Irrigation water salinity, as an essential factor influencing salt transport in soil, can significantly affect the distributions of soluble salt and ions in soil. The accumulation of soluble salt not only can modify the physical and chemical properties of soil but also indirectly impact crop growth and development. In the present study, c_salt, c_Na, and c_Cl exhibited an initial rise followed by a subsequent decline among all treatments in the 0–40 cm soil profile (Figure 2). As the irrigation water salinity increased, the solute concentration at the wetting front and the depth at the peak solute concentration also increased, indicating that higher irrigation water salinity could promote the solute front downward movement in unsaturated red soil. This can possibly be attributed to the solutes present in irrigation water, accelerating soil colloidal flocculation and aggregate structure formation to increase the proportion of soil macropores, resulting in a downward movement of the solute front [31]. This finding is consistent with the results of Weissman et al. [32]. They also found that irrigation with higher water salinity produced a faster water infiltration rate and higher NO₃⁻ concentration in the deeper layer of unsaturated loess soil.

BTC is a fundamental tool for describing solute transport behavior in soil. The BTCs of soluble salt, Na⁺, and Cl⁻ among all treatments displayed a smooth S-shaped distribution under steady-state flow (Figure 4), with PV_i < 1 and PV_c > 1, demonstrating earlier breakthrough and shorter trailing during the transport of the three solutes in the red soil. This finding aligns with the results reported by Senthil Kumar et al. [33] that the earlier initial and complete breakthrough of BTCs occurred in the influent with higher Cu(II) ion concentrations. In our study, convection played a dominant role in solute transport (Table 4), probably causing preferential flow. Furthermore, the surfaces of soil particles had an adsorption capacity for solutes, and more micropores in disturbed soil intensified the convection-dispersion of solutes, contributing to the tailing of BTCs [34]. These two common phenomena were observed in diverse solute transport processes in various soils. Moreover, our study found that higher water salinity could diminish the physical non-equilibrium process. As the salinity of the replacement solution (NaCl, 1–10 g/L) increased, the PV_i and PV_c for each solute and their differences decreased, presenting a leftward shift and steeper slope of the BTCs. This result is in accordance with the findings of Cui et al. [10] in CaCl₂ transport (0.25–0.50 mol/L) in homogeneous heavy loams. Specifically, our findings indicate greater retardation in Na⁺ transport in acidic red soil, in comparison to soluble salt and Cl⁻. One possible reason is the higher net negative charge in acidic red soil [11], which attracts and retains more Na⁺ on soil particles. Another reason could be the displacement of Mg²⁺ and Ca²⁺ from the soil surface by Na⁺, slowing down Na⁺ movement [35]. Saygin et al. [3] also pointed out that anions exhibit more mobile behavior than cations, resulting from the repulsion of anions by the electrostatic bilayer colloid. In comparison to R_salt and R_Cl, the higher R_Na estimated from the MIM in our study provides further evidence to support this result.

Our findings demonstrate that MIM can provide an accurate description of solute transport in red soil under low-quality water irrigation (Table 4). As the irrigation water salinity increased, D decreased as a power function, while v and R decreased linearly. This can also demonstrate that higher salinity of irrigation water could decline the adsorption of soluble salt, Na⁺, and Cl⁻, as well as their dispersion in red soil. Similarly, Li and Wang [36] observed a decrease in D for a CO(NH₂)₂ concentration of 1% compared to 0.1% at the same soil moisture content in brown loam. In the present study, β and ω can quantify the physical non-equilibrium process for the solute transport in red soil, showing a non-monotonic relationship with the irrigation water salinity. S1 slightly increased β by 1.0–18.0% compared to S2–S10, and ω approached zero in S1 and S10 for soluble salt transport, which possibly contributed to the difference in the irrigation water salinity and absolute breakthrough time [37,38]. Additionally, our findings showed that Pe > 30.6 and α_s approached zero in all BTCs, reflecting the importance of convection in solute transport in red soil under low-quality water irrigation. Based on the temporally relaxed theory of Fick’s Law, Lin et al. [39] introduced two relaxation times to develop a novel convection-dispersion equation for evaluating solute transport. Their findings indicate that the relaxation times can be proportional to Pe, suggesting that an increase in Pe can reflect the higher probability of particle collisions and adsorption in porous media. In our study, higher Pe was detected in Na⁺ and Cl⁻ BTCs for S10 and in salt BTC for S2, but the largest R for three solutes was found in S1 (Table 4). This indicates that the Pe could not reliably reflect the particle collisions and adsorption in the red soil under low-quality water irrigation.

As for our unsaturated soil infiltration experiment, there was a good linear relationship between the irrigation water salinity and the soil solute concentrations. Similarly to the distributions of the absolute soil solute concentrations, C_r-salt, C_r-Na, and C_r-Cl initially increased and then decreased among all treatments in the 0–40 cm soil profile, with more pronounced variations (Figure 3). This nonlinear behavior may be caused by leaching and changes in the soil structure. Dispersed clay particles tend to migrate with water, facilitating the formation of micropores [18,40]. Prolonged immersion can lead to the swelling, loosening, and loss of soil particles in the upper soil layers [41]. Furthermore, the combined effects of water pressure and soil gravity potential can cause an increase in bulk density in the deeper soil column [42]. Yan et al. observed a reduction in bulk density at the bottom soil column due to soil particle loss through small pores [18]. A significant negative correlation was observed between C_r and irrigation water salinity in the 0–20 cm soil layer. This could be attributed to weaker adsorption of soluble salt, Na⁺, and Cl⁻ and smaller dispersion from the higher salinity of irrigation water in the red soil, promoting solute transport to deeper layers. The variations in C_r, v, D, and R for the three solutes were lower in the irrigation water with higher salinity. In our soil solute breakthrough experiment under steady flow, the irrigation water salinity showed a significant negative correlation with PV_i, PV_c, and PV_Cf = 0.5. It should be noted that there is a difference in the relative concentrations of unsaturated and saturated soil. Actually, due to the negligible impact of molecular diffusion on dispersion in miscible displacement, combined with a small λ (0.045 cm < λ < 0.490 cm), there was no significant difference between C_r and C_f [23].

Many studies have demonstrated that solute transport and its parameters in soil are affected by both the irrigation water salinity and the solute type [43,44]. The complex process cannot be fully explained by a single parameter but depends instead on the integrated interaction of several factors. In this study, the primary factors affecting salt distribution in the 0–40 cm soil profile were v > soil pH > β_salt (Figure 5a), while the key factors were R_Cl > soil pH > sorptivity > β_salt in the 0–20 cm soil profile (Figure 5b). Similarly, Koestel et al. [45] found that v and soil porosity were the important factors influencing soil solute transport under multiple steady-state conditions. v could alter the PV_i, PV_c, and their difference (PV_c − PV_i), and higher v increased the outflow concentration [46]. Moreover, salt transport in soil is driven by water movement, and sorptivity can reflect the influence of the soil matric potential on water movement [47]. Irrigation water salinity can modify soil porosity, producing a difference in the infiltration capacity to affect salt transport [48]. It should be noted that the infiltration can be governed by the soil matric potential at the initial stage of infiltration and can vary with salt concentration [49]. Therefore, the sorptivity could be a key factor relative to soil salt distribution in the 0–20 cm soil profile.

Furthermore, non-specific adsorption of inert Cl⁻ often exhibits negative adsorption in soils dominated by constant charge surfaces [50], potentially influencing soil salt distribution. Typically, Cl⁻ is considered a conservative solute with R_Cl < 1.0 in most soils (e.g., Pei et al. [51]). However, R_Cl > 1.0 observed in our study (Table 4) is inconsistent with this typical result. This discrepancy may be attributed to the presence of both positive and negative charges of red soil, generating electrostatic attraction, and resulting in Cl⁻ adsorption [11]. In this study, the adsorption of salt transport in red soil under low-quality water irrigation appears to be driven by the adsorption of Cl⁻ (R_salt = 0.118 R_Na + 0.839 R_Cl + 0.068).

Soil pH could be a factor influencing salt distribution in the 0–40 cm soil profile and 0–20 cm soil layer. Some studies [12,52] noted that soil pH can affect ion or salt adsorption by altering the electric charge of soil. Our previous study [14] showed that irrigation water salinity significantly affected pH in the 0–35 cm soil profile and the soil pH initially decreased and then increased with increasing irrigation water salinity. The variation in soil pH induced by the irrigation water solute could potentially modify the charge of red soil. In general, higher soil pH is associated with lower adsorption of Cl⁻ in variable charge soils (such as red soil) and negative adsorption can be observed at high pH levels [50]. Our finding is consistent with these results. In our study, S10 had the lowest R_Cl (Table 4), corresponding to the highest pH in the 0–20 cm soil layer [14], probably explaining the variation in the soil salt distribution in this soil layer.

β_salt played a crucial role in salt distribution at the soil depths of 0–40 cm, 0–20 cm, and 20–40 cm (Figure 5), demonstrating that mobile water for salt transport is related to the soil salt distribution. Our study detected β_salt as the sole variable accounting for 53.5% of the variation in soil salt distribution in the 20–40 cm soil layer (Figure 5c). The possible reason is that the change in C_r-salt in the 20–40 cm soil layer was more stable than that in the upper soil layer (Figure 2a). Moreover, higher irrigation water salinity had a lower soil water potential gradient due to a reduction in the soil solute potential, impeding water movement through the micropores between soil particles [53]. Moreover, irrigation water with higher salinity contained more solutes, probably blocking soil pores to reduce the fraction of mobile water in the soil. Consequently, in unsaturated red soil irrigated with higher water salinity, there was a reduction in water movement at the deeper soil layer, as evidenced by the lower sorptivity [14]. Additionally, this influence of β_salt in the soil salt distribution at various depths can also support that the MIM can provide more accuracy in the actual solute transport in red soil. However, the solute type used in this study was relatively simple, and the effects of irrigation water salinity on soil organic matter and microbial activity could influence solute transport in red soil [54,55]. Further research is needed to investigate additional factors, such as the composition of multiple ions, soil organic matter, and microbial activity. Additionally, the long-term effects of low-quality water irrigation on soil properties and solute transport should also be considered.