The sandy soil used in the experiment was obtained from the dunes at the Gurbantunggut Desert in Xinjiang, China. The distribution of the particle size of the natural sandy soils is shown in Table 1. To ensure the uniformity and representativeness of the sandy soil, we selected soil with a particle size range of 100–200 μm for the experiment. To remove the interference of soluble salts in the soil, the soil was cleaned with deionized water, and the cleaned sand was placed in an oven at 105°C for drying. The salts used for the experiment were CaCl₂, KCl, Na₂SO₄, NaCl, Na₂CO₃, and NaHCO₃ (purity ≥99.99%). Six gradients of salt content were set for each salt, 0.2, 0.5, 1, 2, 3, and 4% (the value of soluble salt mass as a percentage of the soil mass). The solvent was deionized water.

For the experiment, the mass of the soil was fixed at 750 g, and the corresponding mass of salt at all salt content is shown in Table 2. 100 ml of deionized water was added, stirred well, and then poured into a mold (a cylindrical plastic cup of 3.5 cm height and 4 cm radius). The surface of the mold was then smoothed, and the bulk density is 1.43 g cm⁻³, after which the samples were placed in an oven at 75°C until all the water in the samples evaporated and the sample mass reached constant weight. Ten replicates were carried out for all salt content.

The dried samples were cooled to room temperature (the average temperature is 24.41°C and the average humidity is 20.85% indoor). Photographs were taken from above and on the side of the samples to observe the apparent morphological characteristics of the crusts. The process of taking pictures lasts about 2–3 min, and then the compressive strength was immediately measured. The crust compressive strength was tested with a digital electronic manometer (HANDPI, HP-1000N, accuracy: 0.1 N, range: 0–1000 N). The manometer was fixed on a holder, and the speed of the manometer probe was set to 1 mm s⁻¹ and the diameter of the manometer probe is 10 mm. The stress-time curves were automatically recorded by the manometer connected to a computer. The compressive strength of a sample was the value of the maximum compressive strength on the stress-time curve of the sample. The average and standard deviation of the compressive strength of each salt content were calculated, and the average of the compressive strength was represented as the compressive strength of the samples at the same salt content. The energy absorbed by the crust per unit area is the toughness of the crust, that is, the ratio between the work done by the manometer probe and the area of the manometer probe when the compressive strength of the crust reaches its maximum.

Figure 1 shows the appearance of the crusts of the six salts at different salt contents. As can be seen from Figure 1, there were differences in the appearance of crusts formed by the different salts. When the salt content was 0.2%, there were no macroscopic salt crystals on the surface of all six salt crusts, and the color of the crusts was brownish-red. There were a few cracks at the surface of Na₂CO₃ and NaHCO₃ crusts, while cracks were not seen on other salt crusts. When the salt content was 1%, macroscopic salt crystals appeared on the surface of the salt crusts, except for Na₂SO₄. The salt crusts formed by Na₂CO₃ had the largest area of salt crystals on the surface and had some fine polygonal fractures. When the salt content was 4%, the salt crystals did not appear on the surface of Na₂SO₄ salt crusts still, and the salt crystals on the surface of the other five salt crusts increased noticeably. Polygonal fractures existed on the surface of both salt crusts of Na₂CO₃ and NaHCO₃, and an obvious expansion could be seen in the crust formed by Na₂CO₃.

The side view photographs of Na₂CO₃ and NaHCO₃ crusts are shown in Figure 2. When the salt content was 0.2%, the surface of the crusts formed by Na₂CO₃ and NaHCO₃ had cracked but did not exhibit swelling. When the salt content was 1%, obvious swelling could be seen in the Na₂CO₃ samples, and swelling was seen only in the surface edge area in the NaHCO₃ samples. When the salt content was 4%, the swelling phenomenon inside both samples was further intensified, the crusts formed by Na₂CO₃ had separated from the soil thoroughly, and the internal loose sand under the crust could be seen. The crusts formed by NaHCO₃ were lifted noticeably from the sample.

Figure 3 showed the variation of compressive strength with salt content for the six salt crusts. The compressive strength of different salt crusts varied with the salt content variation. With the increase in the salt content, the compressive strength of the Na₂SO₄ crust increased first and then decreased, reaching a maximum of 5.12 MPa at a salt content of 3%. The compressive strength of the NaCl crust increased first and then remained stable, and the compressive strength of the crust of NaCl reached 3.59 MPa when the salt content was 3% and 3.76 MPa when the salt content was 4%, which was only an increase of 4.7%. The compressive strength of KCl and CaCl₂ crusts increased almost exponentially with the increase of the salt content. The compressive strength of NaHCO₃ crusts increased slightly but then decreases, reaching a maximum at 0.5% salt content, but its compressive strength was only 0.40 MPa. The compressive strength of Na₂CO₃ crusts was almost stable, and the maximum was only 0.03 MPa.

When the salt content was ≤3%, the compressive strength of the crust formed by Na₂SO₄ was significantly higher than that of the other five salt crusts (p < 0.05), and the compressive strength reached 1.15 MPa when the salt content was 0.2%. When the salt content was >3%, the compressive strength of the crusts formed by CaCl₂ and KCl salts was higher than that of the crusts formed by Na₂SO₄. The compressive strength of CaCl₂ and KCl salt crusts reached 5.36 and 4.98 MPa at 4% salt content, respectively, while the compressive strength of Na₂SO₄ and NaCl was slightly less, 4.53 and 3.76 MPa, respectively.

Figure 4 showed the stress-time curves of four salts, Na₂SO₄, NaCl, CaCl₂, and KCl at different salt contents (the stress-time curve of Na₂CO₃ and NaHCO₃ crusts was not considered because the compressive strength of their crusts was very small). The energy absorbed by the crust per unit area is the toughness of the crust, and Figure 5 is the graph of the crust toughness of four salts. From Figure 4 and Figure 5, it could be seen that the crust compressive strength and crust toughness did not always have a positive correlation, and there were differences in the toughness of different salt crusts. The toughness of the Na₂SO₄ crusts and the NaCl crusts all increased and then decreased. The crust toughness of KCl and CaCl₂ increased gradually as the salt content increased, with the crust toughness of KCl being slightly higher than that of CaCl₂. When the salt content is ≤3%, the crust toughness of Na₂SO₄ was higher than that of CaCl₂, KCl, and NaCl, while in the range of the salt content between 2 and 3%, the difference between the crust toughness of Na₂SO₄ and NaCl was not obvious. When the salt content is 4%, the crust toughness of CaCl₂ and KCl was greater than that of Na₂SO₄ and NaCl.

Under these experimental conditions, Na₂SO₄ formed the subflorescence, with most of the salt crystals distributed inside the soil and fewer salt crystals distributed on the surface of the soil, which was the same as the results of the previous studies (Espinosa et al., 2007; Shahidzdeh-bonn et al., 2010; Gentilini et al., 2012; Li et al., 2021). NaCl, CaCl₂, KCl, Na₂CO₃, and NaHCO₃ formed the efflorescences. As the salt content increased, the crystals of salts gradually increased on the surface of a sample. The crusts formed by Na₂CO₃ and NaHCO₃ produced the phenomena of cracks and swelling on the samples, which was related to the crystallization process of Na₂CO₃ and NaHCO₃. Both solutions could produce hydrate crystals. The hydrate crystals were all bigger than the crystals of Na₂CO₃ and NaHCO₃. The saturated Na₂CO₃ solution produced Na₂CO₃-10H₂O crystals, and the Na₂CO₃-10H₂O crystals would increase in volume, and the formation of hydrate crystals in the soil would produce stress to the surrounding sand grains, causing the soil to swell. The crystals of Na₂CO₃-10H₂O would lose water at temperatures higher than 33°C to form anhydrous Na₂CO₃ crystals. The crystals of anhydrous Na₂CO₃ decreased in volume. Crystal expansion and shrinkage resulted in soil expansion and cracking. The Na₂CO₃-10H₂O crystal layer crystallized on the surface of the soil lost water and then decreased in volume, resulting in the surface crust being cracked (Rijniers et al., 2005; Liu, 2016). The NaHCO₃ solution had a similar phenomenon to the Na₂CO₃ solution, but the volume expansion of NaHCO₃-10H₂O was smaller than that of the Na₂CO₃-10H₂O, so its cracks and expansion were smaller than those of Na₂CO₃. The swelling phenomenon of Na₂CO₃ and NaHCO₃ salt crusts was related to salt content. The greater the salt content, the more obvious the salt swelling phenomenon, which could be verified from Figure 2.

During the evaporation of soil water, the brine in the fine pores evaporates and the salt crystallizes as the salt solution reaches saturation solubility (Rad et al., 2013; Li et al., 2021). In this experiment, the crystals of Na₂SO₄ formed the subflorescence, and most of the salt crystals crystallized in the internal pores of the soil. When the Na₂SO₄ content was ≤3%, with the increase in salt content, the number of salt crystals crystallized in the pores gradually increased, the cementation between salt crystals and soil particles became strong, and the compressive strength of the salt crust increased. With a further increase in salt content (>3%), the crystals gradually filled the soil pores and generated salt swelling pressure, which damaged the soil structure, and the compressive strength of the salt crust formed by Na₂SO₄ decreased obviously. In these experiments, NaCl, KCl, and CaCl₂ formed the efflorescences, with salt crystallization occurring mainly on the soil surface. As evaporation proceeded, the salts gradually accumulated on the soil surface, and the thickness of the salt crust on the surface increased; thus, the increase in salt content only increased the number of the crystals on the surface of the soil and did not affect the internal structure of the soil. Therefore, the compressive strength of these three salts gradually increased with the increase of salt content. The compressive strength of NaCl crusts remained stable at salt content >3%, which might be due to the detachment of the salt crust of NaCl from the soil surface, which hindered water evaporation and inhibited the movement of salt ions to the soil surface (Nachshon et al., 2018). The crystallization process of Na₂CO₃ and NaHCO₃ salts resulted in severe salt swelling, which led to salt crust bulging. The compressive strength of the two salt crusts was extremely low, with the maximum compressive strength of the crusts of Na₂CO₃ being only 0.03 MPa and that of NaHCO₃ being 0.40 MPa.

The crust strength and crust toughness of the crusts affect the wind erosion resistance of the crusts. The greater the crust strength and crust toughness, the greater the potential resistance of the crust to wind erosion, and the more potentially erosive events the soil crust can withstand (Nickling and Ecclesstone, 1981; Nield et al., 2016b). This study compared the wind erosion resistance of six salt crusts by three indicators: crust appearance, crust compressive strength, and crust toughness. The surface of the crust formed by Na₂CO₃ and NaHCO₃ salts had obvious swelling, exposing the loose soil beneath the crust (as shown in Figure 2), and the compressive strength of the crust was extremely low. The compressive strength of the crust formed by the two salts was far from the sand fixation strength of 1 MPa (Han et al., 2000), and the crusts formed by the two salts would be susceptible to wind erosion. The crusts formed by Na₂SO₄, NaCl, KCl, and CaCl₂ were flat relatively, the salt expansion was not observed, and the compressive strength of the salt crust was stronger than that of other salt crusts. These salt crusts could be used for increasing soil resistance. The compressive strength and crust toughness of the Na₂SO₄ crust both increased and then decreased with the increase of the salt content. The compressive strength could reach 1.15 MPa at the salt content of 0.2%, and the compressive strength and crust toughness reached the maximum at the salt content of 3%, with the best resistance to wind erosion. There are two stable crystalline phases during Na₂SO₄ solution crystallization, decahydrate (Na₂SO₄-10H₂O) and anhydrous state (Na₂SO₄); when the temperature is high, anhydrous Na₂SO₄ is generated, and as the humidity in the air increases, anhydrous Na₂SO₄ is transformed into Na₂SO₄-10H₂O. The solubility of Na₂SO₄-10H₂O is less than that of anhydrous Na₂SO₄, resulting in super-saturation of Na₂SO₄-10H₂O, generating a crystallization pressure and causing damage to the medium material (Steiger and Asmussen, 2008; Saidov et al., 2015). Thus, the optimal salt content of the Na₂SO₄ crust for wind erosion resistance under field conditions still needs to be further investigated. The compressive strength of the NaCl crust increased first and then stayed stable with the increase of salt content, reaching the maximum of 3.76 MPa at the salt content of 4%. The toughness of NaCl crust increased first too and then decreased with the increase in salt content. The toughness reached its maximum at the salt content of 3%, and its compressive strength also reached 3.59 MPa, which was slightly lesser than that of 4% (about 0.16 MPa). Thus, for NaCl, the best potential resistance to wind erosion of the salt crust was the salt content of 3%. The compressive strength and crust toughness of KCl and CaCl₂ crusts increased gradually with the increase in salt content, and the compressive strength already reached 1.94 and 1.60 MPa, respectively, at 2% salt content, both having wind erosion resistance. For the crusts formed by Na₂SO₄, NaCl, KCl, and CaCl₂, the compressive strength and toughness of the crust formed by the Na₂SO₄ are higher than those of the remaining salts, and the wind erosion resistance is the best when the salt content is ≤3%; the crust formed by CaCl₂ has the best wind erosion resistance when the salt content is 4%, its compressive strength is the greatest, and the crust toughness is also better.