How rain-formed soil crust affects wind erosion in a semi-arid steppe in northern China

How rain-formed soil crust affects wind erosion in a semi-arid steppe in northern China

Geoderma 249–250 (2015) 79–86 Contents lists available at ScienceDirect Geoderma journal homepage: www.elsevier.com/locate/geoderma How rain-formed...

595KB Sizes 2 Downloads 97 Views

Geoderma 249–250 (2015) 79–86

Contents lists available at ScienceDirect

Geoderma journal homepage: www.elsevier.com/locate/geoderma

How rain-formed soil crust affects wind erosion in a semi-arid steppe in northern China Yuchun Yan a,b,c,⁎, Lianhai Wu d, Xiaoping Xin a,b,c,⁎⁎, Xu Wang a,b,c, Guixia Yang a,b,c a

Hulunber State Station of Grassland Ecosystem Field Observation and Scientific Research, PR China Key Lab of Resources Remote Sensing and Digital Agriculture, Ministry of Agriculture, PR China Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, PR China d Sustainable Soils and Grassland Systems Department, Rothamsted Research, North Wyke, Okehampton EX20 2SB, UK b c

a r t i c l e

i n f o

Article history: Received 23 December 2014 Received in revised form 2 March 2015 Accepted 7 March 2015 Available online 17 March 2015 Keywords: Crust thickness Soil texture Soil organic carbon Wind erosion Simulated rainfall

a b s t r a c t There have been few studies on the formation and resistance of physical crusts to wind erosion for typical steppe soils in Inner Mongolia, China. The objectives of this study were to 1) examine the effects of rainfall quantity on soil crust thickness, 2) investigate the effects of soil crust on wind erosion, 3) determine the crust thickness (crust formed by various rainfall quantities) able to most effectively resist wind erosion, and 4) evaluate the differences between the responses of soils with different treatment histories to crust formation and subsequent wind erosion at given rainfall quantities. To this end, we simulated five light rainfall levels to investigate the impact of light rainfall on soil crusting and subsequent wind erosion for soils of a semi-arid steppe via a unique approach. The results show that the soil crust thickness increases linearly with an increasing amount of rainfall for all four soils. The soil crust formed by rainfall of more than 0.5 mm was able to nearly completely prevent wind erosion during the experimental period; soil losses of only 0.1–2.4% were observed for the high rainfall treatments (N0.5 mm) for all four soils. In contrast, soil losses of 9.4–33.1% occurred in the non-rainfall treatments for the four soils. The results show that the soil loss ratio increased with increasing clay plus silt content and SOC content for the non-rainfall treatment and 0.2 mm rainfall treatment. © 2015 Elsevier B.V. All rights reserved.

1. Introduction A physical crust is one of the major soil structural features in many arid and semi-arid regions of the world. The crusting process begins with the breakdown of aggregates and dispersion of clay when the soils are wetted or exposed to rainfall. As the soils dry after clay dispersion, a thin seal or skin forms (B.C. Feng et al., 2013; G. Feng et al., 2013). The presence of a physical soil crust alters many characteristics of the soil surface; thus, this crust plays an important role in many ecosystem functions. Soil crusts are generally undesirable because they can reduce water infiltration and increase runoff (Belnap, 2001), but they can also be favorable due to their ability to reduce evaporation by capping the soil surface and reducing porosity (Chamizo et al., 2011) in agricultural areas. However, in arid and semi-arid rangeland ecosystems, soil crusts play a critical role in conserving soil resources in regions where wind erosion is predominant over water erosion (Yan et al., 2013). The influence of crusts on erosion has been noted for a long time in many arid regions (Rajot et al., 2003).

⁎ Correspondence to: Y. Yan, Hulunber State Station of Grassland Ecosystem Field Observation and Scientific Research. ⁎⁎ Correspondence to: X. Xin, Key Lab of Resources Remote Sensing and Digital Agriculture, Ministry of Agriculture, PR China. E-mail addresses: [email protected] (Y. Yan), [email protected] (X. Xin).

http://dx.doi.org/10.1016/j.geoderma.2015.03.011 0016-7061/© 2015 Elsevier B.V. All rights reserved.

The existing wind erosion models do not include the effect on erosion of crusts formed by light rainfall, although certain studies have reported the effect of rainfall characteristics on soil crust formation (Farres, 1978; Morrison et al., 1985; B.C. Feng et al., 2013; G. Feng et al., 2013). Usually, only events with more than 10 mm of rainfall are taken into account in these models (Fryrear et al., 2000; Hagen et al., 1995). In view of the lack of studies on the effect of light rainfall in soil crust formation, G. Feng et al. (2013) conducted a field experiment involving 5 soil types on the Columbia Plateau to evaluate this effect and found an increase in crust thickness and strength with increased rainfall amounts. The crust thicknesses were similar among the soils, whereas the crust strengths varied according to the amount of rainfall. Furthermore, the soil crust strength increased with increasing soil clay plus silt contents. Although these findings provided a detailed relationship between crust thickness/crust strength and the quantity of light rainfall, the resistance of the crusted soils to erosion was not measured further. Soil surface seals or crusts often form in unconsolidated soils during rainfall. Crusts have a pronounced effect on the susceptibility of soils to wind erosion because their properties differ from those of unconsolidated soils (Zobeck, 1991). Soil crusts are characterized by an interlocking network of particles, and soil pore spaces become clogged with fine particles that are more compact and mechanically stable, which decreases or eliminates the availability of loose erodible material for saltation

80

Y. Yan et al. / Geoderma 249–250 (2015) 79–86

(Chepil, 1958; B.C. Feng et al., 2013; G. Feng et al., 2013). Furthermore, this crust or seal can protect the unconsolidated soils beneath them, consequently reducing the susceptibility of the soils to erosion when exposed to wind (Zobeck, 1991; Chepil, 1958). This process is particularly important in arid and semi-arid regions where vegetative cover may not sufficiently protect soils from erosion. Zobeck (1991) found that crusts formed on silt loam and clay soils can be much more effective in reducing the total soil erosion under various abrasion conditions than can those formed on sandy loam soil. In addition, soil crusts increase the entrainment threshold, as has been suggested by the large amount of work that has been conducted on how crust disturbances increase dust emission (Belnap and Gillette, 1997; Baddock et al., 2011). Although the soil crust effects on erosion have been quantified in the past (Chepil, 1953, 1958; Zobeck, 1991; Rajot et al., 2003), a majority of that research has focused on crusts formed by heavy rainfall, i.e., precipitation N10 mm, and the effects of soil crusts on abrasion flux. These results do not include quantitative estimates of the effect of a weak crust formed by light rainfall on erosion. In addition, soil erodibility has been observed to change dramatically under small quantities of rainfall (Yan et al., 2013). Therefore, it is important to investigate the effects of light rainfall on soil crust formation in many arid and semi-arid regions. Wind erosion commonly occurs in arid and semi-arid regions where vegetative cover is sparse and may not sufficiently protect soils. This type of erosion has been accelerated by the clearing of vegetation for rangelands (Webb et al., 2012). Over the past decades, wind erosion has been identified as a primary reason for soil degradation in the semi-arid steppes of northern China (Yan et al., 2010). The relevant studies have primarily focused on how to prevent erosion through land-use changes or vegetation coverage (Hoffmann et al., 2008; Yan et al., 2013). However, few field datasets are available on the effect of rain-formed crust on wind erosion. The objectives of this study were the following: 1) to investigate the effects of rainfall quantity on soil crust thickness, 2) to examine the effects of soil crust on wind erosion, 3) to determine the levels of crust formation (crust formed by various rainfall quantities) that can most effectively resist wind erosion, and 4) to evaluate the differences between the responses of soils with

different treatment histories to crust formation and subsequent wind erosion at given rainfall quantities. 2. Materials and methods 2.1. Site description The study was conducted in the Baiyinxile pasture, a component of the Xilingele grassland in Inner Mongolia, China (43°26′N, 116°04′E) (Fig. 1). This semi-arid region comprises basalt plateaus that are primarily covered with fine-sand loess with the typical chestnut and calcic chernozem soil types. The area features a semi-arid steppe climate (cold and dry in winter but mild and humid in summer), with an annual average rainfall of 264 mm (1981–2010). Precipitation is highly variable, with 75% of the total occurring between June and September. The average daily temperature is − 22.3 °C in the coldest month (January) and 18.8 °C in the hottest month (July). Strong winds associated with dust storms occur from March to May, with an average monthly speed of up to 4.9 m s−1. Wind erosion and dust storms are common phenomena and contribute considerably to the dynamics of soil carbon and nutrients through dust emission and deposition in this area (Hoffmann et al., 2008). 2.2. Experimental design The soils were sampled in May 2013 from four fields with various land-use types within an area of 2400 ha (Fig. 1). The sites include: 1) a field that has been ungrazed since 1979 (Y79), featuring greater than 70% vegetation coverage and representing the original climax community in the typical Inner Mongolian steppe, where vegetation is dominated by Stipa grandis, Leymus chinensis community; 2) a field that has been ungrazed since 1983 (T83), featuring well-recovered vegetation and soils (the dominant community type is similar to Y79); 3) a continuously grazed field (TW), with a grazing intensity of more than two sheep ha−1 (overgrazing has degenerated the community to the type dominated by Artemisia frigid), less than 30% vegetation coverage,

Fig. 1. Geographic location of the study area and the distribution of the fields where the soil samples were collected.

Y. Yan et al. / Geoderma 249–250 (2015) 79–86

and less than 5 cm community height; and 4) a field of arable land that has been in use for over 30 years (CL), where main crops like wheat and buckwheat are planted, with a small quantity of residue after harvest season and serious wind erosion. To investigate the effects of rainformed soil crust on wind erosion, soils were collected from the surface 5 cm of each field, air dried, and sieved in the laboratory to remove roots and debris N 2 mm. For each soil, the pre-treated samples were divided into fifteen sub-samples (3 replicates × 5 treatments) of 600 g each; these samples were used to create soil crusts under various simulated rainfall quantities and consecutive wind erosion treatments. The 600 g samples of treated soils were placed in trays (20 × 20 × 3 cm), and the tray and soil were weighed. A similar method was used in previous studies by Yan et al. (2011, 2013). Light rainfall was simulated by hand using pure water, which was gently sprayed from a 50 ml spray bottle onto the surface of the treated soils. The electrical conductivity of the pure water was 3 μs cm−1, and the simulated rainfall rate was 3 mm h−1, which was determined through a calculation of the simulated rainfall amount divided by the spray time, i.e., 1 mm simulated rainfall amount was applied over the course of 20 min. Due to the small amount of water added and the relatively small spray bottle (50 ml), the spraying process was easily controlled. The spray bottle outlet was positioned just below the tray side when the water was sprayed in order to avoid the loss of rainfall by spraying beyond the sides. Because the water was sprayed onto the soil surface in the form of mist, the drop size was difficult to determine. Although there will inevitably be differences between simulated rainfall and natural rainfall conditions, this method ensured that the simulated rainfall quantity could be accurately controlled. The water quantity and its equivalent rainfall quantity in millimeters were calculated according to the tray size and quantity of rainfall required, e.g., 4 ml of pure water is equivalent to 0.1 mm of rainfall. Each soil was treated with five rainfall levels: 0.0 mm, 0.2 mm, 0.5 mm, 0.8 mm and 1.2 mm. These levels of rainfall produced dramatic changes in soil erodibility in a study by Yan et al. (2013). Each treatment included three replicates. After the simulated rainfall treatments, all trays were air dried to a constant weight, and the water content of the air-dried soil was approximately 1.5%. Soil crusts were observed to form in the trays. Then, a total of 60 trays with four soils were exposed to natural wind to study the wind erosion effects of various levels of physical crust. We selected a 24 ha long-ungrazed grassland area with flat terrain and relatively even vegetation coverage (community coverage N 80% and community height = 20 cm). This location provides relatively uniform surface conditions and ensures that there is almost no local dust emission except under the influence of strong dust weather. The trays were mounted in reinforced brackets at a height of 1 m above the ground on this site to minimize the influence of wind-blown sediments from the ground. Thus, soil removal by wind was the only process present, and the soil loss of different treatments could be accurately evaluated. Of course, the wind speed at 1 m above soil surface is certainly higher than that of the soil surface, which results in a more rapid wind erosion process than that of soil surface due to the higher wind speed. The wind speed can be calculated for different heights according to the wind speed profile function. In this study, we provide the actual wind speed at 1 m height above the soil surface. Therefore, the actual wind speed data can be used as reference conditions for relative comparisons with other studies. The 60 trays were divided into three replicate groups of 20 trays (4 soils × 5 simulated rainfall treatments). The trays in each group were arranged randomly in a linear pattern with a spacing of 2 m and were perpendicular to the main wind direction. To assess the potential influence of the deposition of wind-blown sediments from surface erosion on the treated soils, three marble dust collectors (Goossens and Offer, 2000), one for each group, were installed approximately 30 m downwind of the trays to collect wind-blown sediment. In addition, three trays with 600 g soil were placed in a well-ventilated room to

81

examine the possible influence of water vapor on soil weight due to soil moisture levels in the air-dried soils used in our study. These control trays were weighed, and the change in soil weight during the experiment period (142 min) was within 0.3 g, i.e., b0.05% of the total mass. Therefore, there was no significant difference between the weights at different weighing times (p b 0.05). Therefore, the effect of water vapor on soil weight was not considered to be significant in this study. The synchronous wind speed at 1 m above the ground was measured at 1-minute intervals using a wind measurement instrument that consisted of a data logger (FC-2, China), a three-cup anemometer and a wind vane.

2.3. Field experiments and laboratory analyses The wind-blown field experiment was conducted on June 5th, 2013. We anticipated that the observed result might be influenced by the limited soil supply in the trays (due to tray size) if the soils were exposed to wind for an extended duration, especially under low rainfall conditions. Therefore, the duration of the wind erosion treatment was 142 min, i.e., when the soils in the non-rainfall treatment presented their first visual exposure of the bottom of the tray. Before this point, we consider that there was no limitation on the soil supply for any treatments. After the wind erosion treatment, each tray was weighed to determine the corresponding soil loss ratio for the various treatments. A soil loss ratio (SLR, %) was calculated using formula (1). Crust thickness was measured when the first complete crack in the crust occurred, thus allowing the measurement of a complete cross-section of the crust. The crust was gently tilted, and the thickness at three random locations was measured with a ruler (B.C. Feng et al., 2013; G. Feng et al., 2013).

SLRð%Þ ¼

MOS−MRS  100 MOS

ð1Þ

where MOS is the mass of the sub-sample (600 g) and MRS is the mass of the soil left in the tray after the erosion. Sample splits for particle size analysis were dispersed with sodium hexa-metaphosphate (Wang et al., 2006). After sample dispersion, the particle size of the original soils was measured with a Mastersizer 2000 laser particle size analyzer, and the measured size range was 0–2000 μm. The SOC was measured by digestion with potassium dichromate and back-titration with 0.025 mol L−1 ferrous ammonium sulfate (Kalembasa and Jenkinson, 1973). The measurement of total nitrogen (TN) followed the Kjeldahl digestion method (Moore and Chapman, 1986). Total phosphorus (TP) was determined by the ascorbic acid and ammonium molybdate blue method (Bao, 2000). Available N (AN) was determined by the alkaline hydrolysis diffusion method (Bao, 2000).

2.4. Statistical analyses A statistical analysis was carried out using the SPSS package (v10.0, SPSS, Inc.). The significance level of the data was set at 0.05. The standard error of the treatment means was calculated by oneway analysis of variance (ANOVA). To evaluate the effect of crusted levels (formed by different simulated rainfall quantities) on the soil loss of each soil, multiple comparisons among the various rainfall treatments with respect to the soil loss ratio were performed with the Tukey's honest significant difference (HSD) test. The same method was used to evaluate the differences between the responses of different soil to crust formation and subsequent wind erosion for given rainfall amounts. The homogeneity of variance was tested before performing ANOVA.

82

Y. Yan et al. / Geoderma 249–250 (2015) 79–86

3. Results 3.1. Original soil properties and soil crust thickness The main particle size fractions and nutrient content of the four original sampled soils are shown in Table 1. Although these soils contain a range of clay, silt and sand contents, they are all classified as sandy loamy soil under the USDA system. There were no significant differences in particle size distribution between the soils Y79 and T83 (p N 0.05). The CL soil had the highest clay and sand contents and the lowest silt content among the four soils (p b 0.05). The Y79 soil had the highest SOC and nutrient contents among the four soils, followed successively by T83, TW and CL (p b 0.05). The soil crust thickness increased linearly with increasing amounts of simulated rainfall in all of the soils (Table 2). The CL soil had the highest crust thickness (p b 0.05), followed successively by TW, T83 and Y79. The soil crust thickness (Yct, mm) of each soil increased with increasing simulated rainfall quantities, following a positive linear function (Fig. 2): Yct(Y79) = 5.48Xr, (R2 = 0.993, p b 0.0001, n = 5), Yct(T83) = 5.63Xr, (R2 = 0.976, p b 0.0001, n = 5), Yct(TW) = 5.92Xr, (R2 = 0.999, p b 0.0001, n = 5), Yct(CL) = 6.45Xr, (R2 = 0.996, p b 0.0001, n = 5), where Xr is the simulated rainfall (mm). These equations represent the results of the simulated rainfall ranging from 0 mm to 1.2 mm. 3.2. Wind speed conditions during the experimental period The dynamics of wind speed during the experimental period are shown in Fig. 3. The main wind direction was northwest (260–291°) during the experiment (142 min). When the wind speed reached 5 m s−1, soil loss was visually observed. Therefore, we set 5 m s−1 as the threshold of wind speed needed to initiate erosion (Gillette et al., 1980). Throughout the experiment, the total time above the threshold was 42 min, and the highest wind speed was 8.2 m s−1. 3.3. Soil loss ratios after wind erosion under various crust thicknesses The deposition collected at the end of the experimental period was less than 0.03 g, which only accounted for b0.005% of the weight of the treated soil in a tray. Therefore, the influence of deposition on the treated soils was not considered to be significant in this study. The proportion of soil loss for each soil varied with the level of rainfall. Obvious wind erosion occurred for the non-rainfall treatment and 0.2 mm rainfall treatment for all four soils, whereas only minor soil loss was observed for the high rainfall treatments (above 0.5 mm) (Fig. 4 A). Taking Y79 as an example, after the wind erosion treatment, the soil loss ratios were 33% and 28% in the non-rainfall and 0.2 mm rainfall treatments, respectively. These percentages correspond to actual soil loss amounts of 4.95 and 4.2 kg m−2 for the non-rainfall and 0.2 mm rainfall treatments, respectively. There was no significant difference between these values (p N 0.5), but both were significantly greater

Table 2 Thickness (mm) of the soil crust in the four soil types with various levels of simulated rainfall. No crust formed in the non-rainfall treatment. The different letters in a column indicate significant differences between soils at the 5% level (Tukey's test). There are significant differences in crust thickness between the simulated rainfall levels in each soil at the 5% level (Tukey's test), which are not labeled. Field

Rainfall (mm)

Y79 T83 TW CL Mean

0.2

0.5

0.8

1.2

1.1b 1.1b 1.1ab 1.3a 1.1

2.5b 2.8ab 2.9ab 3.4a 2.9

3.9c 4.0c 4.8b 5.4a 4.5

6.8b 7.1ab 7.3a 7.5a 7.2

than those of the treatments with more than 0.5 mm of rainfall (p b 0.5). Soil losses of only 2.5% occurred in treatments with more than 0.5 mm of rainfall. There were significant differences in the soil loss ratios among the four soils in the non-rainfall and 0.2 mm rainfall treatments (p b 0.5) (Fig. 4 B). The soil loss ratios of Y79 and CL were the highest and the lowest, respectively, among the four soils in the non-rainfall and 0.2 mm rainfall treatments (p b 0.5), and the soil loss ratios of the other two soils fell between those of Y79 and CL. In contrast, no significant differences were present among the four soils in the treatments with more than 0.5 mm of rainfall (p N 0.5) (Fig. 4 B). We analyzed the relationships between the soil loss ratio and the original soil properties, including soil texture and the SOC. The results showed that the soil loss ratio increased with increasing silt plus clay contents for the non-rainfall and 0.2 mm rainfall treatments (Fig. 5 A). Similarly, the ratio also increased with increasing SOC content (Fig. 5 B). 4. Discussion 4.1. Effects of simulated rainfall and soil properties on the soil crust thickness Previous studies have suggested that cumulative rainfall is the dominant factor that affects crust formation (G. Feng et al., 2013). We selected four soils with various contents of SOC and nutrients, chosen to represent the local land-use types in the typical Inner Mongolian steppe, to investigate the impact of light rainfall on soil crust formation and soil loss. Our experiment showed that the soil crust thickness linearly increased with increasing amounts of simulated rainfall in all soil types for amounts of rainfall up to 1.2 mm. However, a previous study reported a logarithmic relationship based on a large number of samples with rainfall ranging from zero to N10 mm (G. Feng et al., 2013). The logarithmic relationship implies that crust development in response to rainfall is initially large and subsequently decreases with additional rainfall toward a given equilibrium (G. Feng et al., 2013). Given that light rainfall events up to 1.2 mm represent the actual conditions in the studied region, our simple linear relationship might be better suited within this range of rainfall. In fact, the experimental data of G. Feng et al. (2013) also fit the linear relationship for amounts of rainfall up to 1 mm (Fig. 2 Warden), although they used an alternative logarithmic relationship with a slightly better fit relationship.

Table 1 Soil particle size and chemical properties of the soils collected from the top 5 cm in four fields: Ungrazed since 1979 (Y79), ungrazed since 1983 (T83), continuous grazing (TW) and arable land in use for more than 30 years (CL). Different letters in a column indicate significant differences between soils at the 5% level (Tukey's test). Field

Y79 T83 TW CL

SOC (g/kg)

Total (g/kg) N

P

21.96a 16.31b 11.63c 6.01d

2.12a 1.61b 1.21c 0.61d

0.39a 0.34b 0.28c 0.22d

Available N (mg/kg)

Soil particle size distribution (%) Clay (b0.002 mm)

Silt (0.002–0.05 mm)

Sand (N0.05 mm)

187.64a 143.23b 118.94c 59.95d

2.30b 2.31b 2.44b 3.36a

40.60a 39.09ab 38.14b 28.00c

57.10c 58.59bc 59.42b 68.64a

Y. Yan et al. / Geoderma 249–250 (2015) 79–86

Soil cr us t t hicknes s (mm)

7

TW

T83

Y79

83

CL

Warden

6 5 4

3 2 1

0 0

0.3

0.6

0.9

1.20

Rainfall (mm)

0.3

0.6

0.9

Rainfall (mm)

1.20

0.3

0.6

0.9

Rainfall (mm)

1.20

0.3

0.6

0.9

Rainfall (mm)

1.20

0.3

0.6

0.9

1.2

Rainfall (mm)

Fig. 2. Relationship between rainfall quantity and soil crust thickness. The terms Y79, T83, TW and CL identify our experimental soils, and the panel labeled Warden presents data from G. Feng et al. (2013).

The formation of physical soil crust was also influenced by the soil properties. Soils that are especially susceptible to physical crusting have low structural and aggregate stability, low organic matter, high silt, and/or high salt contents (Belnap, 2001). Under these conditions, the soil thickness of the CL soil was slightly greater than those of the other three soils, i.e., the soil more easily formed a crust. However, the silt plus clay content of this soil is the lowest among the sampled soils (p b 0.05). Considering that the silt plus clay content is still 31%, we believe that this level might be sufficiently high to form a crust. In addition, a previous study reported that the soil crust thickness changed from 0.8 mm at 0.15 mm rainfall to 5.0 mm at 1.0 mm rainfall for a silt loam soil with a silt plus clay content of 87.3% (G. Feng et al., 2013); the variations in crust thickness with rainfall were small compared with our results. The silt plus clay content of the soils that Feng et al. used were much higher than those of our soils. Therefore, the silt plus clay content of the soil was not a primary factor affecting soil crust thickness in our study. Decomposed organic matter or the products resulting from decomposition have been reported to decrease the soil cloddiness and increase the erodibility by wind (Chepil, 1954). Therefore, the reduced SOC content of the CL soil among the soil types should be the main cause underlying the development of a thicker crust. 4.2. Effect of soil crust on the soil mass loss There are few field datasets that provide information regarding the effect on wind erosion of a soil crust formed by rainfall, especially the effect of light rainfall. Previous work has primarily focused on crusts formed by heavy rainfall and the soil crust effect on abrasion flux (Chepil, 1953, 1958; Zobeck, 1991). The stability of crusted soils against abrasion by wind-blown soil material has been measured, and crusted soils have been shown to erode at a rate of approximately 0.04 to

0.4 that of freshly cultivated fields (Chepil, 1953, 1954, 1955, 1958). The resistance of soil crust formed by a total of 32 mm of simulated rainfall to abrasion for 13 mineral soils was examined by Zobeck (1991), who found that the soil loss ratio of crusted to loose soils ranged from 0.0002 to 0.098. These results demonstrate the significant effect of soil crust on abrasion by wind-blown soil material. However, these studies did not assess the effect of weak crusts formed by light rainfall on wind erosion. We anticipated that the effects of crusts on soil loss would be closely related to the penetration resistance of the crust. A previous study found a positive relationship between crust thickness and penetration resistance (G. Feng et al., 2013). Our experiment confirmed that the crust level formed by 0.2 mm of rainfall has only a small effect on decreasing soil loss, whereas the crust levels created in the high rainfall treatments (greater than 0.5 mm) significantly increased the resistance of soil to erosion compared with that of the original unconsolidated soils (p b 0.05). We analyzed the resistance to erosion of various crust levels by comparing the soil loss of crusted soils (following various rainfall treatments) with that of the original unconsolidated soil (non-rainfall treatment). The soils subjected to various rainfall treatments showed various levels of resistance to erosion. For example, in soil Y79, the soil loss was reduced by 14.8%, 92.6%, 96.6% and 99.0% in treatments with 0.2 mm, 0.5 mm, 0.8 mm and 1.2 mm rainfall, respectively, compared with that for the non-rainfall treatment. According to our study, the soil crust formed by a rainfall quantity greater than 0.5 mm has a significant effect on the resistance to wind erosion. It should be noted that our results regarding the resistance of the rain-formed crust vs. wind erosion of the rain-formed crust were produced over a limited duration of wind erosion. We believe that given sufficient exposure time to wind, all crusts would eventually be broken due to the influence of

Fig. 3. Wind speed measured at the end of each minute at a height of 1 m above the ground at the field site during the experimental period (DT: minutes during which the wind speed exceeded the observed threshold).

84

Y. Yan et al. / Geoderma 249–250 (2015) 79–86

Fig. 4. Soil loss ratios of the four soils at various rainfall levels. The original soil weight was 600 g. The bars labeled with different letters differ significantly (p b 0.05) for each soil (A) and for each rainfall level (B).

the abrasion process or other disturbances. Thus, further study is needed to determine how long a rain-formed soil crust remains intact. In this study, we used a unique approach to study the effects of light rainfall on soil crusting and subsequent wind erosion for soils in Inner Mongolia. These methods and experiment design let us observe only the soil removal process by wind, and the soil loss can therefore be calculated quantitatively. This study mainly focused on comparisons between the resistances of different crust thicknesses (formed by different simulated rainfall quantities) to the shear force of wind because the small tray size is not long enough for saltation and abrasion to develop. The resistance to abrasion will be considered in future research work. In addition, the 3 cm-deep tray inevitably results in some wind deflection and turbulence. Therefore, the saltation of particles below a height of 3 cm may have been prevented, resulting in these particles remaining in the tray and not being included in the soil loss mass. We anticipated that all these influences are the same for all treatments (Yan et al., 2013). We therefore believe that our experimental method provides ground for direct comparison among the studied soils. Further research is required to assess how these soils respond to rainfall and wind erosion in the field condition.

4.3. Responses of soils with different land-use histories to wind erosion under various crust thicknesses Unreasonable land use, such as cultivation and overgrazing, has a significant influence on wind erosion via the destruction of natural vegetation and soil conditions, especially in dryland environments. Additionally, wind erosion influences soil quality, such as SOC and nutrient condition (Yan et al., 2010). A number of studies have examined wind erosion effects on soil fine particles and SOC. There is a general consensus that fine soil particles are preferentially depleted by wind erosion, which is significant because disproportionately greater amounts of organic carbon and nutrients are stored in the fine soil particle fraction (Li et al., 2007, 2009; Yan et al., 2013). In addition, a series of results on SOC enrichment in dust emissions have been reported (Goossens, 2004; Webb et al., 2012, 2013; Chappell et al., 2013). These studies showed that the mean SOC enrichment ratio for dust (2 m) for various soil types was 2.1–41.9 for sites with sand-rich soil and 2.1 from areas with clay soil (Webb et al., 2012, 2013; Chappell et al., 2013). These results imply that soils with high quantities of fine particles and SOC may be more erodible.

Fig. 5. Relationships between the soil loss ratio and the original clay plus silt content (A, particle size b0.05 mm) and SOC concentration (B) after the soils were exposed to natural wind for 142 min.

Y. Yan et al. / Geoderma 249–250 (2015) 79–86

Although many studies on the effect of wind erosion on soil properties (the content of fine particles and SOC) have been reported, relatively few studies have described the effects of the proportions of fine particles and SOC on wind erosion, particularly for both natural and unconsolidated (treated) soils with different land use histories. The original soils used in this study were collected from various landuse types. These sites have different vegetation coverage percentages, and the soils have various proportions of fine particles (clay plus silt) and SOC due to the varying land-use histories (Table 1). The site Y79 has greater than 70% vegetation coverage for the whole year due to the exclusion of livestock and has the highest SOC content. In contrast, the site CL has the lowest vegetation coverage due to crop harvesting, resulting in unconsolidated soils that are exposed in the windy season and that feature the lowest SOC content. In natural conditions, the sites Y79 and T83 have lower wind erosion rates than TW and CL due to the higher vegetation coverage and aggregated soils in the ungrazed grasslands (Y79 and T83) (Yan et al., 2010). Similar results have been reported by Xu et al. (2005), e.g., the wind erosion rates of grazing grassland and cropland are 11.5 and 91.8 times higher, respectively, than the erosion rate of long-term ungrazed grassland in the study area. Webb et al. (2013) also showed that sandy soil sites with lower fine particulate and SOC contents can more readily release dust and SOC than sites that are aggregated and have higher silt, clay and SOC contents. Consequently, soils with lower fine particle and SOC contents are typically more erodible in natural conditions because generally high SOC values are associated with high vegetation coverage and aggregated soils, whereas low SOC values are associated with low vegetation coverage and more unconsolidated soils. In addition, this result might also be due to the greater efficiency of the abrasion process for sandy soils with dense saltators (Webb et al., 2013). When soils are treated (air dried and sieved to remove roots and debris N2 mm.), such as in this study, the effects of environment factors, such as vegetation coverage and large soil aggregates, are excluded, and the resistance to wind depends primarily on the soil particles themselves. The results that we found were different from those of earlier studies examining natural conditions: The soil loss ratio increased with increasing proportions of fine particles (silt plus clay) and SOC under non-rainfall and 0.2 mm rainfall treatments, i.e., the Y79 soil is the most erodible and the CL soil is the least erodible in our study (Fig. 4b). These results imply that unconsolidated soils with higher proportions of fine particles and SOC can be more easily eroded. The results also imply that once ungrazed grassland (e.g., Y79) is subjected to a disturbance in the vegetation coverage and soil structure (e.g., vegetation clearing for agriculture or human disturbance), the soil will face a higher wind erosion risk, and the SOC will inevitably be depleted (Webb et al., 2012).

5. Conclusions In this experiment, we simulated five light rainfall levels of 0.0 mm, 0.2 mm, 0.5 mm, 0.8 mm and 1.2 mm to investigate the impact of light rainfall on soil crust formation and its effects on wind erosion via a unique approach. Our results show that soil crust thickness increased linearly (up to 1.2 mm) with increasing rainfall in soils from the semiarid steppe of Inner Mongolia. Future research will include different soil types and higher rainfall amounts. If this similar relationship also exists in other different soils, the simple regression function can easily be used in wind erosion prediction modeling. After a short-term wind treatment (142 min), we found that the weak crust formed by 0.2 mm of rainfall only provided a slight resistance to wind erosion. The thin soil crust thickness (1.1 mm) formed by 0.2 mm of rainfall did not possess enough crust strength to resist the shear force of the wind. When the rainfall quantity reached 0.5 mm, the soil crust thickness was greater than 2.5 mm, which almost completely prevented wind erosion during the experimental period. These findings suggest that small rainfall events (0.5–1.2 mm of precipitation) should also be considered for in

85

wind erosion management and prediction modeling, especially in arid regions. Our results suggest that the soil properties have a significant influence on wind erosion. The influence was shown to differ between natural conditions and unconsolidated experimental soils (air dried and sieved). In our study, the soils Y79 and T83 have lower wind erosion rates than TW and CL, i.e., the soils with lower proportions of fine particles and SOC are more erodible in natural conditions. Our interpretation of the mechanisms is that generally high SOC values are associated with high vegetation coverage and aggregated soils, whereas low SOC values are associated with low vegetation coverage and more unconsolidated soils. However, when the soils were treated to create unconsolidated soils, such as in this study, the results show that the soil loss ratio increased with increasing fine particles and SOC content for non-rainfall and 0.2 mm rainfall treatments. In other words, the soils with high proportions of fine particles and SOC were more erodible. These results imply that once ungrazed grassland (e.g., Y79) is subjected to vegetation clearing for agriculture and human disturbance, these areas will face a higher wind erosion risk, and the SOC would inevitably be depleted (Webb et al., 2012). Acknowledgments This study was funded by the International S & T Cooperation Project of China (2012DFA31290), a National Non-profit Institute Research Grant of CAAS (2014-13), the National Natural Science Foundation of China (40901053), the National Public Benefit (Agricultural) Research Foundation of China (201003061) and the National Science and Technology Support Program (2012BAD13B07). We are grateful to Jason Neff and Joseph Yaw Appiah-Gyapong for their constructive comments and suggestions during revision. References Baddock, M.C., Zobeck, T.M., Scott Van Pelt, R., Fredrickson, E.L., 2011. Dust emissions from undisturbed and disturbed, crusted playa surfaces: cattle trampling effects. Aeolian Res. 3, 31–41. Bao, S.D., 2000. Soil Agrochemical Analysis. China Agriculture Press, Beijing, p. 56 (71–78 (in Chinese)). Belnap, J., 2001. Comparative structure of physical and biological soil crusts. In: Belnap, J., Lange, O.L. (Eds.), Biological Soil Crusts: Structure, Function, and Management, Ecological Studies Series 150. Springer-Verlag, Berlin, pp. 177–191. Belnap, J., Gillette, D.A., 1997. Disturbance of biological soil crusts: impacts on potential wind erodibility of sand desert soils in Southeastern Utah. Land Degrad. Dev. 8, 355–362. Chamizo, S., Cantón, Y., Domingo, F., Belnap, J., 2011. Evaporative losses from soils covered by physical and different types of biological soil crusts. Hydrol. Process. http://dx.doi. org/10.1002/hyp.8421. Chappell, A., Webb, N.P., Butler, H.J., Strong, C.L., Mctainsh, G.H., Leys, J.F., Viscarra Rossel, R.A., 2013. Soil organic carbon dust emission: an omitted global source of atmospheric CO2. Global Change Biol. 19, 3238–3244. Chepil, W.S., 1953. Factors that influence clod structure and erodibility of soil by wind: I. Soil texture. Soil Sci. 75, 473–483. Chepil, W.S., 1954. Factors that influence clod structure and erodibility of soil by wind: III. Calcium carbonate and decomposed organic matter. Soil Sci. 77, 473–480. Chepil, W.S., 1955. Factors that influence clod structure and erodibility of soil by wind: V. Organic matter at various stages of decomposition. Soil Sci. 77, 413–421. Chepil, W.S., 1958. Soil conditions that influence wind erosion. USDA-ARS Tech. Bull. vol. 1185. U.S. Govt. Print. Office, Washington, DC. Farres, R., 1978. The role of time and aggregate size in the crusting processes. Earth Surf. Process. Landf. 3, 243–254. Feng, B.C., Gale, W.J., Guo, C.Q., Fang, W.S., 2013. Process and mechanism for the development of physical crusts in three typical Chinese soils. Pedosphere 23, 321–332. Feng, G., Sharratt, B.S., Vaddella, V., 2013. Windblown soil crust formation under light rainfall in a semiarid region. Soil Tillage Res. 128, 91–96. Fryrear, D.W., Bilbro, J.D., Saleh, A., Schomberg, H., Stout, J.E., Zobeck, T.M., 2000. RWEQ: improved wind erosion technology. J. Soil Water Conserv. 55, 183–189. Gillette, D.A., Adams, J., Endo, A., Smith, D., Kihl, R., 1980. Threshold velocities for input of soil particles into the air by desert soils. J. Geophys. Res. 85, 5621–5630. Goossens, D., 2004. Net loss and transport of organic matter during wind erosion on loamy sandy soil. In: Goossens, D., Riksen, M. (Eds.), Wind Erosion and Dust Dynamics: Observations, Simulations, Modelling. ESW Publications, Wageningen, pp. 81–102. Goossens, D., Offer, Z.Y., 2000. Wind tunnel and field calibration of six aeolian dust samplers. Atmos. Environ. 34, 1043–1057.

86

Y. Yan et al. / Geoderma 249–250 (2015) 79–86

Hagen, L.J., Zobeck, T.M., Skidmore, E.L., Elminyawi, I., 1995. WEPS technical documentation: soil submodel. Wind Erosion Prediction System Technical Description, Proceedings of WEPP/WEPS Symposium, Soil and Water Conservation Society, Des Moines, IA/Ankeny, IA, August 9–11,1995. Hoffmann, C., Funk, R., Li, Y., 2008. Effect of grazing on wind driven carbon and nitrogen ratios in the grasslands of Inner Mongolia. Catena 75, 182–190. Kalembasa, S.J., Jenkinson, D.S., 1973. A comparative study of titrimetric and gravimetric methods for determination of organic carbon in soil. J. Sci. Food Agric. 24, 1085–1090. Li, J.R., Okin, G.S., Alvarez, L., Epstein, H., 2007. Quantitative effects of vegetation cover on wind erosion and soil nutrient loss in a desert grassland of southern New Mexico, USA. Biogeochemistry 85, 317–332. Li, J.R., Okin, G.S., Epstein, H.E., 2009. Effects of enhanced wind erosion on surface soil texture and characteristics of windblown sediments. J. Geophys. Res. http://dx.doi.org/ 10.1029/2008JG000903. Moore, P.D., Chapman, S.B., 1986. Methods in Plant Ecology. Alden Press, Oxford. Morrison, M.W., Prunty, L., Giles, J.F., 1985. Characterizing strength of soil crusts formed by simulated rainfall. Soil Sci. Soc. Am. J. 49, 427–431. Rajot, J.L., Alfaroa, S.C., Gomes, L., Gaudichet, A., 2003. Soil crusting on sandy soils and its influence on wind erosion. Catena 53, 1–16. Wang, H., Mason, J.A., Balsam, W.L., 2006. The importance of both geological and pedological processes in control of grain size and sedimentation rates in Peoria Loess. Geoderma 136, 388–400.

Webb, N.P., Chappell, A., Strong, C., Marx, S.K., Mctainsh, G.H., 2012. The significance of carbon-enriched dust for global carbon accounting. Global Change Biol. 18, 3275–3278. Webb, N.P., Strong, C.L., Chappell, A., Marx, S.K., Mctainsh, G.H., 2013. Soil organic carbon enrichment of dust emissions: magnitude, mechanisms and its implications for the carbon cycle. Earth Surf. Process. Landf. 38, 1662–1671. Xu, Z.Q., Li, W.H., Min, Q.W., 2005. Experimental research on the anti-wind erosion of typical grasslands. Environ. Sci. 26, 164–168 (in Chinese with English summary). Yan, Y.C., Tang, H.P., Zhang, X.S., 2010. Probe on grassland wind erosion based on the analysis of soil particle size. J. Desert Res. 30, 1263–1268 (in Chinese with English summary). Yan, Y.C., Xu, X.L., Xin, X.P., Yang, G.X., Wang, X., Yan, R.R., Chen, B.R., 2011. Effect of vegetation coverage on aeolian dust accumulation in a semiarid steppe of Northern China. Catena 87, 351–356. Yan, Y.C., Xin, X.L., Xu, X.P., Wang, X., Yang, G.X., Yan, R.R., Chen, B.R., 2013. Quantitative effects of wind erosion on the soil texture and soil nutrients under different vegetation coverage in a semiarid steppe of northern China. Plant Soil 369, 585–598. Zobeck, T.M., 1991. Abrasion of crusted soils: influence of abrader flux and soil properties. Soil Sci. Soc. Am. J. 55, 1091–1097.