34 Kehl, Everding, Botschek, Skowronek DOI: 10.1002/jpln.200421382 J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Erosion processes and erodibility of cultivated soils in North Rhine-Westphalia under artificial rain§ I. Site characteristics and results of laboratory experiments Martin Kehl1*, Christoph Everding2, Johannes Botschek1, and Armin Skowronek1 1 2 Institut für Bodenkunde, Rheinische Friedrich-Wilhelms-Universität Bonn, Nuûallee 13, D-53115 Bonn, Germany Gröninger Straûe 7, D-48161 Münster, Germany Accepted September 20, 2004 PNSS P138/2B SummaryÐZusammenfassung Soil erosion by water causes substantial on-site degradation and off-site damages in the densely populated state of North Rhine-Westphalia (Germany). Measures of soil conservation should be adjusted to soil erodibilities and should be based on an understanding of the processes involved in water erosion including aggregate breakdown, rainsplash erosion, surface sealing, and soil loss. For a state-wide assessment of erosion processes and erodibilities, we tested representative cultivated soils of North Rhine-Westphalia in laboratory and field experiments with artificial rain. In the laboratory experiments described in this paper, rainsplash erosion, sealing susceptibility, and interrill erodibility of 25 topsoils filled in 0.5 m2 boxes were investigated. Results of different aggregate-stability tests correlate with organic-matter contents but not with parameters of rainsplash or soil loss. On most soil materials, rainsplash increases or maintains constant rates in the course of the simulation runs indicating that the soil surface did not attain a higher shear resistance. High sealing susceptibilities are found for soils of quite different textures ranging from loam sand to silt clay, whereas other silt clays, clay loams, and some clay silts maintain high infiltration rates. A trend of increasing sealing susceptibility and total soil loss with increasing clay content is observed for the loam sands to sand loams. Dynamics of soil loss is largely governed by runoff rates. Total soil loss is also determined by sediment concentration in surface runoff, which is low on most clayey soils, on loam sands poor in clay, and on a sand loam, and high in the case of highly erodible clay silts, loam sands, and sand loams. The most crust prone soils are not necessarily the most erodible. On most soils, soil-loss rates do not stabilize until the end of the rainfall experiments. For comparing the interrill erodibilities of the soils, total soil loss is preferred instead of interrill erodibility factors (Ki, Kiq) published in the literature. Key words: aggregate stability / rainsplash erosion / surface sealing / interrill erodibility / North Rhine-Westfalia * Correspondence: Dr. M. Kehl; e-mail: kehl@boden.uni-bonn.de § The article is partially based on the Ph.D. thesis by M. Kehl (1997): Experimentelle Laboruntersuchungen zur Dynamik der Wassererosion verschieden texturierter Ackerböden Nordrhein-Westfalens, and on the Ph.D. thesis by C. Everding (1998): Kennzeichnung des Erosionsverhaltens und der Erosionsanfälligkeit verschieden texturierter Ackerböden Nordrhein-Westfalens mit Hilfe von Feldberegnungen. Both: Rheinische Friedrich-Wilhelms-Universität Bonn, Germany. Erosionsverhalten und Erosionsanfälligkeit von Ackerböden in Nordrhein-Westfalen bei simulierten Starkregen I. Untersuchungsstandorte und Ergebnisse aus Laborexperimenten Bodenerosion durch Wasser verursacht erhebliche on- und off-site-Schäden im dicht besiedelten Bundesland NordrheinWestfalen. Maûnahmen des Bodenschutzes sollten an die Erosionsanfälligkeit der Böden angepasst werden und auf einem Verständnis der Teilprozesse der Wassererosion wie Aggregatzerfall, Regentropfenerosion, Oberflächenverschlämmung und Bodenabtrag beruhen. Für eine landesweite Einschätzung der Erosionsprozesse und der Erosionsanfälligkeiten wurden repräsentative Ackerböden NordrheinWestfalens in Labor- und Feldexperimenten mit künstlichem Starkregen untersucht. In den hier beschriebenen Laborversuchen wurden die Dynamik der Regentropfenerosion, die Verschlämmungsneigung und die Erodierbarkeit von 25 in Bodenkästen (0,5 m2) gefüllten Krumenmaterialien erfasst. Ergebnisse von Aggregatstabilitätstests ergeben eine mäûige Korrelation zum Humusgehalt der Krumenmaterialien, aber nicht zu Kennwerten der Regentropfenerosion oder des Bodenabtrags. Auf den meisten Krumenmaterialien steigen die Raten der Regentropfenerosion kontinuierlich an oder erreichen konstante Werte. Eine die Ablösung mindernde Scherfestigkeitssteigerung der Bodenoberfläche ist unwahrscheinlich. Hohe Verschlämmungsanfälligkeiten weisen Materialien unterschiedlicher Textur von Lehmsanden bis Schlufftonen auf, während andere Schlufftone, Tonlehme und einige Tonschluffe hohe Infiltrationskapazitäten aufrecht erhalten. Eine Tendenz steigender Verschlämmungsanfälligkeit und Bodenabträge mit zunehmenden Tongehalten zeigt sich in der Gruppe der Lehmsande bis Sandlehme. Der zeitliche Verlauf des Bodenabtrags wird i. W. durch die Dynamik der Abflussraten bestimmt. Die Gesamtabträge hängen auch von der mittleren Sedimentkonzentration ab, die als niedrig auf den tonreichen Böden, auf tonarmen Lehmsanden und auf einem Sandlehm und als hoch im Fall der erosionsanfälligen Tonschluffe, Lehmsande und Sandlehme einzuschätzen ist. Die am stärksten verschlämmungsanfälligen Böden sind dabei nicht zwangsläufig auch die erosionsanfälligsten. Auf den meisten Krumenmaterialien stabilisieren sich die Bodenabtragsraten nicht bis zum Ende der Beregnungsversuche. Für einen Vergleich der Erodierbarkeiten wird daher der Gesamtabtrag als besser angesehen als die in der Literatur zur Berechnung der Zwischenrillenerodierbarkeit publizierten Parameter (Ki, Kiq). ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1436-8730/05/0102-34 J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Erosion processes and erodibility of cultivated soils, Part I 1 Introduction About 11,250 km2 equaling 1/3 of the total land area of North Rhine-Westphalia, the most densely populated federal state of Germany, is under arable land use. Climatic conditions, topography, and agronomic factors often cause on-site degradation and off-site damages due to soil erosion by water. Because of a high diversity of soil parent materials and topographic conditions, the soils under cultivation have manifold physical and chemical properties. These define soil erodibilities that vary both on a regional and on a local scale (Botschek, 1991). The erodibility of a soil depends on more or less stable and on transient soil properties affecting aggregate stability, detachability of soil particles by rainsplash, sealing susceptibility, and soil loss. The stability of aggregates, for instance, is related to Corg content (Blackman, 1992; Roth, 1992), pH (Auerswald, 1995), aggregate size (Fohrer, 1995; Gäth et al., 1995), antecedent moisture content (Henk, 1989), and microbial activity, where microbial exudates, fungal hyphae, and humic substances promote the formation of aggregates (Tisdall and Oades, 1982; Oades, 1993). The kinetic energy of raindrops hitting the soil surface and fast wetting cause disintegration of aggregates. Rainsplash reflects this disintegration and supplies small particles for entrainment and transport in sheet flow. During a rainfall event, the infiltration capacity of a soil decreases as water content in the subsoil rises, and hydraulic gradient is reduced. Furthermore, soil particles resulting from aggregate breakdown clog pores, and compaction by raindrop impact reduces the porosity of the surface layer. A thin surface seal develops (Bresson and Cadot, 1992; Bielders and Baveye, 1995) with highly reduced water conductivity in comparison to the underlying soil (Morin and Benyamini, 1977; Mualem et al., 1990). Thus, the initiation and dynamics of surface runoff also depend on the sealing susceptibility of the soil surface. Soil loss by surface runoff integrates the effects of aggregate breakdown, rainsplash detachment, sealing, and particle entrainment. Under conditions of sheet flow, soil loss is dominated by ªraindrop induced flow transportº (RIFT; Kinnell, 1988), because stream power of surface runoff is low. In RIFT, the transport of particles decreases with the flow depth of surface runoff and increases with the amount of raindrops hitting the water film on the soil surface. On soils with a low initial surface roughness, fast aggregate breakdown, and leveling of the surface, runoff can spread over a larger area, and a higher amount of particles can be entrained than on soils with a coarse seed-bed and high aggregate stability, where the same amount of runoff is concentrated and attains larger flow depths. Soil loss by sheet flow is thus the result of different processes which likely are affected by diverse soil properties. However, simple relationships between different erosion processes exist, as, for instance, the close correlation between total surface runoff and soil loss reported by Le Bissonnais and Singer (1993) shows. The implementation of soil-conservation measures against water erosion is improved by knowledge about soil erodibil- 35 ities and about the different erosion processes involved. Our aim was to study rainsplash erosion, surface runoff, and related soil loss on cultivated soils covering a broad spectrum of different textures in order to obtain a representative assessment of soil erosion processes and soil erodibilities in North Rhine-Westfalia. In the laboratory experiments reported below, artificial rain was applied on soil material taken from plough layers and packed in soil boxes. The aim was to investigate the dynamics of rainsplash erosion, surface sealing, and soil loss under controlled conditions. Furthermore, topsoil properties governing different erosion processes should be identified. Another series of experiments with artificial rain was conducted in the field on plots under bare fallow. The results of these experiments are presented in a companion paper (Kehl et al., 2005). 2 Materials and methods 2.1 Soil and site characteristics Twenty-eight cultivated soils were selected for field and laboratory investigations, including four soils investigated by Botschek (1991). The soils developed in lithologically different parent materials and were classified as Haplic, Eutric, Dystric, or Stagnic Cambisols, Cutanic Luvisols, Regic Anthrosols, Haplic Regosol, Haplic Podsol, or Rendzic Leptosol (Tab. 1). According to numerous studies, soil texture is an important factor influencing soil erodiblity. Representative texture classes of the topsoils were compiled from the digital soil map of North Rhine-Westphalia in the scale of 1 : 50,000 (Geologisches Landesamt Nordrhein-Westfalen, 1971± 1995). Topsoils rich in silt that belong to the texture groups (and subgroups) clay silt (Lu, Ut4) and loam silt (Ut2, Ut3, Uls) of AG Boden (1994) cover approximately 53% of the total land area. Thirteen sites represent these topsoils of high erosion potential. Approximately 18% of the land area are covered by pure sands (Ss), which are often found on level terrain and are known to have a low erodibility. These were omitted in this study, whereas other sandy soils corresponding to loam sands (Sl2, Sl3, St2) and silt sands (Su2, Su3), which sum up to ~10% of the total land area, are represented by four experimental sites. Covering another ~9% of the state, standard loams (Lt2), silt clays (Lt3, Tu3), and clay loams (Lts) are represented by six sites. The remaining five soils under investigation belong to the texture groups of loam clay (Tu2), standard loam (Ls3, Ls4), or sand loam (Sl4). At most sites, stone cover did not exceed 5%, whereas a stone cover of 7%, 10%, 20%, 25%, and 30% was observed at the sites EN, BM, SO, HB, and MB, respectively. Differences in cropping history, tillage and fertilization practices, and soil-water regime in the soils under investigation are reflected in a range of Corg content from 3.9 to 22.1 g kg±1 in the topsoils of sites GF and ML, respectively (Tab. 1). The pH(CaCl2) values range from 5.3 in the sandy topsoil of RH to 7.7 in the reclaimed soil of GF. Whereas most soils are free of carbonate, the CaCO3 equivalent in the plough layers at the sites WG, HB, GF, and BP amounts to 14, 17, 68, and 74 g ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 36 Kehl, Everding, Botschek, Skowronek J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Table 1: Test sites with soil groups according to FAO (1998) and parent material of soil formation. Soil texture according to AG Boden (1994), grain-size distribution, and physical-chemical properties are given for the topsoils. The results of aggregate-stability tests were obtained on aggregates taken from the seed-bed of the field experiments described by Kehl et al. (2005). Tabelle 1: Versuchsstandorte mit Bodengruppen nach FAO (1998) und Ausgangsmaterialien der Bodenbildung. Bodenart (nach AG Boden, 1994), Korngröûenverteilung und physikalisch-chemische Kennwerte gelten für den Oberboden. Die Ergebnisse der Aggregatstabilitätstests wurden an Aggregaten gewonnen, die aus dem Saatbett der Feldversuche von Kehl et al. (2005) entnommen wurden. Code Soil unit Parent material a) Soil text. b) Sand Silt Clay pH Corg Stable aggregates c) ISd) PSe) air-dried pre-wetted ±1 (g kg ) (%) (ml (5 min)±1) AV Cutanic Luvisol Loess Ut4 25 752 223 7.0 11.5 68 98 54 59 BM Stagnic Cambisol Ut4 57 767 176 6.7 11.6 76 96 60 501 BP Regic Anthrosol Artificial loess deposits Ut4 22 781 197 7.6 5.6 9 89 33 21 CR Stagnic Cambisol Tu3 167 502 331 6.9 15.4 86 96 71 170 Lt2 306 425 269 6.8 16.5 91 97 70 18 Lu 250 547 203 7.4 16.2 64 97 57 353 GF g) Regic Anthrosol Artificial loess deposits Ut4 50 765 185 7.7 n.d. n.d. n.d. GS Colluvial deposit of loess Ut3 101 748 151 7.4 13.1 58 97 63 400 Limestone Tu2 48 486 466 7.4 17.0 n.d. n.d. n.d. n.d. Lts 509 232 259 7.1 8.0 65 94 48 110 CS EN Gelifluction deposit of loess and sandstone Glacial till \ clay marl Haplic Cambisol Marl f) Stagnic Cambisol Eutric Cambisol HB g) Rendzic Leptosol Gelifluction deposits of loess and sandstone 3.9 n.d. HLL Eutric Cambisol Glacial till / marl HLS Cutanic Luvisol Glacial till // marl Sl4 626 239 135 6.2 7.4 43 86 37 184 HV Haplic Cambisol Gelifluction deposits of loess and clay-stone / claysandstone Ls3 403 373 224 7.2 11.6 57 87 43 39 LB Haplic Cambisol Sandy loess Uls 353 522 125 6.4 7.3 36 97 62 147 LM Stagnic Cambisol Gelifluction deposits of loess and sandstone Ut3 42 821 137 6.6 11.9 63 97 61 708 MB Dystric Cambisol Stony* gelifluction deposits of loess, and clay-silt stone Lt2 223 493 284 5.8 15.5 65 99 46 49 ML Stagnic Cambisol Stony* gelifluction deposits of loess and slate Lu 143 626 231 6.7 22.1 64 98 51 609 NU f) Stagnic Cambisol Gelifluction deposits of loess and glacial till / marl Ls3 406 368 226 5.9 13.4 81 97 57 322 RB Loess // loessic sand // sandy marl Ut2 215 675 110 6.2 17.8 70 97 66 697 Aeolian sand // Gelifluction deposits // Clay Sl3 722 180 98 5.3 7.1 n.d. n.d. 48 281 Slu 394 455 151 5.5 19.8 92 93 68 585 Eutric Cambisol RH Dystric Cambisol RO Cutanic Luvisol Sandy loess / calcareous sand ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Erosion processes and erodibility of cultivated soils, Part I 37 Table 1: Continued. Tabelle 1: Fortsetzung. Code Soil unit Parent material a) Soil text. b) Sand Silt Clay pH Corg Stable aggregates c) ISd) PSe) air-dried pre-wetted (g kg±1) (%) (ml (5 min)±1) RÜ Haplic Cambisol Limestone and marl Tu3 66 626 308 6.2 15.7 36 98 39 13 SO Stagnic Cambisol Geliflucation deposits of loess and sandstone // sand- and silt-stone Lu 262 528 210 6.3 15.0 59 95 47 190 TB Plaggic Anthrosol Plaggen / sandy gelifluction deposits Sl3 581 308 111 6.1 14.8 68 88 50 676 Haplic Regosol Fluvial sand Sl3 666 227 107 5.9 8.2 73 83 50 522 Stagnic Cambisol Colluvial deposit of loess Ut4 48 749 203 6.5 12.6 45 90 52 101 WF Haplic Cambisol Gelifluction deposit of loess and slate / argillaceous slate Ut3 54 793 153 6.0 13.2 64 97 60 847 WG Eutric Cambisol Gelifluction deposit of loess and clay marl Tu3 101 512 387 7.4 13.6 83 98 60 131 Fluvial sand Su2 795 175 30 5.7 10.0 n.d. n.d. n.d. n.d. Mean 273 524 203 6.2 12.7 63 94 54 309 Med. 219 517 200 6.6 13.1 65 97 54 190 TE f) WE f) WM g) Haplic Podzol a) Change of parent material in the depth of <30 cm ( \ ), 30±70 cm ( / ), and 70±120 cm (//) * with more than 20% rock fragments in the upper 50 cm of the soil profile b) S, s: sand, sandy; U, u: silt, silty; L, l: loam, loamy; T,t: clay, clayey; 2: low, 3: moderate, 4: high amounts of S, U, L, or T as subdominant particle-size fraction c) Aggregate stability against raindrop impact, tested with air-dried and pre-wetted aggregates d) IS: Immersion stability as mean of immersion stabilities in 0%, 25%, and 50% of ethanol-water mixtures e) PS: Perkolation stability described as percolation rate within 5 min f) Experimental sites of Botschek (1991) g) No erosion tests in the laboratory n.d.: not determined (kg soil)±1, respectively (Everding, 1998). The cation exchange capacities (CEC) of the topsoils are between 5.4 and 31.4 cmolc kg±1 in WM and CS, respectively. The maximum CEC is partly caused by a high portion of smectite in the clay fraction, while the other soils are mostly dominated by Illite. With the exception of the sandy soils WM and RH, base saturation is higher than 50% and exchangeable bases are dominated by Ca2+ ions. Amounts of Mg2+ and K+ are less than 10%, and Na+ is negligible reaching less than 1% of the CEC (Everding, 1998). For presentation and discussion of results, the sites in Tab. 2 were grouped according to texture classes of the topsoils. Group I comprises the loam silts (Uls, Ut2, Ut3) and clay silts (Ut4), group II the loam sands (Sl3, Su2), sand loams (Sl4, Slu), and two standard loams (Ls3), and group III the clay silts (Lu), silt clays (Tu3), two standard loams (Lt2), a clay loam (Lts), and a loam clay (Tu2). 2.2 Soil chemical and physical analyses All standard laboratory analyses were carried out on the fraction of air-dried samples passed through a 2 mm sieve. The Corg content was measured spectrophotometrically after wet ashing with acidified K2Cr2O7 (Nelson and Sommers, 1982). The pH was determined in 0.01 M CaCl2 in a 1:2.5 soil : solution suspension (McLean, 1982). The calcium carbonate equivalent (CaCO3) was measured volumetrically after addition of HCl. The potential cation exchange capacity (CEC) ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 38 Kehl, Everding, Botschek, Skowronek J. Plant Nutr. Soil Sci. 2005, 168, 34±44 c) Finally, the decrease in percolation rate after Sekera and Brunner (1943) modified by Becher and Kainz (1983) was carried out with air-dried aggregates of 1±2 mm in diameter. As a measure of stability, the percolated volume of water was recorded after 5 min. was determined after extraction with 0.2 M BaCl2-Triethanolamin at pH 8.1 according to Mehlich (1942). Cations Ba2+, Ca2+, K+, and Na+ were quantified by inductively coupled plasma-optical emission spectrometry (ICP-OES), whereas Mg2+ was analyzed by flame atomic-absorption spectrometry (AAS). Further laboratory analysis included the determination of aggregate bulk density in the size fraction 6±8 mm (Becher et al., 1990). To study the disintegration of aggregates resulting from raindrop impact and fast wetting, the portion of particles with a diameter of <125 lm in the sealed soil surface (P125) was determined by wet sieving (Loch and Foley, 1994), with nine replicates. Aggregate stability was assessed on air-dried or pre-wetted aggregates taken from the seed-bed of the field plots at the times of the field experiments (Kehl et al., 2005). Three different tests of aggregate stability were conducted: a) The stability of aggregates to resist raindrop impact was tested with 10 g of aggregates of the size fraction 6± 8 mm, which were placed on sieves with 3 mm openings and exposed to 10 mm of artificial rain applied with an intensity of 40 mm h±1. A similar procedure has been proposed by Henk (1989) and Roth (1992). The percentage of ªrain-stable aggregatesº was measured gravimetrically as the oven-dried soil mass remaining on the sieves after rainfall and related to the original weight of the aggregates. This test was carried out in ten replicates each with air-dried and pre-wetted aggregates, which had moisture contents ranging from 12 to 34 g kg±1 and from 89 to 330 g kg±1, respectively (Kehl, 1997). Differences in airdried or pre-wetted stability gave evidence for the susceptibility of the soil aggregates towards air slaking. 2.3 Experimental setup of erosion experiments with soil boxes Soil material was taken from the seed-bed of the field experiments described by Kehl et al. (2005). Because of technical reasons, three sites (GF, HB, and WM) were not included in the laboratory experiments. The soil material was densely packed in four metal boxes (98 cm long, 50 cm wide, 30 cm deep) placed next to each other to allow exchange of soil particles by rainsplash (Fig. 1). The soil boxes were inclined to a slope of 9%. Air-dried aggregates from the seed-bed of the field trials were spread on the surface to prepare a ªseedbedº similar to the field experiments with mean weight diameters (MWD) in the laboratory trials between 3.0 and 10.0 mm. Before the simulation runs, the moisture content of the ªseed-bedº had a maximum of 61 g (kg soil)±1, and soil water pressure head in 15 cm soil depth ranged from ±53 to ±120 hPa, as measured with eight tensiometers. Runoff and sediment concentration were measured in 5 min time intervals to calculate soil-loss rates. The mean sediment concentration was derived by dividing total soil loss values by total runoff. Infiltration rates were calculated by subtracting the measured runoff rates from the average rainfall intensity b) The stability of air-dried aggregates (5 g of 4±6 mm in diameter) was measured after a 5 min immersion in distilled water mixed with 0%, 25%, and 50% ethanol, respectively. The percentage of ªimmersion-stableº aggregates was determined by carefully sieving the immersed samples through a 1 mm sieve and weighing the oven-dried size fraction >1 mm. This weight was related to the weight of the original sample. Q 1 - Q4 R1 - R4 a) + + ° ° ° ° + Surface runoff Exchange of splash particles + °+ R1 Q3 ° ° ° ° b) + + + Q2 Splashcups Tensiometer Drainage outlets R3 Q1 R2 Gutters to collect surface runoff Splashboards with gutters Upper soil layer ('seed-bed') Compacted soil layer Styropore plate with Fleece R4 R2 R4 + Q4 Drainage outlets ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Figure 1: Vertical view on soil boxes used for rain simulation tests in the laboratory. The four boxes were placed next to each other, to allow for exchange of soil particles by rainsplash. Figure 1 b shows a cross-section of the boxes filled with topsoil taken from the plough layers of agricultural fields. Abbildung 1: Skizze der für die Beregnungsexperimente im Labor verwendeten Bodenkästen. Die vier Kästen wurden nebeneinander gestellt, um den Austausch von Bodenmaterial durch Regentropfenerosion zu ermöglichen. Abbildung 1 b zeigt einen Querschnitt durch die mit Ackerkrumenmaterial gefüllten Kästen. J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Erosion processes and erodibility of cultivated soils, Part I applied. The time interval of 5 min was also used to collect rain-detached particles with splashboards (30 cm ” 50 cm) mounted at the long side of each soil box. The total amount of these particles is denoted as Sb. In addition, rain-detached particles were collected with test tubes with 2.3 cm2 openings buried in the soil boxes to have a measure of total rainsplash per unit area (Sc, with eight replicates). Data on rainsplash are means of three replicates, whereas runoff and soil loss measured are means of four replicates. As a measure of soil surface roughness, the length of a chain placed on the soil surface before (R0) and after (R1) the simulation runs was recorded (White, 1983). 2.4 Simulation of rainfall Artificial rain was applied with an intensity of 40 mm h±1 for 90 min in each simulation run using two similar rain simulators for field and laboratory trials equipped with Veejet 80 100 and Veejet 95 100 nozzles, respectively. Applying a pressure of 420 hPa, the artificial raindrops fell from a height of 2.5 m. This setup was tested for the field simulator to produce a drop-size distribution with a median diameter of 2.5 mm, a maximum drop size of 5.4 mm, and a kinetic energy of 23.7 J m±2 per mm of rainfall (Hassel and Richter, 1992) corresponding to a R-factor of 57 kJ m±2 mm h±1 (Wischmeier and Smith, 1978). R-factors of natural rainfall at the sites of the field experiments range from 30 to 80 kJ m±2 mm h±1 (Sauerborn, 1994; Everding, 1998). 3 Results and discussion 3.1 Aggregate stability and aggregate breakdown The topsoils exhibit a wide range of aggregate stabilities towards raindrop impact and fast wetting (Tab. 1). The percentage of air-dried and pre-wetted stable aggregates range from 9% to 92% and from 86% to 99%, respectively. Immersion stability is between 33% and 71%. Rain-stability and immersion-stability tests indicate linearly increasing aggregate stabilities with increasing Corg content. That trend is strongly reflected by the immersion stability in pure water (r2 = 0.32**, n = 24). The stabilizing effect of organic matter (OM) can be related to different mechanisms, among them the reduction of wetting velocities (Zhang and Hartge, 1992). Slow wetting reduces the compression of air in the pore space by inflowing water and the extent of air slaking (Gäth and Frede, 1995). This process is reflected in significantly higher rain stabilities of slowly pre-wetted aggregates than of air-dried ones (Tab. 1). While most pre-wetted aggregates remain stable after a 10 min rainfall, the air-dried ones break down to an extent which is negatively but not significantly correlated to OM content (r2 = 0.28, n = 24). A similar effect might be expected for aggregate bulk density, which ranges from 1.49 to 2.06 g cm±3 for RÜ and TE, respectively. Aggregates with high densities and low pore volume should be wetted slowly and should be more stable against air slaking in comparison to aggregates with low densities. The stability tests reported in this study do not support this hypothesis, though. 39 The results of the percolation-stability test range from 13 to 847 ml (5 min)±1. They neither correlate with the other tests nor with any of the measured soil properties. The close correlation of percolation stability with erodibility parameters as reported by Auerswald (1995) was not found. 3.2 Rainsplash erosion As shown in Fig. 2 a and b, the splash rates of most of the soils either steadily increase or reach a plateau which is kept until the end of the simulation run. These linear or plateaushaped dynamics (Tab. 2) indicate an increasing constantly high availability of splash-transportable particles. The soils BP, WE, and TE show declining splash rates after an early maximum had been reached. Declining splash rates were reported by Bradford et al. (1987), Farres (1987), Henk (1989), Moore and Singer (1990), and Roth (1992) and have been attributed to the formation of a dense surface seal with high shear resistance as described by Al-Durrah and Bradford (1982). In the case of topsoil TE, declining splash rates can be explained by the accumulation of large sand grains at the soil surface, that are too heavy to be transported onto the splash boards by air splash. In consecutive simulation runs applying the same amount of rainfall on a crusted and dried soil surface, a further increase of splash rates or constantly high rates have been recorded (Kehl, 1997). Thus, the formation of a dense, splash-decreasing surface seal on the soils is unlikely. Constant low rates of rainsplash erosion were observed for the silt clay of CR. These sum up to the smallest splash totals recorded (Sb or Sc in Tab. 2), which can be attributed to the high cohesion and high aggregate stability of this clay-rich soil. Assuming that aggregate breakdown of an air-dry seed-bed is most pronounced during the wetting phase with a high potential for air slaking, the increase of splash rates at the beginning of the simulation runs (DS value of Tab. 2) should be negatively correlated with parameters of aggregate stability like, e.g., the rain stability of air-dried aggregates (RSd), and with the surface roughness before the onset of rainfall (R0). This hypothesis is confirmed by the following multiple regression: DS = 3.38** ± 0.016* RSd ± 0.099* R0, r2 = 0.34 (1) However, a higher coefficient of determination is obtained if the percentage of fine aggregates (<1 mm in diameter, S01) in the upper soil layer (ªseed-bedº) and Corg are considered: DS = 1.0 + 0.033** S01 ± 0.652* Corg, r2 = 0.61 (2) Comparing the total splash rates measured with splashboards (Sb) between the textural groups, the silty soils have highest values (mean = 16.1 g (90 min)±1), followed by sandy soils (mean = 14.3 g (90 min)±1). Soils rich in clay (group III) show lowest Sb values (mean = 9.0 g (90 min)±1), which can be explained by their high aggregate stability and partly by stones covering the soil surface, which were not separated ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 40 Kehl, Everding, Botschek, Skowronek J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Rainsplash erosion (g) 2 2 aA b 1,6 1,6 1,2 1,2 0,8 0,8 0,4 0,4 0 0 0 15 30 45 60 75 0 90 15 30 45 60 75 90 30 45 60 75 90 -1 Surface runoff (mm h ) 40 40 c d 30 30 20 20 10 10 0 0 0 15 30 45 60 75 90 0 15 Soil loss (g m -2) 50 100 e f 40 80 30 60 20 40 10 20 0 0 0 15 30 45 60 75 90 0 15 30 45 60 75 time (min) time (min) RB WF BM HLS RO HV BP WE TE SO ML CR while filling the soil boxes. The mean weight diameter (MWD) of splash particles is highest in this group (mean = 0.30 mm), while the sandy soils (group II) have medium (mean = 0.28 mm) and the silty soils (group I) smallest splash particles (mean = 0.23 mm). Regression analyses were carried out to identify which textural and chemical soil properties are related to Sb or Sc, which are closely correlated (r2 = 0.51**, n = 24). A positive correlation of r2 = 0.43** (n = 25) was found between Sb and the f-index of Bloemen (1980). The f-index decreases with increasing clay content and with the homogeneity of particlesize composition in the soil. Therefore, it reflects the observaã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 90 Figure 2: Dynamics of rain splash (a, b), surface runoff (c, d), and soil loss (e, f) on 12 topsoil materials from North Rhine-Westphalia (results of laboratory experiments; bars represent standard deviations). Abbildung 2: Zeitlicher Verlauf der Regentropfenerosion (a, b), des Oberflächenabflusses (c, d) und des Bodenabtrags von zwölf Ackerkrumenmaterialien aus Nordrhein-Westfalen (Ergebnisse von Laborexperimenten; Fehlerbalken geben Standardabweichungen wieder). tion that soils rich in clay are more resistant against raindrop impact than silty or sandy soils. 3.3 Surface runoff and sealing susceptibility According to their infiltration behavior, eight soils in the present study have low sealing susceptibilities. They show slowly increasing runoff rates (Fig. 2 c and d) which result in total runoff volumes not exceeding ~1/6 of the total rainfall (60 mm, Tab. 2). In most cases, they have final infiltration rates of more than ~25 mm h±1 equaling approximately 2/3 of the rainfall intensity. Most of these soils (CR, CS, EN, ML, WG) belong to the texture group III of clayey soils. In addition, J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Erosion processes and erodibility of cultivated soils, Part I 41 Table 2: Parameters of rain-splash erosion, surface runoff, and soil loss (with standard deviations in parantheses). Tabelle 2: Kennzahlen der Regentropfenerosion, des Oberflächenabflusses und des Bodenabtrags (mit Standardabweichungen in Klammern). Splash DSb) dynamicsa) g Sb c) g (90 min)±1 Sc d) MWDe) kg m±2 (90 min)±1 mm 1.50 (0.37) 0.26 Total surface runoff I 1000f) l m±2 mm h±1 Final infiltration rateg) g l±1 Kiqi) Total soil loss ” 10±6 g m±2 kg s±1 m±4 Sediment concentr.h) Group I (silty soils) AV plateau 0.33 17.3 (1.4) 12.1 (1.1) 26.7 21.5 (1.0) 16.8 (0.7) 204 (8) 3.5# (0.2) BM plateau 0.39 20.8 (3.8) 1.98 (0.55) 0.29 14.2 (1.6) 24.9* 20.1 (1.6) 14.4 (1.4) 205 (2) 3.3 (0.5) BP decreasing 0.57 18.5 (3.3) 1.99 (0.34) 0.18 17.7 (2.6) 21.1* 17.3 (2.4) 20.4 (1.0) 361 (58) 4.4 (0.3) GF n.d. n.d. GS linear 0.30 n.d. 17.8 (3.1) n.d. n.d. 1.22 (0.29) 0.19 n.d. 8.8 n.d. (2.6) 31.0 n.d. 26.4 n.d. (3.1) 25.3 (4.2) n.d. 222 (90) n.d. 6.2# (1.0) LB linear 0.41 17.3 (2.5) 2.54 (0.86) 0.21 24.6 (4.6) 13.0* 11.0 (2.1) 31.2 (9.1) 769 (375) 8.7 (1.3) LM plateau 0.30 12.1 (0.9) 1.02 (0.34) 0.21 14.4 (1.5) 24.6 17.6 (1.9) 11.4 (2.8) 164 (52) 2.3 (1.0) RB linear 0.08 10.7 (0.7) 1.48 (0.26) 0.21 28.3 (2.5) 11.9* 10.8 (1.8) 14.0 (1.8) 396 (18) 4.3# (0.7) WE decreasing 0.39 15.5 (2.0) 1.83 (0.37) 0.24 12.2 (1.4) 26.5* 20.7 (2.5) 15.8 (0.6) 193 (28) 3.6 (0.2) WF plateau 0.30 14.7 (2.3) 2.16 (0.87) 0.31 21.2 (3.8) 17.3* 14.4 (3.6) 14.0 (1.5) 298 (81) 3.3 (0.6) Group II (sandy soils) HLL linear 0.16 13.7 (0.8) 2.05 (0.56) 0.26 8.1 (1.6) 27.4 22.8 (2.2) 15.4 (2.4) 126 (40) 4.0# (0.4) HLS plateau 0.19 15.3 (1.3) 1.83 (0.39) 0.23 16.5 (1.8) 21.0* 16.1 (2.1) 16.8 (2.7) 277 (26) 3.4 (0.3) HV plateau 0.52 16.0 (3.2) 1.80 (0.29) 0.31 33.4 (2.1) 7.9* 6.3 (1.0) 18.3 (2.7) 610 (73) 3.9 (0.4) # NU linear 0.09 10.5 (1.6) 1.41 (0.26) 0.33 6.5 (0.3) 35.2 25.2 (0.6) 6.7 (0.9) 44 (7) 1.7 (0.3) RH plateau 0.21 10.5 (1.7) 1.64 (0.13) 0.32 14.6 (5.2) 23.8 21.0 (4.4) 9.6 (0.5) 140 (55) 2.3 (0.3) RO linear 0.30 15.9 (3.3) 1.49 (0.24) 0.18 24.3 (2.2) 15.9* 15.9 (1.8) 32.8 (3.3) 791 (79) 9.9 (0.7) TB linear 0.63 17.0 (0.9) 1.96 (0.47) 0.25 14.3 (6.7) 24.8* 23.0 (6.2) 32.3 (7.4) 461 (264) 7.1 (2.0) TE decreasing 0.71 15.5 (2.0) 1.57 (0.26) 0.36 28.5 (0.5) 11.7* 10.6 (0.2) 8.3 (0.4) 236 1.7# (0.1) WM n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. (10) n.d. n.d. Group III (clayey soils') CR constant ±0.03 4.4 (0.2) 0.31 (0.11) 0.33 0.4 (0.4) >40.0 38.6 (1.8) 7.6 (2.8) 3 (4) 0.9# (1.1) CS linear 0.11 8.3 (0.1) n.d. 0.31 0.0 (0.0) >40.0 >40.0 (0.0) 0.0 (0.0) 0 (0) 0.0 (0.0) EN plateau 0.16 9.9 (0.5) 0.99 (0.25) 0.22 6.3 (1.9) (3.8) 10.5 (2.1) 66 (24) 2.7# HB n.d. n.d. n.d. n.d. MB plateau 0.08 8.7 (0.3) 0.98 (0.27) 0.17 13.4 (2.7) 28.1 17.3 (3.3) 12.6 (1.0) 168 (45) 2.9# (0.2) ML plateau 0.06 9.0 (0.7) 1.13 (0.21) 0.39 5.1 (0.9) 35.0 30.6 (2.5) 7.6 (3.3) 39 (20) 2.0# (0.6) RÜ plateau 0.38 12.9 (2.5) 1.88 (0.68) 0.34 24.4 (0.9) 14.7* 12.8 (1.2) 10.7 (2.8) 261 (59) 2.0 (0.3) # n.d. n.d. n.d. 35.4 n.d. 29.1 n.d. n.d. n.d. (0.7) n.d. SO linear 0.13 9.5 (1.9) 1.07 (0.21) 0.35 11.4 (2.3) 28.8 22.4 (1.9) 8.6 (2.1) 98 (6) 2.0 (0.5) WG linear ±0.02 9.3 (2.0) 1.57 (0.33) 0.28 3.6 (0.7) 36.4 32.7 (1.3) 11.6 (6.4) 42 (17) 3.1# (1.6) a) Course of splash rates during the simulation runs: linear = linearly increasing; plateau = reaching a plateau; decreasing = reaching a maximum rate and decreasing afterwards; constant = constant rates b) Difference between rain splash of the fourth and the first time interval (15th±20th and 0th±5th min) c) Arithmetic mean of total rain splash measured with splash boards d) Total rain splash measured with splash cups, median of eight replicates with standard deviation of arithmetic means e) Mean weight diameter of rain-detached particles collected with splashboards f) Infiltration rate after a cumulative rainfall kinetic energy of 1 000 kJ m±2 (= 42 mm of rainfall), where soils marked with an asterisk (*) can be fitted to the infilration model of Morin and Benyamini (1977) yielding coefficients of determination >0.8 g) Difference between rainfall and surface runoff as a mean of the last three time intervals (80th±90th min) of the simulation runs h) Sediment concentration = total soil loss / total surface runoff i) Interrill erodibility according to Kinnell (1993a). The Kiq values have been calculated with the mean of soil loss rates from the 75th to the 90th min. The sites marked with # did not reach steady soil-loss rates at the end of the simulation runs. See text for further explanations. n.d.: not determined ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 42 Kehl, Everding, Botschek, Skowronek J. Plant Nutr. Soil Sci. 2005, 168, 34±44 the loam silt of GS (group I), the clay loam of HLL, and the standard loam of NU (both group II) fall into this class of low sealing susceptibility. The soil of NU has an almost identical grain-size distribution compared to the sandy loam of HV, but a much higher infiltrability. This difference can be attributed to smaller particles resulting from aggregate breakdown on HV that can form a denser surface seal (Kehl, 1997). Under the experimental conditions chosen, the transport capacity of surface runoff is low reaching a maximum stream power (Bagnold, 1977; Huang, 1995) of 0.009 W m±2. In this range, shear velocity is not sufficient for the formation of rills (Rose, 1988), and particle transport is mainly driven by RIFT (Kinnell, 1988). Moderate sealing susceptibilities can be ascribed to soils with total surface runoff of more than ~10 mm and less than ~20 mm (equaling to 1/6 and 1/3 of total rainfall, respectively) and final infiltration rates between 1/3 and 2/3 of rainfall intensity (corresponding to ~15 mm h±1 and 25 mm h±1). This class is composed mainly of silty soils and some sandy soils with linear infiltration decreases (AV, LM, MB, SO) or with a Hortonian type of infiltration behavior (BM, BP, WE, HLS, RH, TB). The latter can be described with the infiltration model of Morin and Benyamini (1977) and is characteristic for the class of soils exhibiting highest sealing susceptibilities with total runoff volumes exceeding 1/3 of the applied rainfall and final infiltration rates of less than 1/3 of applied rainfall intensity. Seven soils with quite different textural and chemical properties fall within this category (LB, RB, WF, HV, RO, TE, RÜ). Some of them have lower aggregate stabilities or higher splash rates compared to less susceptible soils of the same texture group. Surprisingly, a high sealing susceptibility has also been recorded for the silty clay of RÜ (texture group III), which might be attributed to its comparatively low aggregate stability. 3.4 Sediment concentration, soil loss, and interrill erodibility Roth (1992) proposed to use the infiltration rate at a cumulative kinetic rainfall energy of 1000 J m±2 (I1000 in Tab. 2) as a parameter for sealing susceptibility. In general, the soils of the present study exhibit higher I1000-values and therefore lower sealing susceptibilities than the silty soils from loess and sand loams or loam sands from glacial till tested by Roth et al. (1995). There is a similar trend of lower infiltration rates in sand loams than in loam sands. The close positive correlation between I1000 and the f-index of Bloemen (1980) reported by Roth et al. (1995) was not found, though. In contrast, most of the soils with small f-indices exhibit low sealing susceptibilities (final infiltration rates >25 mm h±1) in our investigation. This contradiction can be attributed to the broader textural spectrum and probably to the higher aggregate stabilities of the soils investigated in this study. The percentages of fine particles with diameters <125 lm in the surface seal (P125) varied from 32% to 71% in TE and BP, respectively, indicating a trend of increasing total runoff with rising P125. According to Loch and Foley (1994), steady infiltration rates of soils as calculated from runoff rates and hydraulic conductivities of surface seals show a very close negative correlation with P125. This relationship was established in erosion trials with artificial rainfall of high intensity (100 mm h±1) and high kinetic energy (29.5 J m±2 mm±1) where all soils reached constant infiltration rates, and compaction of the soil surface by raindrop impact was high. In the present study, most soils did not attain constant infiltration rates, and aggregate breakdown was presumably less intense. ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Sediment concentration in surface runoff of most soils changed significantly with time (data not shown). Several soils, including some with a high sealing susceptibility (HV, TE, RÜ), displayed declining sediment concentrations subsequent to an early maximum. Sediment concentrations of other soils either reached a maximum that was maintained till the end of the experiment or increased steadily. On most clayey soils, sediment concentrations remained at a more or less constant low level. The dynamics of sediment concentration is not correlated to those of surface runoff. Instead, there is some connection to the rates of rainsplash erosion. The mean sediment concentration (Tab. 2) displays a close correlation to the total amount of splash particles collected with splashboards (Sb, r2 = 0.45**, n = 25). The dynamics of soil loss (Fig. 1) is mainly controlled by surface runoff. Total soil loss relates to mean sediment concentration (r2 = 0.68**, n = 25) and to total runoff (r2 = 0,60**, n = 25). In exceptional cases only, a comparatively high soil loss in combination with a moderate or low total runoff (TB, GS) or a low soil loss with high runoff values (TE), respectively, is recorded. In the latter case, this can be attributed to the enrichment of medium and coarse sand at the soil surface, which could not be transported by thin sheet flow. On this soil only, soil loss decreased in the course of the simulation run though runoff rates remained high. As a measure of soil erodibility, the interrill erodibility Kiq proposed by Kinnell (1993a) was calculated based on the soil loss and surface-runoff rates of the last three time steps of the erosion tests. Kiq considers the influence of differential infiltration behavior of the soils by including the runoff rate Q in the equation Kiq = Di / I Q Sf where Di is the mean of the interrill-erosion rates, I the rainfall intensity, and Sf the slopecorrection factor of the Water Erosion Prediction Project (WEPP, cf., Huang and Bradford, 1993). Sf takes a constant of 0.456 for the 9% slope used in the experiments. Thus, the Kiq values reflect the sediment concentration at the end of the simulation runs, which is a less dynamic property than the soil-loss rates used for calculation of interrill erodibilities of WEPP (Laflen et al., 1991). Most of the measured Kiq are slightly lower than those reported by Kinnell (1993b), while Kiq of TB, LB, and RO exceed the published values substantially. The Kiq values give a different ranking of the soils than if total soil loss (Tab. 2) would be taken as a measure of interrill erodibility. This effect can be illustrated by comparing the soils of GS and HV which have a total soil loss and Kiq of 222 g and 6.2 ” 10±6 kg s±1 m±4 or 610 g and 3.9 ” 10±6 kg s±1 m±4, respectively. The relationship between Kiq and total soil J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Erosion processes and erodibility of cultivated soils, Part I loss is close (r2 = 0.71***, n = 25) and can be improved, if soils not having reached stable soil-loss rates are excluded (r2 = 0.78***, n = 12). 4 Conclusions The dynamics of rainsplash erosion, sealing, and soil loss recorded in the experiments are diverse. Rainsplash dynamics indicates that on most topsoils, the formation of a surface seal with splash-decreasing shear resistance is unlikely. About 1/3 of the soils showed a Hortonian-type infiltration curve reaching low final infiltration rates. Only these soils are prone to sealing under the experimental conditions chosen. Splash totals are positively correlated with sediment concentrations in surface runoff, whereas soil loss relates to sediment concentration and to the amount of surface runoff. Soil loss is thus limited by the availability of fine particles resulting from raindrop impact or slaking and, possibly more important, by the availability and possibly aerial distribution of the transporting-agent surface runoff. In this context, the flow pattern of surface runoff should be monitored in future studies, because it defines the number of raindrops hitting the flow and triggering the entrainment of particles in RIFT. Relationships between the experimentally derived erodibility parameters and stable soil properties are observed, e.g., of increasing aggregate stability with humus content, or between splash totals and the f-index of Bloemen (1980), which might be helpful in describing the erodibility of these and of other soils. However, the variance recorded could not be explained to a sufficient degree. This is partly caused by the large variation of soil characteristics and might be improved, if a larger number of soils within textural groups is investigated. Since many of the clay silts, clay loams, and silt clays have proven to be sealing-resistant and less erodible, the focus should be on the silty soils which show medium to high sealing susceptibilities and soil losses. Furthermore, the group of loam sands and sand loams exhibit a very diverse erosion behavior, which deserves further research. For a relative ranking of soil erodibilities, total soil loss is preferred instead of interrill-erodibility parameters published in the literature, because most soils did not reach constant soilloss rates at the end of the experiments. Acknowledgments We are grateful to many colleagues, technicians, master students, and research assistants and last but not least to the farmers for their assistance in conducting field and laboratory rainfall experiments. The Institute of Agricultural Engineering of the University of Bonn provided the rain simulators including transport to and from the experimental sites. The Ministerium für Umwelt, Raumordnung und Landwirtschaft Nordrhein-Westfalen (MURL-NRW) generously financed our research within the project ªErosionsgefährdete Böden in Nordrhein-Westfalenº. 43 References AG Boden (1994): Hannover. Bodenkundliche Kartieranleitung. 4th ed., Al Durrah, M. M., Bradford, J. M. (1982): The mechanism of raindrop splash on soil surfaces. Soil Sci. Soc. Am. J. 46, 1086±1090. Auerswald, K. (1995): Percolation stability of aggregates from arable topsoils. Soil Sci. 159, 142±148. Bagnold, R. A. (1977): Bedload transport by natural rivers. Water Resour. Res. 13, 303±311. Becher, H. H., Kainz, M. (1983): Auswirkungen einer langjährigen Stallmistdüngung auf das Bodengefüge im Löûgebiet bei Straubing. Z. Acker- Pflanzenbau 152, 152±158. Becher, H. H., Müller, K., Schneider, E. (1990): Bestimmung der Dichte kleiner Einzelaggregate durch Tauchwägung nach Paraffintränkung. Z. Pflanzenernähr. Bodenkd. 153, 369±371. Bielders, C. L., Baveye, P. (1995): Processes of structural crust formation on coarse-textured soils. Eur. J. Soil Sci. 46, 221±232. Blackman, J. D. (1992): Seasonal variation in aggregate stability of downland soils. Soil Use Manage. 8, 142±150. Bloemen, G. W. (1980): Calculation of hydraulic conductivities of soils from texture and organic matter content. Z. Pflanzenernähr. Bodenkd. 143, 581±605. Botschek, J. (1991): Bodenkundliche Detailkartierung erosionsgefährdeter Standorte in Nordrhein-Westfalen und Überprüfung der Bodenerodierbarkeit (K-Faktor). Hamburger Bodenkundl. Arb. 16, 1±130. Bradford, J. M., Ferris, J. E., Remley, P. A. (1987): Interrill soil erosion processes: I. Effect of surface sealing on infiltration, runoff, and soil splash detachment. Soil Sci. Soc. Am. J. 51, 1566±1571. Bresson, L. M., Cadot, L. (1992): Illuviation and structural crust formation on loamy temperate soils. Soil Sci. Soc. Am. J. 56, 1565±1570. Everding, C. (1998): Kennzeichnung des Erosionsverhaltens und der Erosionsanfälligkeit verschieden texturierter Ackerböden Nordrhein-Westfalens mit Hilfe von Feldberegnungen. Bonner Bodenkundl. Abh. 24, 1±195. FAO (1998): World Reference Base for soil resources. World Soil Resour. Reports 84. Farres, P. J. (1987): The dynamics of rainsplash erosion and the role of aggregate stability. Catena 14, 119±130. Fohrer, N. (1995): Auswirkungen von Bodenfeuchte, Bodenart und Oberflächenbeschaffenheit auf Prozesse der Flächenerosion durch Wasser. Bodenökologie und Bodengenese 19, 1±183. Gäth, S., Frede, H.-G. (1995): Mechanisms of air-slaking, in Hartge, K.-H., Stewart, B. A. (eds.): Soil Structure ± Its Development and Function. Advances in Soil Science. CRC-Lewis Publishers, Boca Raton, pp. 159±174. Gäth, S., Görk, H., Frede, H.-G. (1995): Soil surface roughness as the result of aggregate size distribution. 2. Report: changes in aggregate size classes caused by erosive rainfalls. Z. Pflanzenernähr. Bodenkd. 158, 37±41. Geologisches Landesamt Nordrhein-Westfalen (1971±1995): Bodenkarte von Nordrhein-Westfalen 1 : 50 000. Krefeld. Hassel, J. M., Richter, G. (1992): Ein Vergleich deutscher und schweizerischer Regensimulatoren nach Regenstruktur und kinetischer Energie. Z. Pflanzenernähr. Bodenkd. 155, 185±190. Henk, U. (1989): Untersuchungen zur Regentropfenerosion und Stabilität von Bodenaggregaten. Landschaftsgenese und Landschaftsökologie 15, 1±197. ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 44 Kehl, Everding, Botschek, Skowronek J. Plant Nutr. Soil Sci. 2005, 168, 34±44 Huang, C. (1995): Empirical analysis of slope and runoff for sediment delivery from interrill areas. Soil Sci. Soc. Am. J. 59, 982±990. Morin, J., Benyamini, Y. (1977): Rainfall infiltration into bare soils. Water Resour. Res. 13, 813±817. Huang, C., Bradford, J. M. (1993): Analyses of slope and runoff factors based on the WEPP erosion model. Soil Sci. Soc. Am. J. 57, 1176±1183. Mualem, Y., Assouline, S., Rohdenburg, H. (1990): Rainfall induced soil seal. Critical review of observations and models. Catena 17, 185±203. Kehl, M. (1997): Experimentelle Laboruntersuchungen zur Dynamik der Wassererosion verschieden texturierter Ackerböden Nordrhein-Westfalens. Bonner Bodenkundl. Abh. 21, 1±120. Nelson, D. W., Sommers, L. E. (1982): Organic carbon, in Page, A. L., Miller, R. H., Keeney, D. R. (eds.): Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties. 2nd ed., ASA, Madison, WI., pp. 570±571. Kehl, M., Everding, C., Botschek, J., Skowronek, A. (2005): Erosion processes and erodibility of cultivated soils in North Rhine-Westphalia under artificial rain. II. Results of field experiments and comparison to laboratory trials. J. Plant Nutr. Soil Sci. 168, 45±52. Kinnell, P. I. A. (1988): The influence of flow discharge on sediment concentrations in raindrop induced flow transport. Aust. J. Soil Res. 26, 375±382. Kinnell, P. I. A. (1993a): Runoff as a factor influencing experimentally determined interrill erodibilities. Aust. J. Soil Res. 31, 333±342. Kinnell, P. I. A. (1993b): Interrill erodibilities based on the rainfall intensity ± flow discharge erosivity factor. Aust. J. Soil Res. 31, 319±332. Laflen, J. M., Elliott, W. J., Simanton, J. R., Holzkey, C. S., Kohl, K. D. (1991): WEPP ± Soil erodibility experiments for rangeland and cropland soils. J. Soil Water Conservation 49, 36±42. Le Bissonnais, Y., Singer, M. J. (1993): Seal formation, runoff, and interrill erosion from seventeen California soils. Soil Sci. Soc. Am. J. 57, 224±229. Loch, R. J., Foley, J. L. (1994): Measurement of aggregate breakdown under rain: comparison with tests of water stability and relationships with field measurements of infiltration. Aust. J. Soil Res. 32, 701±720. McLean, E. O. (1982): Soil pH and lime requirement, in Page, A. L., Miller, R. H., Keeney, D. R. (eds.): Methods of Soil Analysis, Part 2, Chemical and Microbiological Properties. 2nd ed., ASA, Madison, WI., pp. 199±224. Mehlich, A. (1942): Rapid estimation of base-exchange properties of soil. Soil Sci. 53, 1±14. Moore, D. C., Singer, M. J. (1990): Crust formation effects on soil erosion processes. Soil Sci. Soc. Am. J. 54, 1117±1123. ã 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Oades, J. M. (1993): The role of biology in the formation, stabilization, and degradation of soil structure. Geoderma 65, 377±401. Rose, C. W. (1988): Research progress on soil erosion processes and a basis for soil conservation practices, in Lal, R. (ed.): Soil Erosion Research Methods. Soil Water Conservation Soc., Ankeny, pp. 119±138. Roth, C. H. (1992): Die Bedeutung der Oberflächenverschlämmung für die Auslösung von Abfluû und Abtrag. Bodenökologie und Bodengenese 6, 1±179. Roth, C. H., Helming, K., Fohrer, N. (1995): Oberflächenverschlämmung und Abfluûbildung auf Böden aus Löû und pleistozänen Sedimenten. Z. Pflanzenernähr. Bodenkd. 158, 43±53. Sauerborn, P. (1994): Die Erosivität der Niederschläge in Deutschland ± Ein Beitrag zur quantitativen Prognose der Bodenerosion durch Wasser. Bonner Bodenkundl. Abh. 13, 1±189. Sekera, F., Brunner, A. (1943): Beiträge zur Methodik der Gareforschung. Z. Pflanzenernähr. Bodenkd. 29, 169±212. Tisdall, J., Oades, J. (1982): Organic matter and water stable aggregates in soils. J. Soil Sci. 33, 141±163. White, E. M. (1983): Simple measurements of soil surface microrelief characteristics. Water Res. Bull. 19, 231±233. Wischmeier, W. D., Smith, D. D. (1978): Predicting rainfall erosion losses. A guide to conservation planning. USDA Agric. Handbook No. 537. Zhang, H., Hartge, K. H. (1992): Zur Auswirkung organischer Substanz verschiedener Humifizierungsgrade auf die Aggregatstabilität durch Reduzierung der Benetzbarkeit. Z. Pflanzenernähr. Bodenkd. 155, 143±151.