Preamble

Arid Zone Research Vol. 42 No. 7 Jul. 2025

Simulation and Regulation of Groundwater Depth Effects on Soil Water-Salt Transport in Arid Regions: A Case Study of Representative Farmland in the Riparian Zone of the Lower Aksu River

LU Li¹,², GUO Jianhua³, WANG Younian⁴

¹Institute of Water Resources and Environment, Jilin University, Changchun, Jilin, China
²College of Geology and Mining Engineering, Xinjiang University, Urumqi, Xinjiang, China
³Haikou Marine Geological Survey Center, China Geological Survey, Haikou, Hainan, China
⁴Xinjiang Water Resources and Hydropower Survey and Design Institute, Urumqi, Xinjiang, China

Abstract: Salinization in downstream irrigation areas of arid watersheds exacerbates soil degradation, reduces crop yields, and increases river salinity, thereby constraining agricultural production and ecological stability. These problems are influenced by groundwater depth and improper irrigation and drainage management, making scientifically formulated soil water-salt regulation measures essential. This study conducted field experiments in a typical farmland plot in the riparian zone of the lower Aksu River. Based on dynamic observations and field survey data, an unsaturated model was established using HYDRUS-1D software to simulate soil water-salt transport patterns during the cotton growing season, determine reasonable regulation schemes, and explore the relationship between stable groundwater evaporation depth and riparian soil structure. Results showed that the identification and validation accuracies for soil moisture content and total dissolved solids were 0.862 and 0.752, respectively, with root mean square errors of 0.033 and 0.008, indicating high model reliability. Irrigation infiltration recharge accounted for 85.66% of total soil water recharge, introducing 127.164 mg·cm⁻² of salt, while soil water discharge to the phreatic zone accounted for 59.67% of total discharge, removing 267.78 mg·cm⁻² of salt. The water balance error was 9.2% and the desalination rate was 33.89%. Considering crop water requirements and dynamic soil salinity changes, setting the irrigation water depth at 70 cm while controlling groundwater depth at approximately 220 cm can effectively reduce root zone soil salinity. In sandy-loam interbedded structures, the position of the loam layer has minimal effect on the critical groundwater evaporation depth (150 cm), but significantly influences the stable evaporation depth and actual evaporation rate. The closer the loam layer is to the surface, the shallower the stable evaporation depth and the smaller the actual evaporation. These findings provide a reference for salinization prevention and rational water resource allocation in arid regions.

Keywords: water-salt regulation; cotton; arid regions; riparian zone; groundwater depth; stable evaporation depth

Introduction

Northwest China's arid inland basins primarily feature alluvial-proluvial landforms, with water resources derived from alpine snowmelt that infiltrates and evaporates in desert oases. As groundwater flows through aquifers rich in soluble minerals, it undergoes leaching with host rocks, gradually increasing soluble salt content during downstream discharge. In oasis agricultural areas, irrigation represents another major pathway affecting field salt distribution, particularly in regions with well-developed watersheds and shallow groundwater depths, where strong evaporation drives capillary rise of water and salts to the surface, causing soil salinization. The Aksu region, located at the southern foothills of the central Tianshan Mountains and the northwestern margin of the Tarim Basin, features the Aksu River as its main watercourse and serves as Xinjiang's largest long-staple cotton production base.

Since the 1990s, market-oriented policies have stimulated farmers to reclaim wasteland, converting abandoned saline-alkali lands in the lower Aksu River into cultivated farmland. To mitigate salt stress on crops, excessive irrigation water is often used for "salt leaching," which temporarily suppresses salinization but raises groundwater levels and alters the water-salt cycle in the vadose zone vegetation, causing problems such as farmland "salt return" and Aksu River salinization. Clearly, rational irrigation systems and orderly groundwater regulation are critical for effective water resource utilization and solving soil salinization in the study area, necessitating research on hydrological cycling and soil salt migration during irrigation and water level regulation processes.

Previous studies have used HYDRUS models to simulate groundwater level effects on soil water-salt distribution and proposed improved irrigation management and groundwater regulation strategies. While these simulations and experiments help understand irrigation impacts on soil water-salt cycling, they lack crop yield constraints, limiting practical feasibility. Although some research has explored water-salt regulation effects on soil salt distribution and crop yield through modified irrigation volumes and groundwater depths, these measures primarily focused on irrigation methods and strategies. Few scholars have investigated soil salinity and crop yield changes under different groundwater depths and irrigation strategies, revealing the critical role of groundwater regulation in salinization prevention and water resource management. Furthermore, for the variable soil structures in riparian zones, clarifying soil texture effects in water-salt regulation models—particularly regarding critical and stable evaporation depths—remains a key issue. Soil physical properties such as pore structure and hydraulic conductivity significantly influence vertical water migration, capillary rise, and dynamic water-salt distribution, thereby determining critical and stable depths of groundwater evaporation. Incorporating dynamic soil texture simulation can deepen understanding of water-salt migration mechanisms in riparian zones, improve model adaptability under changing conditions, and provide more refined support for salinization prevention and water resource optimization.

This study established an unsaturated model using HYDRUS-1D based on local cultivation calendars, soil salinity surveys, and cotton yield investigations. The objectives were: (1) to calibrate the model using field monitoring data, (2) to simulate soil water-salt transport patterns during the cotton growing season, (3) to propose field water-salt regulation schemes balancing water-saving irrigation and salinization prevention, (4) to quantify groundwater critical and stable evaporation depths under different water level conditions based on the calibrated model, and (5) to conduct sensitivity analysis on how layered soil structures affect these evaporation depths. While results are specific to the experimental area, the methodology's physical process considerations may benefit other regions.

1 Study Area and Research Methods

1.1 Study Area Overview

The study was conducted in the riparian zone of the lower Aksu River where groundwater is shallow and soil salinization is severe. The experimental site is located in Xinkailing Town, Regiment 13, Aral City, on the alluvial plain of the lower Aksu River (80°47′43.90″E, 40°31′06.60″N). The area experiences a typical warm temperate continental arid climate with an average annual temperature of 11.2–12.2°C, frost-free period of 193–310 days, annual precipitation of 60.5–98.2 mm, and sunshine duration of 2761–2827.7 hours. The experimental field area is 4.5×10⁴ m².

The experimental soil has a layered structure dominated by sandy and loamy soils, classified as crusted solonchak, salinized meadow soil, and meadow solonchak. The main crop is cotton, which has long relied on shallow groundwater for agricultural irrigation. Due to channel sedimentation and aging infrastructure, combined with the backwater effect of the Aksu River, groundwater depth during the crop growing season ranges 104–156 cm with mineralization of 2.35–2.57 g·L⁻¹. Root zone (0–60 cm) soil salinity maintains approximately 0.68%, classified as moderately salinized with surface salt accumulation according to Xinjiang soil salinization classification standards.

The soil profile consists of four layers: silt loam (0–0.35 m), loam (0.35–0.8 m), fine sand (0.8–1.8 m), and fine sand below 1.8 m, representing typical riparian zone characteristics. Sensors were installed at 15 cm, 40 cm, 70 cm, 110 cm, and 150 cm depths to monitor soil moisture content, electrical conductivity, and temperature.

1.2 Experimental Design

The cotton variety "Tuonong 1" was planted in a "wide-narrow row" pattern with 30 cm narrow row spacing and 50 cm wide row spacing. Water level and quality monitoring wells were arranged in both wide and narrow rows. To reduce calculation errors from water outflow during irrigation, field ridges were covered with plastic film (ridge width 25 cm), while the cotton planting area remained uncovered. The experiment ran from sowing on April 20 to harvest on October 8, 2021. Irrigation timing was set for May 25, June 25, July 25, and August 25, with each irrigation depth of 30 cm using flood irrigation with local groundwater (dissolved solids 1.048 g·L⁻¹).

1.3 Observation Content and Methods

1.3.1 Meteorological and Three-Parameter Data

Meteorological data were obtained from an automatic weather station (HOBO H21-001, Onset) including maximum/minimum temperature, relative humidity, wind speed, solar radiation, air pressure, precipitation, and sunshine duration. Reference evapotranspiration (ET₀) was calculated daily using the Penman-Monteith formula. Soil moisture content, electrical conductivity, and temperature were monitored in real-time using WET-2 sensors (Delta T Devices) at 5-minute intervals. Annual rainfall and ET₀ values were derived from these measurements. Precipitation was concentrated in June-August, with daily average temperature showing a clear decreasing trend during this period.

1.3.2 Soil Physical Parameters

Before the experiment, soil texture distribution and layer positions were determined through drilling. Soil samples were collected at depths of 0–0.35 m, 0.35–0.8 m, and 0.8–1.8 m. Soil bulk density and saturated hydraulic conductivity were measured using the ring knife method, while particle size distribution was determined using a laser particle size analyzer (Microtrac Inc., USA). The Rosetta function predicted soil water retention parameters (Table 1). Soil particle fractions showed: 0–0.35 m layer with 74.65% sand, 23.63% silt, and 1.72% clay; 0.35–0.8 m layer with 65.82% sand, 30.26% silt, and 3.92% clay; 0.8–1.8 m layer with 88.58% sand, 10.96% silt, and 0.46% clay.

1.3.3 Groundwater Monitoring

PVC pipes wrapped with gauze and perforated served as "wells" for groundwater level and quality monitoring. Groundwater depth ranged 104–156 cm with mineralization of 2.35–2.57 g·L⁻¹. Shallower depths corresponded to higher mineralization.

1.4 Soil Water-Salt Dynamic Model

HYDRUS-1D was selected for its user-friendly interface, modular functionality, suitability for secondary development, and embedded Fortran programming, with wide applications in agriculture, hydrogeology, water conservancy, and environmental fields. The simulation considered crop root growth and water uptake, using the Galerkin finite element method for equation solving. The inversion module employed the Levenberg-Marquardt parameter optimization algorithm to solve soil hydraulic and solute transport parameters, excluding heat transfer. The simulation process is illustrated in Figure 5.

1.4.1 Soil Water Movement Equation

Assuming isotropic, rigid soil matrix, the Richards equation with source-sink terms was used:

$$\frac{\partial \theta}{\partial t} = \frac{\partial}{\partial z}\left[K(h)\left(\frac{\partial h}{\partial z} + 1\right)\right] - S$$

where $\theta$ is volumetric water content (cm³·cm⁻³), $t$ is time (d), $h$ is pressure head (cm, positive for saturated, negative for unsaturated), $z$ is vertical coordinate (positive upward), $K(h)$ is unsaturated hydraulic conductivity (cm·d⁻¹), and $S$ is the source-sink term representing root water uptake.

The soil water characteristic curve and unsaturated hydraulic conductivity function $K(h)$ were described using the van Genuchten-Mualem model:

$$\theta(h) = \theta_r + \frac{\theta_s - \theta_r}{[1 + |\alpha h|^n]^m}$$

$$K(h) = K_s S_e^l [1 - (1 - S_e^{1/m})^m]^2$$

where $K_s$ is saturated hydraulic conductivity (cm·d⁻¹), $\theta_s$ is saturated water content (cm³·cm⁻³), $\theta_r$ is residual water content (cm³·cm⁻³), $S_e$ is effective saturation, and $\alpha$, $m$, $n$ are empirical shape parameters.

The root water uptake model defines the source-sink term $S$ as:

$$S = b(x) \cdot \alpha(h) \cdot T_p$$

where $b(x)$ is the normalized water uptake distribution function, $\alpha(h)$ is the water stress response function, and $T_p$ is potential transpiration. The Feddes model provided threshold parameters: $h_1 = -10$ cm, $h_2 = -25$ cm, $h_{3H} = -200$ cm, $h_{3L} = -600$ cm, $h_4 = -14000$ cm, with maximum root depth of 60 cm.

1.4.2 Soil Solute Transport Equation

Based on mass conservation, the convection-dispersion equation considering soil adsorption but excluding root uptake was used:

$$\frac{\partial (\theta c_l + \rho_b c_s)}{\partial t} = \frac{\partial}{\partial z}\left(\theta D_e \frac{\partial c_l}{\partial z} - q_w c_l\right) + \phi$$

where $c_s$ is adsorbed solute concentration (mg·g⁻¹), $\rho_b$ is soil bulk density (g·cm⁻³), $D_e$ is hydrodynamic dispersion coefficient (cm²·d⁻¹), $c_l$ is solute concentration in water (mg·mL⁻¹), $q_w$ is water flux (cm·d⁻¹), and $\phi$ represents source-sink terms.

1.4.3 Canopy Transpiration and Soil Evaporation Calculation

Reference evapotranspiration ($ET_0$) was calculated using the Penman-Monteith formula. Potential crop transpiration ($T_p$) and soil evaporation ($E_p$) were partitioned using leaf area index and extinction coefficient:

$$E_p = ET_c \cdot e^{-K_c \cdot LAI}$$

$$T_p = ET_c \cdot (1 - e^{-K_c \cdot LAI})$$

$$ET_c = ET_0 \cdot K_c$$

where $ET_c$ is actual crop evapotranspiration without water stress (mm), $LAI$ is leaf area index, $K_c$ is crop coefficient, and the extinction coefficient $\lambda$ for cotton is typically 0.85.

1.4.4 Soil Water-Salt Balance Calculation

The water balance equation in a control volume is:

$$P + I + U - ET - D - R = \Delta S_w$$

where $P$, $I$, $U$, $ET$, $D$, $R$, and $\Delta S_w$ represent precipitation, irrigation, groundwater recharge, evapotranspiration, deep drainage, runoff, and soil water storage change, respectively (all in mm). No runoff occurred during the experiment.

The salt balance equation is:

$$I_s + U_s - L_s - S_s = \Delta S_s$$

where $I_s$, $U_s$, $L_s$, $S_s$, and $\Delta S_s$ represent salt input from irrigation, salt from groundwater recharge, salt discharged by deep drainage, soil salt storage change, and salt leaching, respectively (mg·cm⁻²). Positive $S_s$ indicates salt accumulation.

1.5 Model Setup

1.5.1 Initial and Boundary Conditions

Initial condition: $\theta(z,0) = \theta_0(z)$

Upper boundary (atmospheric): $-K(h)\left(\frac{\partial h}{\partial z} + 1\right) = q_s(t)$

Lower boundary (variable head): $h(Z,t) = h_b(t)$

The upper boundary was a concentration flux boundary: $q_s c_s(t)$, where $c_s$ is irrigation water concentration (mg·mL⁻¹). The lower boundary was a concentration boundary: $c(Z,t) = C_b$, where $C_b$ is phreatic water concentration (mg·mL⁻¹).

1.5.2 Spatial and Temporal Discretization

The simulation depth was 180 cm, discretized into 70 layers with 1 cm node spacing based on measured soil textures. The simulation period covered the cotton growing season (April 20–October 8, 201 days) with daily time steps, minimum step of 0.01 d, and maximum step of 1 d.

1.5.3 Parameter Acquisition

Initial soil hydraulic parameters ($\theta_s$, $\theta_r$, $K_s$, $\alpha$, $n$, $l$) were predicted using the Rosetta function based on soil particle content (Table 1). Root water uptake thresholds followed the Feddes model with literature values for cotton. The maximum root depth was set at 60 cm, with logistic growth model assuming roots reach maximum depth at mid-growing season.

1.5.4 Model Calibration and Validation

The trial-and-error method was used for coarse parameter adjustment, followed by inverse modeling for fine-tuning. Irrigation period data were used for calibration/validation due to greater water-salt variation. Model accuracy was evaluated using $R^2$ and RMSE—lower RMSE indicates higher precision.

Simulated and observed soil moisture and TDS at different depths are shown in Figure 8. Most data points fell near the 1:1 line, with minor deviations primarily due to surface soil structure sensitivity to climate and human activities. Table 2 shows validation results: $R^2$ for moisture ranged 0.351–0.682; for TDS, 0.322–0.612. Lower $R^2$ values at depth resulted from small variations relative to baseline values. Overall, the model demonstrated high accuracy with calibrated parameters meeting requirements (Table 3).

2 Results

2.1 Soil Water Balance Analysis

Using the calibrated model, soil water balance during the growing season was analyzed (Table 4). Total irrigation was 1269.40 mm, accounting for 85.66% of soil water recharge, while groundwater recharge contributed 14.34% (212.42 mm). Soil water discharge occurred primarily through bottom boundary recharge to the phreatic zone (59.67%, 802.94 mm), with actual evapotranspiration accounting for 40.33% (542.61 mm). The water balance error was 9.2%. Post-irrigation, soil water mainly discharged as groundwater recharge, reducing groundwater depth and subsequently causing water loss through open drains or lateral flow, leading to secondary water waste and river salinization. Therefore, determining rational irrigation volumes for "water-saving and salt-control" is essential.

2.2 Soil Salt Balance Analysis

The calibrated model revealed salt balance dynamics (Table 5). Spring irrigation effectively controlled seasonal salt accumulation, with salt discharge reaching 44.258 mg·cm⁻² in March. Total salt input through irrigation and groundwater recharge was 199.99 mg·cm⁻², while total salt discharge was 267.78 mg·cm⁻², achieving a desalination efficiency of 33.89%. Although irrigation ceased in September-October, strong evaporation with shallow groundwater continued driving salt accumulation via capillary rise.

2.3 Model Application and Field Irrigation System Optimization

Water-salt balance analysis indicates that irrigation serves dual purposes: meeting crop water demand and leaching salts to reduce soil salinity. Excessive irrigation raises groundwater levels, increases ineffective surface evaporation, and causes secondary salinization, while high infiltration water laterally recharges the Aksu River, wasting water and deteriorating water quality. Rational irrigation and groundwater regulation are thus the most economical and effective methods for preventing secondary salinization.

Using 2021 field data and the Aksu Agricultural Technology Extension Center's cultivation calendar, crop water consumption patterns were determined (Figure 10). Total crop transpiration was 422 mm and inter-plant evaporation was 121.15 mm, requiring 543.15 mm irrigation water for the entire growing season. Spring irrigation ("spring leaching") was completed before April 25, calculated as the difference between total irrigation and growing season consumption (700 mm - 543.15 mm = 156.85 mm).

The relationship between irrigation depth and root zone salinity was simulated (Figure 9). When irrigation depth was below 70 cm, root zone salinity decreased rapidly; beyond 70 cm, salinity reduction was minimal, indicating water waste. Therefore, 70 cm is the optimal irrigation depth. Based on this optimization and cotton salt tolerance thresholds (Figure 12), the reasonable groundwater control depth was determined to be 220 cm, which effectively reduces root zone salinity while ensuring crop yield.

2.4 Yield-Groundwater Depth Based Water-Salt Regulation Mode

Groundwater depth and quality are the most active factors in salinization processes, serving as both salt carriers and transport drivers. The relationship between root zone soil salinity and groundwater depth was established (Figure 11). When soil water mineralization was below 2.03 g·L⁻¹, soil salinity remained below the 2.5 g·kg⁻¹ tolerance threshold. At the regulation end, soil water mineralization was 2.59 g·L⁻¹, slightly above the threshold but with minimal yield impact. Therefore, maintaining groundwater depth at approximately 220 cm under the optimized irrigation system can effectively reduce root zone salinity while ensuring crop production.

2.5 Stable and Critical Evaporation Depths in Riparian Zones

During non-irrigation periods (spring/autumn), shallow groundwater (~120 cm) drives soil salt return through evaporation. Determining stable and critical evaporation depths is crucial for understanding salinization physics. Riparian soil structure varies due to floodplain transverse slopes, causing the loam interlayer position to change with distance from the riverbank (Figure 13). To study this effect, simulations were conducted with the loam layer at different positions (40 cm, 50 cm, 60 cm, 70 cm, 80 cm, 90 cm, 100 cm, 110 cm, 150 cm) under varying water table depths.

Previous studies indicate the limiting evaporation depth for single-structure sandy loam in Aksu is approximately 2.5 m. Simulations started from 2.5 m with 10 cm increments. The relationship between actual evaporation ($E_a$) and potential evaporation ($E_p$) versus groundwater depth is shown in Figure 14. When the water table was below the loam layer, stable evaporation depth increased with deeper loam position but with diminishing increments, reaching maximum at 80 cm. Beyond this, stable evaporation depth remained constant regardless of loam position. Actual evaporation also increased with deeper loam layers but with decreasing increments (Figure 15). The loam layer position had minimal effect on critical evaporation depth (150 cm).

In the vadose zone, water content and unsaturated hydraulic conductivity may be discontinuous, but soil suction must be continuous. When groundwater depth is below the loam layer, water from underlying fine sand moves to the loam interface. At the same water content, fine sand suction ($S_1$) is lower than loam suction ($S_2$), allowing water entry into the loam layer. However, fine sand hydraulic conductivity exceeds loam conductivity, slowing the wetting front in the loam layer and controlling water transmission rates from underlying sand. As water continues moving upward, silt loam suction ($S_3$) is lower than loam suction ($S_2$) at the same water content, and loam conductivity is lower than silt loam, causing water to accumulate in the loam layer with difficulty moving upward. Therefore, evaporation-available water is primarily controlled by the upper silt loam layer. As the loam layer moves down, thicker silt loam enhances water supply capacity and increases evaporation, but capillary rise limitations cause evaporation to stabilize with continued thickness increase.

Stable evaporation depth is defined as the maximum depth where potential evaporation equals actual evaporation, with continuous water supply being critical. When stable evaporation depth is below the loam layer, silt loam water replenishment is limited by loam permeability. The loam's low permeability restricts water supply, causing silt loam water consumption to exceed replenishment. As the loam layer moves upward, silt loam water supply capacity further decreases, intensifying water deficit. To maintain new water balance, stable evaporation depth must decrease to enhance groundwater replenishment and compensate for surface water shortage. Therefore, farmland management should combine groundwater level control with measures like straw mulching and deep tillage to create evaporation barriers, improve water retention, and inhibit salt accumulation.

3 Conclusions

Based on field dynamic monitoring and survey data, this study used the HYDRUS model to simulate groundwater depth effects on water-salt transport in salinized soils of riparian irrigation areas in arid regions, proposing rational regulation schemes and discussing stable and critical evaporation depths under different scenarios. The main conclusions are:

1. During the simulation period, irrigation infiltration accounted for 85.66% of soil water recharge (1269.40 mm), while groundwater recharge contributed 14.34% (212.42 mm). Soil water discharge occurred primarily through bottom boundary recharge to the phreatic zone (59.67%, 802.94 mm), with actual evapotranspiration accounting for 40.33% (542.61 mm). The water balance error was 9.2%, indicating limited soil water storage capacity.

2. Spring irrigation effectively controlled seasonal salt accumulation, discharging 44.258 mg·cm⁻² of salt. Total salt input through irrigation and groundwater recharge was 199.99 mg·cm⁻², while total salt discharge was 267.78 mg·cm⁻², achieving a desalination efficiency of 33.89%.

3. Based on the relationship between root zone salinity and irrigation depth, the optimal irrigation depth for cotton was determined to be 70 cm. Using the optimized irrigation system and considering the relationship between groundwater depth and soil salinity during regulation, the reasonable groundwater control depth for downstream irrigation areas was established as 220 cm.

4. In sandy-loam interbedded structures, the loam layer position minimally affects critical evaporation depth (150 cm) but significantly influences stable evaporation depth and actual evaporation. The closer the loam layer is to the surface, the shallower the stable evaporation depth and the smaller the actual evaporation.

References

[1] Lu Li, Ge Yanyan, Li Sheng, et al. Hydrochemical characteristics and enrichment mechanisms of high arsenic groundwater in the Aksu River Basin, Xinjiang[J]. Arid Zone Research, 2025, 42(2): 258-273.

[2] Lei Y, Liu Y, Sun Z, et al. Influences of paleoclimatic environment and hydrogeochemical evolution on groundwater salinity in an arid inland plain in northwestern China[J]. Applied Geochemistry, 2023, 154: 105688.

[3] Lu L, Li S, Wu R, et al. Study on the scale effect of spatial variation in soil salinity based on geostatistics: A case study of Yingdaya River Irrigation Area[J]. Land, 2022, 11(10): 1697.

[4] Lu L, Li S, Gao Y, et al. Analysis of the characteristics and cause analysis of soil salt space based on the basin scale[J]. Applied Sciences Basel, 2022, 12(18): 9022.

[5] Xie T, Liu X, Sun T. The effects of groundwater table and flood irrigation strategies on soil water and salt dynamics and reed water use in the Yellow River Delta, China[J]. Ecological Modelling, 2011, 222(2): 241-252.

[6] Dou X, Shi H, Li R, et al. Simulation and evaluation of soil water and salt transport under controlled subsurface drainage using HYDRUS-2D model[J]. Agricultural Water Management, 2022, 273: 107900.

[7] Guo S, Li X, Šimůnek J, et al. Experimental and numerical evaluation of soil water and salt dynamics in a corn field with shallow saline groundwater and crop season drip and autumn post-harvest irrigations[J]. Agricultural Water Management, 2024, 305: 109119.

[8] Che Z, Wang J, Li J. Effects of water quality, irrigation amount and nitrogen applied on soil salinity and cotton production under mulched drip irrigation in arid Northwest China[J]. Agricultural Water Management, 2021, 247: 106738.

[9] Wang H, Feng D, Zhang A, et al. Effects of saline water mulched drip irrigation on cotton yield and soil quality in the North China Plain[J]. Agricultural Water Management, 2022, 262: 107405.

[10] Li W, Wang Z, Zhang J, et al. Soil salinity variations and cotton growth under long-term mulched drip irrigation in saline-alkali land of arid oasis[J]. Irrigation Science, 2022, 40(1): 103-113.

[11] Chen S, Mao X, Shang S. Response and contribution of shallow groundwater to soil water/salt budget and crop growth in layered soils[J]. Agricultural Water Management, 2022, 266: 107574.

[12] Xue J, Huo Z, Wang F, et al. Untangling the effects of shallow groundwater and deficit irrigation on irrigation water productivity in arid region: New conceptual model[J]. Science of the Total Environment, 2018, 619: 1170-1182.

[13] Zhu Huiyi. Incentive of land use change: A case study on the variations of agricultural factor productivity in Xinjiang[J]. Acta Geographica Sinica, 2013, 68(8): 1029-1037.

[14] Sheng Y, Tian X, Qiao W, et al. Measuring agricultural total factor productivity in China: Pattern and drivers over the period of 1978-2016[J]. Australian Journal of Agricultural and Resource Economics, 2020, 64(1): 82-103.

[15] Yan W, Li F, Zhao Y. Determination of irrigation water quantity and its impact on crop yield and groundwater[J]. Agricultural Water Management, 2022, 273: 107900.

[16] Wang W, Chen Y, Wang W, et al. Water quality and interaction between groundwater and surface water impacted by agricultural activities in an oasis desert region[J]. Journal of Hydrology, 2023, 617: 128937.

[17] Foster S, Pulido-Bosch A, Vallejos A, et al. Impact of irrigated agriculture on groundwater recharge salinity: A major sustainability concern in semi-arid regions[J]. Hydrogeology Journal, 2018, 26(8): 2789-2800.

[18] Feng G, Zhang Z, Wan C, et al. Effects of saline water irrigation on soil salinity and yield of summer maize (Zea mays L.) in subsurface drainage system[J]. Agricultural Water Management, 2017, 193: 205-213.

[19] Ma Z, Wang W, Zhang Z, et al. Assessing bare soil evaporation from different water table depths using lysimeters and a numerical model in the Ordos Basin, China[J]. Hydrogeology Journal, 2019, 27(7): 2707-2718.

[20] Fahle M, Dietrich O. Estimation of evapotranspiration using diurnal groundwater level fluctuations: Comparison of different approaches with groundwater lysimeter data[J]. Water Resources Research, 2014, 50(1): 273-286.

[21] Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences. Research on Xinjiang Soil and Land Resources[M]. Beijing: Science Press, 1991: 1-37.

[22] Xinjiang Agricultural Department, Xinjiang Soil Survey Office. Xinjiang Soil[M]. Beijing: Science Press, 1996: 151-521.

[23] Wang Lu. Study on the Law of Water and Salt Migration and Dynamic Regulation of Diving Level in the Vadose Zone in the Salinized Area of the Aksu River Basin[D]. Urumqi: Xinjiang University, 2020.

[24] Qi L, Fan J, Shao M, et al. Simulation and verification of soil moisture of root distribution functions for alfalfa[J]. Transactions of the Chinese Society of Agricultural Engineering, 2009, 25(4): 24-29.

[25] Hoffman G, Genuchten M. Soil properties and efficient water use: Water management for salinity control[C]//Taylor H, Jordan W, Sinclair T. Limitations to Efficient Water Use in Crop Production. Madison: American Society of Agronomy, 1983: 73-85.

[26] Šimůnek J, Suarez D. Modeling of carbon dioxide transport and production in soil: 1. Model development[J]. Water Resources Research, 1993, 29: 487-497.

[27] Zhang Lizhen, Cao Weixing, Zhang Siping, et al. Characterizing root growth and spatial distribution in cotton[J]. Chinese Journal of Plant Ecology, 2005, 29(2): 266-273.

[28] Belmans C, Wesseling J, Feddes R. Simulation model of the water balance of a cropped soil[J]. Journal of Hydrology, 1983, 63(3-4): 271-286.

[29] Zhang Y, Zhang F, Fan J, et al. Effects of drip irrigation technical parameters on cotton growth, soil moisture and salinity in Southern Xinjiang[J]. Transactions of the Chinese Society of Agricultural Engineering, 2020, 36(24): 107-117.

[30] Feddes R, Kowalik P, Kolinska-Malinka K, et al. Simulation of field water uptake by plants using a soil water dependent root extraction function[J]. Journal of Hydrology, 1976, 31(1-2): 13-26.

[31] Husan Tumarbay, Wu Z, Su L, et al. Numerical simulation of soil water-salt movement on drip irrigation cotton under film[J]. Soils, 2012, 44(4): 665-670.

[32] Liu Hongguang, Bai Zhentao, Li Kaiming. Soil salinity changes in cotton field under mulched drip irrigation with subsurface pipes drainage using HYDRUS-2D model[J]. Transactions of the Chinese Society of Agricultural Engineering, 2021, 37(2): 130-141.

[33] Moriasi D, Arnold J, Van L, et al. Model evaluation guidelines for systematic quantification of accuracy in watershed simulations[J]. Transactions of the ASABE, 2007, 50(3): 885-900.

[34] Hu Hongchang, Zhang Zhi, Tian Fuqiang, et al. Response of soil salinity and crop growth to irrigation methods in Xinjiang[J]. Journal of Tsinghua University (Science and Technology), 2016, 56(4): 373-380.

[35] DB65/T 4565-2022. Requirements for Main Agronomic Traits of Machine-Picked Cotton[S]. Urumqi: Xinjiang Uygur Autonomous Region Market Supervision Administration, 2022.

[36] Lu Li, Li Sheng, Gao Yuan, et al. Spatial distribution and variation characteristics of soil salinity in the oasis of Weigan and Kuga Rivers[J]. Journal of Arid Land Resources and Environment, 2022, 36(3): 136-142.

[37] Toonen W H J, Kleinhans M G, Cohen K M. Sedimentary architecture of abandoned channel fills[J]. Earth Surface Processes and Landforms, 2012, 37(4): 459-472.

[38] Li Qiang, Liu Tianchao, Li Tong, et al. Hydrogeological and Environmental Geological Survey Report of Aksu River Basin in Xinjiang[R]. Urumqi: Xinjiang Fangyuan Geological and Mineral Exploration Institute (Co., Ltd.), 2019.

[39] Li Yan, Hu Shunjun, Song Yudong. Experimental analysis of phreatic evaporation patterns in the Weigan River Irrigation Area[J]. Journal of Arid Land Resources and Environment, 2004, 18(3): 92-96.

[40] Lai Jianbin, Wang Yongping, Jiang Qinghua, et al. Study on phreatic evaporation under different soil textures[J]. Journal of Northwest A&F University (Natural Science Edition), 2003, 31(6): 153-157.

[41] Gong C, Wang W, Zhang Z, et al. Comparison of field methods for estimating evaporation from bare soil using lysimeters in a semi-arid area[J]. Journal of Hydrology, 2020, 590: 125334.

[42] Zhang Z, Wang W, Wang Z, et al. Evaporation from bare ground with different water table depths based on an in situ experiment in Ordos Plateau, China[J]. Hydrogeology Journal, 2018, 26(5): 1683-1697.