different geographic regions. Catena 25, 21 – 32. Schofield, R.K., Samson, H.R., 1954. Flocculation of kaolinite due to the attraction of oppositely charged crystal faces. Faraday Discuss. Chem. Soc. 18, 135 – 145. Schulze, D.G., 1989. An introduction to soil mineralogy. In: Dixon, J.B., Weed, S.B. (Eds.), Minerals in Soil Environments. Soil Sci. Soc. Am., Madison, pp. 1 – 34. Shainberg, I., Letey, J., 1984. Response of soils to sodic and saline conditions. Hilgardia 52, 1 – 57. Sharma, P.P., Gupta, S.C., Foster, G.R., 1995. Raindrop-induced soil detachment and sediment transport from interrill areas. Soil Sci. Soc. Am. J. 59, 727 – 734. Singer, A., 1994. Clay mineralogy as affecting dispersivity and crust formation in aridisols. In: Etchevers, J.D. (Ed.), Transactions of the 15th World Congress of Soil Science. Acapulco, Mexico. Int. Soc. Soil Sci. and Mexican Soc. Soil Sci., Acapulco, Mexico, vol. 8a. Six, J., Elliott, E.T., Paustian, K., 2000. Soil structure and soil organic matter: II. A normalized stability index and the effect of mineralogy. Soil Sci. Soc. Am. J. 64, 1042 – 1049. Stern, R., Ben-Hur, M., Shainberg, I., 1991. Clay mineralogy effect on rain infiltration, seal formation and soil losses. Soil Sci. 152, 455 – 462. van Olphen, H., 1977. An Introduction to Clay Colloid Chemistry, 2nd ed. Interscience Publications, New York. Wakindiki, I.I.C., Ben-Hur, M., 2002. Soil mineralogy and texture effects on crust micromorphology, infiltration and erosion. Soil Sci. Soc. Am. J. 66, 897 – 905. Warrington, D., Shainberg, I., Agassi, M., Morin, J., 1989. Slope and phosphogypsum effect on runoff and erosion. Soil Sci. Soc. Am. J. 53, 1201 – 1205. Watson, D.A., Laflen, J.M., 1986. Soil strength, slope, and rainfall intensity effects on interrill erosion. Trans ASAE 29, 98 – 102.