Water and Salt Dynamics in Cultivated, Abandoned, and Lake Systems Under Irrigation Reduction in the Hetao Irrigation District

Abstract

The shifting irrigation reduction in the Hetao Irrigation District and the inability to effectively discharge salts from the system have led to significant changes in salt migration patterns. Based on the integration of long-term field observations (2017–2023) with soil hydrodynamics and solute transport models, this study explored the impact of irrigation reduction on water and salt migration in a cropland–wasteland–lake system. The results indicated that before and after the reduction in irrigation and decline in groundwater levels, the migration rates of groundwater from croplands to wastelands and from wastelands to lakes remained relatively stable, averaging 78% and 40%. During the crop growth period, after irrigation reduction and groundwater level decline, the volume of groundwater recharging lakes from wastelands decreased by 80–120 mm, causing a water deficit in the lakes of 679–789 mm. After irrigation reduction and groundwater level decline, during the crop growth period, 1402 kg/ha of salt remained in the wasteland groundwater, and 597–861 kg/ha of salt accumulated in the cropland groundwater, exceeding previous levels, leading to salinization in the cropland and wasteland groundwater. This study provides insights relevant to managing groundwater and soil salinity in irrigation areas.

1. Introduction

Water scarcity and soil salinization are two major limiting factors for agricultural production and ecological sustainability in arid irrigation regions [1]. The typical characteristics of these regions include low groundwater levels and fragmented land cover, as seen in irrigation areas, such as the Yellow River basin in China [2], the Fergana Valley in Central Asia [3], and the Indo-Gangetic Plain in India [4]. Meanwhile, mismatches between irrigation and drainage systems, coupled with inadequate irrigation management practices, have led to severe soil salinization, threatening agricultural productivity and sustainability [5,6]. As one of China’s three major gravity-fed irrigation zones, Inner Mongolia’s Hetao Irrigation District has reduced its annual water diversion from 5.2 billion m³ (1998) to 4.7 billion m³ under China’s unified Yellow River water management [7]. Since 1998, water-saving projects have been implemented in the HID to improve water distribution and conveyance capacity [8,9]. With the construction of hydraulic engineering projects, significant changes have occurred in the region’s soil, groundwater, and ecological environment [10,11,12]. To address these challenges, the analysis of significant changes in water and salt migration under changing environments is necessary. This study specifically explores water and salt transport under variations in irrigation regimes and irrigation–drainage conditions and validates the efficacy of current irrigation regimes to inform adaptive management decisions for controlling salt transport and build-up in soils.

Numerous studies, both domestically and internationally, have explored water and salt transport in the HID. Yue et al. [13] developed a water–salt balance model for non-farm areas, farmlands, and water bodies and quantitatively analyzed salt migration. Their findings showed that 75% of the desalinized water from farmlands migrated to non-farm areas via groundwater. Wu et al. [14] investigated the salt transport process through the drainage system in the HID and found that excessive irrigation in croplands facilitated salt migration from the cropland groundwater to wastelands. Wang et al. [15] studied salt removal in the Yonglian irrigation area and found that groundwater in wastelands responded strongly to lateral recharge from cropland groundwater. Their water–salt balance analysis revealed that the amount of water migrating to wastelands was 4 times the amount of artificially drained water, while the salt load was 7.7 times that of artificial salt drainage. Ren et al. [4] used the HYDRUS-dualKc model to simulate soil water and salt transport in different vegetation plots and found that the average root zone salt content increased by 52% during the growing season, with wastelands retaining nearly 50% of the salt introduced by irrigation. Hao et al. [16] established a distributed model based on the HYDRUS-EPIC agricultural hydrological model and used it to assess the changes in the root zone soil moisture and salinity during the growing season and their impact on the crop yield. Mao et al. [17] used the SaltMod model and concluded that salt accumulated in the root zone of well-irrigated areas and in non-irrigated land in canal irrigation zones, with salt migrating from the canal irrigation area to well-irrigated areas. However, their study did not address groundwater salt migration between irrigated and non-irrigated lands under canal irrigation. Wang et al. [18] proposed a parameterization scheme for the large-scale land water cycle model LWCMPS_1D and simulated the water cycle and salt transport characteristics of the GS-SPAC system in the HID. Xu et al. [19] coupled the SWAP and MODFLOW models to reveal the interaction between surface water and groundwater, as well as the characteristics of soil water and salt transport.

Several agricultural hydrological models, such as Hydrus [20], SWAP [21], RZWQM2 [22], SaltMod [23], SahysMod [24], and SWAT-Salt [25], have been extended to the irrigation district scale to study the water and salt transport process in other similar semi-arid salinity-impacted watersheds. Hosseini and Bailey [25] used the SWAT-Salt model to identify the key environmental and hydrologic factors that govern the fate and transport of salts in Arkansas River Valley in southeastern Colorado. Yin et al. [26] used a model for vegetation–salinity–groundwater interactions to quantify the soil water and salt exchange fluxes and groundwater dynamics among the cropland–shelterbelt–desert field in the Sangong River watershed, Xinjiang Province of Northwestern China. Kanzari et al. [27] studied water flow and salt transport using the Hydrus_1D model in a semi-arid region in Tunisia. Kutangila [28] and Faye [29] assessed the transport processes of mineral micropollutants in a multi-layered aquifer by constructing a groundwater solute transport model and applying hydrogen and oxygen isotopes. Zha et al. [30] were the first to propose a generalized moisture-based Richards equation, which can be applied to heterogeneous soils by coupling a one-dimensional water balance model with a three-dimensional groundwater model for simulating and predicting soil water and groundwater movement over large areas. Zhu et al. [31] simplified unsaturated zone water movement and solute transport to one-dimensional vertical movement and coupled it with three-dimensional groundwater flow, solute transport, and nitrogen transformation equations to reconstruct a quasi-three-dimensional water–salt transport model (WSMS_Q3D). However, the applicability of these models is often limited due to various constraints, such as the focus on vertical hydrological processes within a simulation unit, canal drainage and salt removal processes, groundwater dynamics, and the extensive parameter requirements, which hinder their broader application. For example, both SWAP and HYDRUS apply to the variably saturated zone and numerically solve the Richards equation for water flow and the advection–dispersion equation (ADE) for solute transport. The SWAP model is very useful for the simulation of generic crop growth and optimal irrigation, while it requires many more parameters for crops. Moreover, the HYDRUS-1D model also has some inadequacies in determining the actual soil evaporation [4]. The land types in the HID are diverse and complex, generally categorized into cropland, non-cropland (wasteland), sand dunes, and lakes. These land types are interspersed, with wastelands located between croplands and around sand dunes and lakes. These four land types are key areas for salt migration and transformation and are crucial for future soil salinization control in the district [32]. Therefore, the application of the aforementioned models to the cropland–wasteland–lake system in the HID may encounter challenges, such as scale mismatches, incomplete model parameters, and boundary conditions not reflecting current realities [33,34]. To address these challenges, this study, based on the HID, aims to develop a groundwater migration model which breaks through the limitations of traditional single-ecosystem research, established at the interfaces of a farmland–wasteland–lake composite ecosystem. Furthermore, this study aims to elucidate the migration patterns of water and salt and reveal the mechanisms of water consumption and salt redistribution in the district, providing theoretical support for water–salt regulation and sustainable ecological development in the HID.

2. Materials and Methods

2.1. Study Area Overview

The Hetao Irrigation District (HID) is an area where agriculture depends entirely on irrigation. The current farmland area is 5.74 × 10⁵ hm², saline–alkali land covers 2.09 × 10⁵ hm², water bodies occupy 1.3 × 10⁴ hm², and sandy land spans 1.469 × 10⁵ hm² [11]. The region’s key landscape features are farmland, wasteland, sand dunes, and lakes, which are integral to the desert oasis ecosystem of the Hetao Plain in Inner Mongolia. The study area is located in the Hetao Irrigation District’s Jiefangzha irrigation area, specifically near Zhangliansheng Lake (40°54′36.24″ N, 107°15′59.07″ E; elevation: 1035 m) (Figure 1a).

The experimental site covers about 120 acres of farmland and wasteland, with the lake area spanning 768 acres. The farmland, wasteland, and lake are adjacent to one another (Figure 1b). The site’s long-term average temperature is 7.5 °C, with a frost-free period of 130 to 150 days. The annual rainfall ranges from 90 to 144.2 mm, with precipitation concentrated between June and August, accounting for 70% of the total. The mean annual evaporation is 2237 mm. RTK measurements in the study area show that the maximum elevation difference in the farmland is 15 cm, with the farmland being on average 45 cm higher than the wasteland. The main crops in the region are sunflowers and corn. Sunflowers are planted in early June and harvested in early October, with a growing season of 128 days. The soil texture in the 0–300 cm layer is primarily sandy loam and sandy soil.

2.2. Experimental Setup and Data Collection

2.2.1. Groundwater and Irrigation Data

Seventeen groundwater monitoring wells were installed in the study area, including seven key wells and ten general wells. The key wells were equipped with automatic groundwater sensors (CTD-10, METER Group, Pullman, WA, USA), and the groundwater levels and salinity were recorded every hour using an EM50 data logger. The general wells were manually measured every seven days for groundwater depth, and groundwater samples were collected every ten days for salinity analysis. Three key observation points were selected: Point A (farmland), Point B (wasteland), and Point C (lake boundary) (Figure 2).

Groundwater Table

Groundwater levels at these points fluctuate similarly, as shown in Figure 3. During the crop growing season in 2017 and 2018, the groundwater depth in the study area varied between 100 and 220 cm in the farmland; 120 and 190 cm in the wasteland; and 50 and 150 cm at the lake boundary. In 2022 and 2023, the groundwater depths ranged from 140 to 260 cm in the farmland; 140 to 247 cm in the wasteland; and 100 to 180 cm at the lake boundary.

Irrigation Volume

During the 2017 growing season, irrigation events occurred on 22 May (206 mm), 28 June (135 mm), 20 July (112 mm), and 29 August (90 mm). In 2018, irrigation was applied on 23 May (186 mm), 19 June (104 mm), and 2 July (96 mm). For 2022 and 2023, single irrigation events were recorded: 27 May 2022 (253 mm) and 23 May 2023 (247 mm) (Figure 3). This reduction in irrigation frequency and volume during crop growing seasons resulted from water conservation initiatives deeply implemented in the Hetao Irrigation District starting in 2022, thereby limiting irrigation to one annual event in these two years.

Groundwater Electrical Conductivity

The average groundwater TDS value in 2017 and 2018 was 655 mg/L and 670 mg/L for the farmland, 972 mg/L and 1365 mg/L for the wasteland, 1660v mg/L and 1835 mg/L for the lake boundary, and 2619 mg/L and 2926 for the lake. For 2022 and 2023, the average groundwater TDS value was 1775 mg/L and 1899 mg/L for the farmland, 1633 mg/L and 1827 mg/L for the wasteland, 1663 mg/L and 1835 mg/L for the lake boundary, and 2346 mg/L and 2202 mg/L for the lake (

Table 1. Mineralization of different water samples.

Time	Irrigation Water (mg/L)	Farmland Groundwater (mg/L)	Wasteland Groundwater (mg/L)	Groundwater of Lake Boundary (mg/L)	Lake (mg/L)
2017	548	655	972	1259	2619
2018	580	670	1365	1403	2926
2022	600	1775	1663	1663	2346
2023	600	1899	1827	1835	2202

).

2.2.2. Soil Monitoring

A total of 63 general soil observation points were established in the study area, spaced 50 m apart. Soil samples were collected at a depth of 100 cm, with samples taken every 20 cm for five layers. Soil sampling was conducted every 10 days, with additional samples taken before and after irrigation. Soil moisture was measured using the oven-drying method, and soil salinity was determined using a conductivity meter (DDS-307A, Youke Instrument Co., Shanghai, China) to analyze the 1:5 soil–water extract. Additionally, seven key soil monitoring points were established, with soil samples collected at a depth of 200 cm, with samples taken every 20 cm per layer. At these key points, automatic soil sensors (5TE, METER Group, Pullman, WA, USA) were installed to monitor the soil moisture, salinity, and temperature at 1 h intervals, with data recorded using EM50 loggers (Figure 2).

2.2.3. Water Volume Measurement and Sample Collection

Water volumes for irrigation were measured using trapezoidal weirs, and evaporation was measured using 20 cm diameter evaporation pans. Rainwater, irrigation water, and lake water samples were collected every ten days, with three replicate samples taken each time. The electrical conductivity and salinity of the water samples were measured using a conductivity meter (DDS-307A). The relationship between the salinity and EC for water samples was calculated as follows: TDS (g/L) = 0.69 ECw (mS/cm).

2.2.4. Soil Physical Properties in the Study Area

Soil bulk density and saturated hydraulic conductivity were measured at key observation points (Figure 2) in the 0–300 cm soil layer using the ring knife method. Soil particle size distribution was determined using a laser particle size analyzer (HELOS & RODOS, Sympatec GmbH, Clausthal-Zellerfeld, Germany). According to the results, the soil at Point A (farmland) was relatively uniform, with sandy loam throughout the 0–300 cm layer. The soil at Point B (wasteland) could be divided into two layers: sandy loam in the 0–80 cm layer and sandy soil in the 80–300 cm layer. The soil at Point C (lake boundary) could also be divided into two layers: sandy loam in the 0–20 cm layer and sandy soil in the 20–300 cm layer. Based on the soil texture and water retention properties, VG parameters were determined using the neural network calculation submodule in the Hydrus model [35] (

Table 2. Soil physical properties of typical samples in the study area.

Sample Point	Soil Layer (cm)	Soil Physical Properties					VG Parameter
		Clay	Powder	Sand	Dry Density (g/cm3)	Saturated Hydraulic Conductivity (cm/d)	${\mathit{\theta }}_{\mathit{r}}$	${\mathit{\theta }}_{\mathit{s}}$	$\mathsf{\alpha }$ (-)	$\mathbf{n}$ (-)
		<0.02 mm	0.02–0.5 mm	>0.5–2 mm
		%	%	%
Farmland (A)	0–300	3.16	45.28	51.56	1.66	19.67	0.0266	0.3091	0.0341	1.1228
Wasteland (B)	0–80	2.18	42.65	55.17	1.68	22.15	0.0258	0.3597	0.0409	1.1228
	80–300	2.49	7.68	89.83	1.69	230.82	0.0444	0.3296	0.0382	2.0541
Lake boundary (C)	0–20	5.612	14.324	80.064	1.52	197.72	0.0423	0.3798	0.0403	1.7865
	20–300	0.42	5.88	94.12	1.73	315.15	0.0347	0.31	0.036	3.1947

).

2.2.5. Meteorological Data

Daily meteorological data, including air temperature, relative humidity, sunshine duration, wind speed, and precipitation, were collected. Using these data and the Penman–Monteith formula, the reference crop evapotranspiration (ET₀) was estimated. The variation in ET₀ and precipitation for 2017, 2018, 2022, and 2023 is shown in Figure 4. From 2017 to 2023, precipitation ranged from 70 to 145 mm, with growing season (1 May to 1 October) rainfall of between 60 mm and 137 mm. The average daily evapotranspiration was 2.9 mm/day, and during the growing season, it ranged from 4.4 mm/day to 4.7 mm/day.

2.3. Water Balance Model for Farmland, Wasteland, and Lake Systems

In the research area, water balance equations for the farmland, wasteland, and lake were established (Figure 5). The atmospheric boundary was defined as the upper boundary condition, while the groundwater table was designated as the lower boundary condition.

2.3.1. Water Balance Calculation Model for Farmland During the Crop Growing Season

The water balance calculation model for the farmland included parameters such as the irrigation volume, precipitation, actual evapotranspiration, canal seepage, groundwater exchange, and changes in the soil unsaturated zone and groundwater storage. This method applies to irrigated areas with uniform terrain and synchronized irrigation processes. It assumes that lateral groundwater inflows and outflows are equal. The water balance calculation for the farmland is as follows:

$$E{T}_{cf}=R+I+Q_{fcp}-D_f-\Delta W_f$$

(1)

$$Q_{fcp}=m\cdot (1-\eta )$$

(2)

where ET_cf is the actual evapotranspiration of the farmland (mm), R is the precipitation (mm), I is the irrigation volume (mm), D_f is the drainage (mm), and ΔW_f is the change in the unsaturated zone and groundwater storage during the growing season (negative values indicate a decrease in storage) (mm). R and I were observed using an automatic weather station and trapezoidal weirs, respectively. If there is no drainage system in the region, D_f can be neglected. Q_fcp is the canal seepage (mm), m is the irrigation quota (mm), and η is the irrigation efficiency, taken as 0.8.

2.3.2. Water Balance Calculation Model for Wasteland During the Crop Growing Season

Since the wasteland is not irrigated and is far from the canal system, the water balance model for the wasteland does not consider irrigation or canal seepage. The wasteland water balance calculation is as follows:

$$E{T}_{wc}=R+M_{wsg}-\Delta W_w$$

(3)

where ET_ws is the actual evapotranspiration of the wasteland (mm), R is the precipitation (mm), M_wsg is the increase in groundwater after each irrigation event (mm), and ΔW_w is the change in the unsaturated zone and groundwater storage during the growing season (negative values indicate a decrease in storage) (mm).

2.3.3. Soil Water Balance Model for the Lake Boundary During the Crop Growing Season

The formula is as follows:

$$E{T}_{lbc}=R+M_{lb}-\Delta W_{lb}$$

(4)

where ET_lbc represents the actual evapotranspiration at the lake boundary (mm), R is the rainfall (mm), M_lb is the increase in groundwater at the lake boundary after each irrigation event (mm), and ΔW_lb is the change in soil moisture storage within the unsaturated zone and groundwater storage at the lake boundary during the growing season (mm), with negative values indicating a reduction in water storage.

2.3.4. Water Balance Model for the Lake During the Crop Growing Season

The parameters for calculating the lake’s water balance included precipitation, groundwater recharge, and evaporation, which led to the following lake water balance equation:

$$\Delta W_l=R+M_{lgw}-\sigma E_{\varphi 20}$$

(5)

where ΔWl represents the change in the lake water volume (mm), R is the rainfall (mm), M_lgw is the groundwater recharge to the lake (mm), and E is the evaporation measured using a Φ20 evaporation pan (mm). The correction factor σ should be applied to the evaporation pan data, depending on the regional evaporation conditions.

The calculation for water storage change (ΔW) in the unsaturated zone and groundwater storage during a given period is as follows:

$$\Delta W=({\theta }_{fg,s}-{\theta }_{ig})\cdot \Delta H+\Delta \theta \cdot L\approx S_y\cdot \Delta H$$

(6)

where θ_fg and θ_ig are the soil water content at saturation and the actual water content at the end of the period (cm³/cm³), respectively, S_y is the specific yield, ΔH is the change in the groundwater level (mm), Δθ is the change in the soil water content in the unsaturated zone, and L is the thickness of the unsaturated zone (mm).

Sophocleous et al. pointed out that in areas with shallow groundwater depths, the unsaturated zone moisture content changes are closely linked to groundwater fluctuations. Consequently, the calculation of ΔS need not separately account for variations in the unsaturated zone moisture content. Instead, a unified specific yield concept can be applied: the specific yield (S_y) is used to represent the volume of water released from the unsaturated zone to the ground surface per unit area of soil when the groundwater level declines by one unit.

The specific yield (S_y) was calculated using the following formula:

$$S_y=S_{\mathrm{yu}}-{\displaystyle \frac{S_{yu}}{{\left[1+{\left(\alpha \left(\frac{Z_i+Z_f}{2}\right)\right)}^n\right]}^{1-\frac{1}{n}}}},S_{yu}={\mathsf{\theta }}_{\mathrm{s}}-{\mathsf{\theta }}_r$$

(7)

where θ_s is the soil water content at saturation (cm³/cm³), θ_r is the residual soil water content (cm³/cm³), Z_i is the initial groundwater depth (cm), and Z_f is the final groundwater depth (cm). Parameters α and n of the van Genuchten model, which are used to describe soil characteristics, can be determined either by pressure membrane measurements or by predicting soil characteristics based on sand, silt, and clay content and bulk density using the Rosetta pedotransfer function.

2.4. Groundwater Flow Model for Farmland, Wasteland, and Lake Systems

Groundwater flow models for the farmland, wasteland, and lake were established based on the principles of groundwater recharge and discharge. The atmospheric boundary was defined as the upper boundary condition, while the groundwater table was designated as the lower boundary condition. In this system, the lake receives groundwater recharge from the wasteland, the wasteland is recharged by groundwater from the farmland, the farmland’s groundwater is replenished by irrigation water and canal seepage.

In arid and semi-arid regions, an increase in lake water volume occurs only during irrigation and rainfall periods. (Since this section focuses on groundwater migration, rainfall contribution is not considered.) During irrigation, we considered the groundwater increment at the lake boundary as the lake water level increment.

2.4.1. Groundwater Migration Model for Farmland

The formula was as follows:

$$Q_{fip}=W_f+G{W}_{f\to w}$$

(8)

$$W_f=S_{fy}\Delta H_f$$

(9)

where Q_fip is the amount of irrigation and canal seepage replenishing the farmland’s groundwater (mm), GW_f→w is the groundwater volume moving from the farmland to the wasteland (mm), W_f is the increment in farmland groundwater (mm), ΔH_f is the rise in farmland groundwater after each irrigation event (mm), and S_fy is the specific yield of the farmland soil.

2.4.2. Groundwater Migration Model for Wasteland

The formula was as follows:

$$G{W}_{f\to w}=G{W}_{w\to l}+W_w$$

(10)

$$W_w=S_{y\mathrm{w}}\Delta H_w$$

(11)

where GW_w→l is the groundwater volume moving from the wasteland to the lake (mm), W_w is the increment in wasteland groundwater (mm), ΔH_w is the rise in wasteland groundwater after each irrigation event (mm), and S_yw is the specific yield of the wasteland soil.

2.4.3. Groundwater Migration Model for Lake Boundary

The formula was as follows:

$$G{W}_{w\to l}=W_l$$

(12)

$$S_l=S_{yl}\Delta H_l$$

(13)

where W_l is the increment in water at the lake boundary (mm), S_yl is the specific yield of the soil at the lake boundary, and ΔH_l is the rise in the groundwater level at the lake boundary after each irrigation event (mm).

2.5. Groundwater Migration Rate Between Farmland, Wasteland, and Lake Systems

2.5.1. Groundwater Migration Rate for Farmland

$${\lambda }_f={\displaystyle \frac{G{W_f}_{\to w}}{Q_{fip}}}$$

(14)

where λ_f is the groundwater migration rate for the farmland, GW_f→w is the groundwater volume moving from the farmland to the wasteland (mm), and Q_fip is the amount of irrigation and canal seepage replenishing the farmland’s groundwater (mm).

2.5.2. Groundwater Migration Rate for Wasteland

$${\lambda }_w={\displaystyle \frac{G{W_w}_{\to l}}{G{W_f}_{\to w}}}$$

(15)

where λ_w is the groundwater migration rate for the farmland, GW_w→l is the groundwater volume moving from the wasteland to the lake (mm), and GW_f→w is the groundwater volume moving from the farmland to the wasteland (mm).

2.6. Solute Transport Calculation for Farmland, Wasteland, and Lake Systems

The adsorption effects of varying soil textures on salts are not considered in groundwater salt transport model development.

2.6.1. Solute Transport Model

Solutes enter the farmland through irrigation water or rainfall. A portion of these solutes is stored within the farmland, while another portion migrates with groundwater to the wasteland. The solutes in the wasteland groundwater mix with those already present. Similarly, a portion of the solutes in the wasteland is stored, while another portion migrates to the lake with the groundwater. This solute transport model is based on the balance theory of groundwater recharge and discharge. The following formulae describe solute transport between these systems, with the units of parameters determined by the solutes being studied:

(1) Solute Change in Farmland

$$\Delta D_{fc}=D_I-D_{f\to wgw}$$

(16)

where ΔD_fc is the solute change in the farmland, D_I is the solute introduced into the farmland, and D_f→wgw is the solute discharge from the farmland to the wasteland.

(2) Total Solute Introduced into Farmland

$$D_I=C_I\cdot V_I$$

(17)

where V_I is the irrigation water volume introduced into the farmland, and C_I is the solute concentration in the irrigation water.

(3) Solute Increase in Farmland Groundwater D_fgw

$$D_{fgw}=\sum _{k=1}^{n}G{W}_{fk}{C}_{fk}$$

(18)

where GW_fk is the groundwater replenishment in the farmland during the k-th irrigation event (mm), and C_fk is the solute concentration in the farmland groundwater during the k-th irrigation event.

(4) Solute Migration from Farmland to Wasteland Groundwater D_f→wgw

$$D_{f\to wgw}=\sum _{l=1}^{n}G{W}_{f\to wl}{C}_{fl}$$

(19)

where GW_f→w is the groundwater volume moving from the farmland to the wasteland during the l-th irrigation event (mm), and C_fl is the solute concentration in the farmland groundwater during the l-th irrigation event.

(5) Solute Migration from Wasteland Groundwater to Lake D_w→lgw

$$D_{w\to lgw}={\sum }_{m=1}^{n}G{W}_{w\to lm}{C}_{wm}$$

(20)

where GW_w→lm is the groundwater volume moving from the wasteland to the lake during the m-th irrigation event (mm), and C_wm is the solute concentration in the wasteland groundwater during the m-th irrigation event.

For all events k = l = m ∈ (1, 2, 3, 4…N), k, l, and m represent irrigation event numbers.

2.6.2. Groundwater Solute Migration Rates

The solute migration rate for farmland groundwater (μ_f) is the ratio of the solute that moves from farmland to wasteland groundwater to the total solute introduced into farmland groundwater via irrigation and canal seepage:

$${\mu }_f={\displaystyle \frac{D_{f\to wgw}}{D_{fgw}}}$$

(21)

The solute migration rate for wasteland groundwater (μ_w) is the ratio of the solute that moves from wasteland groundwater to the lake to the total solute that migrates from farmland groundwater to wasteland groundwater:

$${\mu }_w={\displaystyle \frac{D_{w\to lgw}}{D_{f\to wgw}}}$$

(22)

2.7. Study Flowchart

The flowchart below (Figure 6) describes the study logic processes, main research content, and key issue addressed.

3. Results

3.1. Direction of Groundwater Migration During Irrigation and Non-Irrigation Periods

Based on the groundwater level contour map generated using Surfer 2.0 software (Grid Vector Map), the groundwater level changes are illustrated in Figure 7. The direction of groundwater movement was determined from the variations in groundwater levels. The years 2017, 2018, 2022, and 2023 were considered normal water years. Taking 2018 as an example, the irrigation dates in the study area were 24 May, 21 June, and 4 July, with relatively short intervals between the second and third irrigation events. To effectively capture the changes in groundwater levels, detailed contour maps were created around these three dates. After 15 July, the groundwater level changes were analyzed monthly.

During the irrigation period, the farmland groundwater receives additional water from irrigation, leading to an increase in the groundwater levels. This creates a hydraulic gradient between the irrigated farmland and the non-irrigated wasteland and lake, prompting groundwater movement. For instance, as shown in Figure 7d, the maximum groundwater level was 1033.4 m, while the minimum was 1032.5 m, resulting in a water level difference of 0.9 m and a hydraulic gradient of 0.0045 (

Table 3. Groundwater migration parameters in the irrigation and non-irrigation periods.

Period	Maximum (m)	Mini Mum (m)	Water Level Difference (m)	Distance (m)	Hydraulic Gradient	Groundwater Migration Direction
Irrigation period	1033.4	1032.5	0.9	200	0.0045	Farmland towards the lake
Non-irrigation period	1032.7	1032.16	0.54	200	0.0027	Farmland towards the lake

). In contrast, during the non-irrigation period, the groundwater levels decrease due to crop uptake and soil evaporation. The groundwater table in the irrigated farmland is deeper than that in the wasteland (as illustrated in Figure 3), indicating lower groundwater consumption in the farmland compared to in the wasteland and lake, which experience greater moisture loss [36]. This creates a hydraulic gradient between the farmland, wasteland, and lake. For example, in Figure 7f, the maximum groundwater elevation in the experimental area was 1032.7 m, with a minimum of 1032.16 m, resulting in a water level difference of 0.54 m and a hydraulic gradient of 0.0027 (Table 3). Notably, the hydraulic gradient during the irrigation period was twice that of the non-irrigation period. Therefore, throughout the entire crop growing season, the direction of groundwater movement in the study area was from the farmland toward the wasteland and subsequently toward the lake.

3.2. Groundwater Migration and Water Balance Calculation in Farmland–Wasteland–Lake System

3.2.1. Specific Yield (Sy) and Soil Unsaturated Zone and Groundwater Storage Changes (ΔS)

The specific yield (Sy) was determined using Equation (7), while the changes in the unsaturated soil zone and groundwater storage were calculated using Equation (6). The parameters and results are shown in

Table 4. Calculation of S_y and ∆S at Point A (farmland), Point B (wasteland), and Point C (lake boundary) during the crop growth period in 2017.

Date	Groundwater Level (cm)			Groundwater Level Difference (cm)			A		B		C
	A	B	C	A	B	C	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$
05/01	153.30	137.30	77.50
05/22	173.20	144.70	75.10	−19.90	−7.40	2.40	0.06	−1.14	0.23	−1.73	0.25	0.59
05/28	88.40	103.20	49.10	84.80	41.50	26.00	0.05	4.39	0.23	9.44	0.23	5.99
06/26	183.60	163.50	120.80	−95.20	−60.30	−71.70	0.05	−5.01	0.23	−13.88	0.25	−18.07
07/05	127.70	133.10	101.50	55.90	30.40	19.30	0.06	3.13	0.22	6.78	0.26	5.06
07/20	182.60	159.50	127.40	−54.90	−26.40	−25.90	0.06	−3.07	0.22	−5.81	0.26	−6.81
07/24	146.20	140.10	112.60	36.40	19.40	14.80	0.06	2.08	0.23	4.52	0.26	3.91
08/29	206.20	184.10	139.40	−60.00	−44.00	−26.80	0.06	−3.53	0.24	−10.54	0.27	−7.11
09/02	177.60	171.20	128.00	28.60	12.90	11.40	0.06	1.74	0.24	3.16	0.27	3.04
09/30	212.50	193.50	151.70	−34.90	−22.30	−23.70	0.06	−2.14	0.25	−5.48	0.27	−6.34
Total				−59.20	−56.20	−74.20		−3.55		−13.53		−19.74

,

Table 5. Calculation of S_y and ∆S at Point A (farmland), Point B (wasteland), and Point C lake boundary) during the crop growth period in 2018.

Date	Groundwater Level (cm)			Groundwater Level Difference (cm)			A		B		C
	A	B	C	A	B	C	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$
05/01	140.30	124.20	64.50
05/23	166.10	135.80	65.30	−25.80	−11.60	−0.80	0.06	−1.43	0.23	−2.71	0.23	−0.19
05/28	92.00	97.20	34.80	74.10	38.60	30.50	0.05	3.81	0.23	8.78	0.21	6.30
06/20	165.50	147.10	91.70	−73.50	−49.90	−56.90	0.05	−3.78	0.23	−11.48	0.23	−13.21
06/25	118.40	129.60	69.70	47.10	17.50	22.00	0.05	2.53	0.22	3.90	0.25	5.49
07/02	154.80	136.30	83.10	−36.40	−6.70	−13.40	0.05	−1.92	0.22	−1.47	0.25	−3.30
07/05	120.10	120.50	73.10	34.70	15.80	10.00	0.05	1.84	0.23	3.68	0.25	2.47
08/30	196.20	172.70	126.20	−76.10	−52.20	−53.10	0.06	−4.28	0.24	−12.50	0.26	−13.74
09/02	179.20	156.80	112.70	17.00	15.90	13.50	0.06	1.03	0.24	3.89	0.26	3.57
09/30	199.50	181.70	138.70	−20.30	−24.90	−26.00	0.06	−1.23	0.25	−6.12	0.27	−6.90
Total				−59.20	−57.50	−74.20		−3.44		−14.03		−19.51

,

Table 6. Calculation of S_y and ∆S at Point A (farmland), Point B (wasteland), and Point C (lake boundary) during the crop growth period in 2022.

Date	Groundwater Level (cm)			Groundwater Level Difference (cm)			A		B		C
	A	B	C	A	B	C	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$
05/01	192.00	158.00	119.00				Sy	S	Sy	S	Sy	S
05/26	205.00	172.00	130.00	−13.00	−14.00	−11.00	0.06	−0.74	0.23	−3.27	0.25	−2.71
06/03	135.00	135.00	104.00	70.00	37.00	26.00	0.05	3.62	0.23	8.41	0.23	5.99
08/01	206.00	197.00	144.00	−71.00	−62.00	−40.00	0.05	−3.74	0.23	−14.27	0.25	−10.08
08/04	191.00	183.00	133.00	15.00	14.00	11.00	0.06	0.84	0.22	3.12	0.26	2.89
08/14	206.00	197.00	145.00	−15.00	−14.00	−12.00	0.06	−0.84	0.22	−3.08	0.26	−3.16
08/17	198.00	190.00	142.00	8.00	7.00	3.00	0.06	0.46	0.23	1.63	0.26	0.79
09/30	223.00	206.00	164.00	−25.00	−16.00	−22.00	0.06	−1.47	0.24	−3.83	0.27	−5.84
Total				−31.00	−48.00	−45.00		−1.87		−11.28		−12.12

and

Table 7. Calculation of S_y and ∆S at Point A (farmland), Point B (wasteland), and Point C (lake boundary) during the crop growth period in 2023.

Date	Groundwater Level (cm)			Groundwater Level Difference (cm)			A		B		C
	A	B	C	A	B	C	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$	${\mathit{S}}_{\mathit{y}}$	$∆\mathit{S}$
05/01	221.00	199.00	134.00
05/22	231.00	210.00	144.00	−10.00	−11.00	−10.00	0.06	−0.57	0.23	−2.57	0.25	−2.46
05/29	165.00	177.00	132.00	66.00	33.00	12.00	0.05	3.42	0.24	7.92	0.25	3.00
09/30	258.00	242.00	178.00	−93.00	−65.00	−46.00	0.06	−5.58	0.25	−16.25	0.27	−12.42
Total				−37.00	−43.00	−44.00		−2.74		−10.90		−11.88

. The Sy values ranged from 0.05 to 0.06 for the farmland, from 0.22 to 0.25 for the wasteland, and from 0.25 to 0.27 for the lake boundary. Zhang et al. [37] found that the Sy values for silt loam soils ranged between 0.04 and 0.06, while Cai et al. [38] reported Sy values for sandy soils of 0.15, 0.26, 0.263, and 0.274, confirming that the calculated Sy values in this study aligned with actual measurements and provided a solid foundation for constructing a high-precision water–salt balance model.

During the entire growing season in 2017 and 2018, the soil and groundwater storage in the farmland, in the wasteland, and at the lake boundary decreased by 34.4–35.5 mm, 135.3–140.3 mm, and 195.1–197.4 mm, respectively. For 2022 and 2023, the decrease was 18.7–27.4 mm for the farmland, 109.0–112.8 mm for the wasteland, and 118.8–121.2 mm for the lake boundary. The reduction in the farmland was smaller than that in the wasteland and lake boundary, as irrigation replenished the soil water and groundwater in the farmland, while no irrigation occurred in the wasteland and at the lake boundary, leading to higher water consumption in those areas due to shallower groundwater depths and increased evaporation and lateral flow. Throughout the crop growth period, the farmland, wasteland, and lake boundary experienced net water losses. In 2022 and 2023, the amount of water diverted to the Hetao Irrigation District was reduced, irrigation quotas were lowered, autumn irrigation was canceled in some areas, and cropping patterns were adjusted. As a result, the groundwater levels in the study area decreased compared to in 2017–2018, with groundwater levels dropping by 40 cm in the farmland, 40 cm in the wasteland, and 50 cm at the lake boundary. This caused the water storage in the farmland, in the wasteland, and at the lake boundary to decrease by an average of 12 mm, 27 mm, and 77 mm, respectively, in 2022–2023 compared to 2017–2018.

3.2.2. Groundwater Migration Volumes

During the irrigation period, vertical recharge to the farmland’s groundwater far exceeded its consumption, resulting in higher groundwater levels in the farmland compared to in the wasteland and lake, causing groundwater to migrate horizontally from the farmland to the wasteland and the lake. The calculation results (

Table 8. Calculation of the groundwater migration of the farmland–wasteland–lake system during the crop growth period in 2017, 2018, 2022, and 2023.

Years	Flooding Incident	Calculation Period	Groundwater Increment (cm)			The Amount of Groundwater Replenishing Wasteland from Farmland	Total Amount of Groundwater Supplied by Leakage from Fields and Canal Systems to Farmland	Translocation Rate
			Farmland	Wasteland	Lake Boundary			Groundwater Migration Rate in Farmland (%)	Groundwater Migration Rate in Wasteland (%)
2017	1	05/22–05/28	4.4	9.48	5.99	15.47	19.87	78	39
	2	06/29–07/03	3.13	6.81	5.06	11.87	15	79	43
	3	07/21–07/24	2.1	4.56	3.91	8.47	10.57	80	46
	4	08/30–09/02	1.72	3.13	3.07	6.2	7.92	78	50
	Average value							78.75	44.3
2018	1	05/22–05/28	3.82	8.8	6.28	15.8	19.62	81	40
	2	06/20–06/25	2.52	3.88	5.51	9.39	11.91	79	59
	3	07/01–07/05	1.85	3.7	2.45	6.15	8	77	40
	Average value							79	46.3
2022	1	05/27–06/03	3.62	8.41	5.99	14.4	18.02	80	41
2023	1	05/23–05/29	3.42	7.92	3.00	10.92	14.34	76	27

) showed that in 2017 and 2018, the average groundwater migration rate from the farmland to the wasteland was 78.75% and 79%, respectively, while the average migration rate from the wasteland to the lake was 44.3% and 46.3%. In 2020, Wang et al. [11] used a binary element model to calculate the different types of water conversion relationships in a farmland–wasteland–sea subsystem before and after irrigation (26 May~31 May, 21 June~28 June, and 4 July~9 July); the study found that the contribution rates of irrigation water and rainfall to the groundwater level of farmland were 94% and 6% respectively; the average contribution rates of farmland groundwater and rainfall to wasteland groundwater were 71% and 29%, respectively; the average contribution rates of wasteland groundwater and rainfall to lake were 43% and 57%, respectively. These results are similar to the groundwater migration model constructed in this study, indicating that the model construction is reasonable and the calculation results have high credibility.

In 2022 and 2023, the average groundwater migration rate from the farmland to the wasteland was 80% and 76%, respectively, while from the wasteland to the lake, it was 41% and 27%. The migration rates over the two years were relatively stable. However, the migration rate from the wasteland to the lake in 2023 was 14% lower than in 2022, primarily due to a 40 cm drop in the groundwater levels. Given the larger area and irrigation volume in the farmland, more water infiltrated into the farmland’s groundwater, leading to a 30–50% higher migration rate from the farmland to the wasteland compared to that from the wasteland to the lake.

3.2.3. Evapotranspiration in Farmland, in Wasteland, and at Lake Boundary

Using the water balance Equations (1)–(15), the evapotranspiration (ET) from May to September in 2017, 2018, 2022, and 2023 for the farmland (A), wasteland (B), and lake boundary (C) in the study area was calculated, as shown in

Table 9. Calculation of the evapotranspiration of the farmland–wasteland–lake system during the crop growth period in 2017 and 2018.

Years	2018				2017
Sample	Rainfall (mm)	Irrigation + Canal Leakage Increment/Water Storage Increment (mm)	Change (mm)	Evapotranspiration (mm)	Rainfall (mm)	Irrigation + Canal Leakage Increment/Water Storage Increment (mm)	Change (mm)	Evapotranspiration (mm)
A	113.4	422.8	−34.4	570.6	53.4	597.3	−35.5	686.2
B	113.4	163.8	−140.3	417.5	53.4	239.8	−135.3	428.5
C	113.4	142.4	−195.1	450.9	53.4	180.3	−197.4	431.1

and

Table 10. Calculation of the evapotranspiration of the farmland–wasteland–lake system during the crop growth period in 2022 and 2023.

Years	2023				2022
Sample	Rainfall (mm)	Irrigation + Canal Leakage Increment/Water Storage Increment (mm)	Change (mm)	Evapotranspiration (mm)	Rainfall (mm)	Irrigation + Canal Leakage Increment/Water Storage Increment (mm)	Change (mm)	Evapotranspiration (mm)
A	85.4	296.4	−27	408.8	137	303.6	−19	459.6
B	85.4	79.2	−109	273.6	137	131.6	−113	381.6
C	85.4	30	−119	234.4	137	96.7	−121	354.7

. In 2017–2018, the farmland ET ranged from 570 to 686 mm, the wasteland ET from 417 to 428 mm, and the lake boundary ET from 431 to 451 mm. The average ET in the farmland was 32% higher than that in the wasteland and 29% higher than that at the lake boundary. In 2022–2023, the farmland ET ranged from 408 to 459 mm, with 2022 having a 50 mm higher ET than 2023 due to 50 mm more rainfall in 2022. The wasteland ET ranged from 273 to 381 mm, while the lake boundary ET ranged from 234 to 354 mm. The wasteland ET was approximately 30–40 mm higher than the lake boundary ET, likely due to the sandy loam soil in the wasteland compared to the sandy soil at the lake boundary, with sandy soils having lower water retention and capillary rise, leading to reduced soil evaporation.

In 2022 and 2023, there was only 1 irrigation event, compared to 2–3 irrigation events in 2017–2018. The water increment in the farmland was 130–290 mm less than in 2017–2018, while the water storage increment in the wasteland was 80–100 mm less, and at the lake boundary, it was 80–110 mm less. Additionally, the groundwater depths in the farmland, in the wasteland, and at the lake boundary in 2022–2023 were 40 cm, 20–50 cm, and 30–50 cm lower, respectively, than in 2017–2018, reducing the water storage (ΔS) in the wasteland and at the lake boundary by 30 mm and 80 mm, respectively. The evapotranspiration in the farmland, in the wasteland, and at the lake boundary in 2022–2023 was reduced by 170–230 mm, 40–140 mm, and 100–200 mm, respectively, compared to 2017–2018.

In 2022–2023, the lake’s water level changed by 679–789 mm, compared to 631–706 mm in 2017–2018, a difference of 50–80 mm. The reduced groundwater supply from the wasteland to the lake by 80–120 mm (

Table 11. Calculation of the water change in the lake based on water balance theory during the crop growth period in 2017, 2018, 2022, and 2023.

Years	Rainfall (mm)	Supply (mm)	Evaporation (mm)	$\mathbf{Change}{∆\mathit{W}}_{\mathit{w}}$ (mm)
2017	53.4	180.3	940	−706.3
2018	113.4	142.4	887	−631.2
2022	137	60	876	679
2023	85.4	30	905	789.6

) contributed to this variation. Without further water replenishment, the lake faces the risk of drying out.

3.3. Calculation of Groundwater Salt Migration in Farmland–Wasteland–Lake System

3.3.1. Groundwater Salt Migration Rate in Farmland–Wasteland–Lake System

Using Equations (16)–(22), the salt migration rate of groundwater from the farmland to the wasteland during the 2017 crop growth period was calculated to be 79%, while the salt migration rate from the wasteland to the lake was 110%. This indicated that all of the salt transferred from the farmland groundwater to the wasteland subsequently migrated to the lake, which also had a desalination effect on the wasteland groundwater. In the 2018 crop growth period, the migration rate of groundwater salt from the farmland to the wasteland was also 79%, while the migration rate from the wasteland to the lake was 95% (

Table 12. Calculation of the groundwater salt migration of the farmland–wasteland–lake system during the crop growth period in 2017, 2018, 2022, and 2023.

			Salt Increment in Groundwater (g/m2)					Translocation Rate
Years	Flooding Incident	Calculation Period	Farmland	Wasteland	Lake Boundary	The Amount of Salt from Farmland Groundwater Recharging Wasteland (g/m2)	The Amount of Salt Supplied by Leakage from Fields and Canal Systems to Groundwater in Farmland (g/m2)	Migration Rate in Farmland Groundwater (%)	Migration Rate in Wasteland Groundwater (%)
2017	1	05/22–05/28	34.49	117.28	74.11	121.26	155.75	78	61
	2	06/29–07/03	23.07	126.87	94.27	87.47	110.54	79	108
	3	07/21–07/24	18.07	121.77	104.41	72.88	90.95	80	143
	4	08/30–09/02	16.67	77.97	76.47	60.11	76.78	78	127
	Average value							79	110
2018	1	05/22–05/28	35.45	136.74	97.58	146.63	182.08	81	67
	2	06/20–06/25	24.71	90.73	128.85	92.07	116.78	79	140
	3	07/01–07/05	19.75	78.45	51.95	65.65	85.39	77	79
	Average value							79	95
2022	1	05/27–06/03	55.39	126.15	91.05	220.32	311.75	71	41
2023	1	05/23–05/29	63.95	133.85	52.50	204.20	268.16	76	26

). The lower rate in 2018 compared to 2017 was due to a reduction in the irrigation volume by 157 mm, which resulted in less groundwater recharge from the wasteland to the lake, and some of the salt transferred from the farmland to the wasteland groundwater was not fully discharged into the lake.

During the 2022 and 2023 crop growth periods, the salt migration rates from the farmland to the wasteland were 71% and 76%, respectively, while the rates from the wasteland to the lake were 41% and 26%. The farmland groundwater salt migration rates were similar to the average rates for 2017 and 2018, likely due to similar irrigation volumes of approximately 200–240 mm for the first irrigation and the consistent soil texture of the farmland, which maintained stable saturated hydraulic conductivity over the years. However, the salt migration rate from the wasteland groundwater to the lake in 2022 and 2023 was significantly lower than in 2017 and 2018. This was mainly due to the reduced number of irrigation events, the cancellation of autumn irrigation, and a 40 cm decrease in the groundwater depth in wasteland compared to 2017–2018. Furthermore, the salt concentration of the farmland groundwater in 2022 and 2023 was three times higher than in previous years, resulting in 80–100 g/m² more salt being transferred from the farmland to the wasteland groundwater than in 2017–2018. Since the groundwater recharge to the lake and its salt concentration increased less than in 2017–2018, the salt migration rate from the wasteland to the lake decreased.

3.3.2. Groundwater Salt Migration Volume in Farmland–Wasteland–Lake System

During the crop growth periods of 2017 and 2018, the average salinity of the farmland groundwater increased by 861 kg/hm². The average amount of salt migrated from the farmland to the wasteland groundwater was 3231 kg/hm², while 3139 kg/hm² was transferred from the wasteland groundwater to the lake. On average, 92 kg/hm² of salt accumulated annually in the wasteland groundwater. The salinity in the wasteland groundwater was primarily discharged through natural salt leaching, effectively maintaining the salinity balance in the wasteland groundwater (Figure 8).

In 2022 and 2023, the average salinity in the farmland groundwater increased by 597 kg/hm², while the average amount of salt transferred from the farmland groundwater to the wasteland groundwater was 2122 kg/hm², and the average amount of salt transferred from the wasteland groundwater to the lake was 720 kg/hm². The total amount of salt introduced in 2022–2023 was 1800 kg/hm², 800 kg/hm² less than in 2017–2018. This reduction in salinity was due to the decreased irrigation volume, as only one irrigation event occurred in the study area during 2022–2023. Although less salt was introduced, 1402 kg/hm² of the salt transferred from the farmland groundwater was retained in the wasteland groundwater. Compared to 2017–2018, the amount of salt migrated from the wasteland groundwater to the lake decreased by approximately 2400 kg/hm² in 2022 and 2023 (Figure 8), causing the salt concentration in the wasteland groundwater to increase by 0.5–0.7 g/L compared to 2017–2018, while the salt concentration in the lake decreased by 0.3–0.7 g/L (Table 1).

Reducing the frequency of irrigation, cutting back on irrigation volumes, and canceling autumn irrigation lowered the groundwater levels, but it also led to salt accumulation in the groundwater of the farmland and the wasteland. This accumulation poses risks to the groundwater environment, crop growth, and soil salinization in the irrigation area. Moreover, the lake is facing a severe risk of drying out due to insufficient water replenishment. Without ecological water supplementation, the water environment in the irrigation area will face significant challenges.

4. Discussion

4.1. Water and Salt Migration Patterns in Different Land Types

In recent years, due to the implementation of water-saving projects, the diversion volume in the Hetao Irrigation District has gradually decreased, disrupting the long-established water–salt balance system and leading to a redistribution of salinity within the district. Li et al. [39] pointed out that wastelands are influenced by horizontal soil water infiltration from adjacent irrigated farmlands and vertical groundwater recharge. This study found that during the crop growth period, the groundwater migrated from the farmland to the wasteland and lake, increasing the water storage in the wasteland. Ren et al. [40] reported that during the irrigation period, intense lateral groundwater exchange occurred between irrigated and non-irrigated fields, indicating a close hydraulic connection between the farmland and the wasteland in the irrigation district. The surface soil salinity in the farmland leached to deeper layers, resulting in salt accumulation in the deep soil. Our study further quantified this deep-soil salt accumulation in farmlands, with the saturated zone accumulating between 597 kg/hm² and 861 kg/hm² of salts. Yue et al. [13] developed a water and salt migration and balance model for non-agricultural areas, agricultural areas, and water bodies. They found that 75% of the salt desalinized from agricultural areas was transferred to non-agricultural areas through the groundwater. In our study, we constructed separate water and salt balance models for the farmland–wasteland–lake system, revealing that 79% of the salt in the farmland groundwater migrated to the wasteland, consistent with previous findings. However, our study also discovered that when the frequency of irrigation exceeded three times, no salt accumulation occurred in the wasteland groundwater, and the salt transferred from the farmland groundwater was entirely discharged into the lake. Conversely, with only one irrigation, 1402 kg/hm² of salt remained in the wasteland groundwater. Wang et al. [15] found that dry drainage accounted for 83% of the salt introduced by irrigation between 2007 and 2011, and Wen et al. [41] showed that the ratio of dry drainage to irrigation-introduced salt in JFZ (Jiefangzha irrigation area) was 80%. Our results are similar, with a dry drainage salt removal rate of 80%. However, our study revealed that when irrigation exceeded three times, 110% of the salt from the wasteland groundwater migrated to the lake, which served as a reservoir for salts. In contrast, with only one irrigation, 28–40% of the salt in the wasteland groundwater migrated to the lake, turning the wasteland into a salt storage zone while reducing the lake’s salinity. Wen et al. [41] found that groundwater table control is key for water and salt regulation in the HID. In the irrigation and drainage conditions described, the groundwater level remained relatively stable with a slight decreasing trend in the study period, i.e., 12.9 mm/year in irrigated land and 5.5 mm/year in non-irrigated land [41]. In our study, we observed that the groundwater levels in the study area decreased by an average of 40 cm in 2022 and 2023 compared to 2017–2018. This decline can be attributed to reduced irrigation frequency and volume over the past six years, as well as the cancellation of autumn irrigation in 2022, altering the previously established patterns of water and salt migration.

4.2. Impact of Water and Salt Migration Processes in Different Land Types on Water-Saving and Salt Control in the Hetao Irrigation District

Wang et al. [8] found that the Hetao Irrigation District has a strong capacity for dry drainage, which can alleviate the risk of salinization in farmlands by utilizing parts of abandoned wastelands under the current irrigation and drainage infrastructure. Our study discovered that in the farmland–wasteland–lake system, the salt balance in the wasteland was maintained throughout 2017–2018. However, in 2022–2023, due to the reduction in the irrigation water volume, salts migrated from the farmland groundwater to the wasteland groundwater, but the wasteland groundwater could not fully transfer these salts to the lake. As a result, the wasteland experienced salt accumulation. To mitigate the salinization of farmland, a portion of the wasteland could be used for dry drainage. However, considering the groundwater environment and sustainable development of the irrigation district, a phased increase in the irrigation water volume and frequency is recommended to adjust the groundwater salinity levels. Wu et al. [14] noted that the ratio of dry drainage to total salt introduction is related to land use patterns and can vary with spatial scale. In our study, which focused on the farmland–wasteland–lake land use type, the dry drainage ratio was relatively high, as we considered both the salt discharge from the wasteland groundwater to the lake and the role of the lake in water and salt migration within the irrigation district. Wen et al. [41] also found that dry drainage played a key role in controlling soil salinization in irrigated lands. However, it can also lead to salt accumulation in non-irrigated lands, which may negatively impact the soil environment and natural ecosystems. Our study reached similar conclusions but further revealed that with reducing irrigation water and lowering groundwater levels, the lake faced a water deficit of 679–789 mm during the crop growth period due to the lack of sufficient water replenishment. Without adequate water supply, the lake is at risk of drying up, leading to vegetation degradation and threatening the ecological safety of the irrigation district. In the future, the research will be expanded to the entire Hetao Irrigation District to identify how many areas will exhibit similar land types. A water and salt migration evaluation model suitable for different scales of farmland–wasteland–lake systems will be developed, providing a theoretical basis for uncovering the mechanisms of water and salt redistribution in the Hetao Irrigation District.

This study mainly focused on the migration process of groundwater and salt. Since the aim of this study was to obtain the groundwater migration volume in the farmland, wasteland, and lake system, this study did not specifically consider the impact of soil heterogeneity on groundwater and salt transport. An uncertainty analysis and sensitivity analysis of model construction considering soil heterogeneity is the next research direction.

5. Conclusions

(1) Before and after the reduction in the irrigation water volume and the lowering of the groundwater table, the groundwater from the farmland continued to migrate toward the wasteland and lake. The migration rates of the farmland groundwater to the wasteland and the wasteland groundwater to the lake remained relatively stable, averaging 78% and 40%, respectively.

(2) Following the reduction in the irrigation water volume and the drop in groundwater levels during the crop growth period, the groundwater supply from the wasteland to the lake decreased by 80–120 mm, resulting in a water deficit of 679–789 mm in the lake. Without sufficient water replenishment, the lake will gradually dry up over time. Additionally, the evapotranspiration of the farmland, wasteland, and lake decreased by 170–230 mm, 40–140 mm, and 100–200 mm, respectively, compared to the period before the reduction in irrigation, which effectively reduced ineffective water consumption in the irrigation district.

(3) During the crop growth period, prior to the reduction in the irrigation water volume, the salt supplied annually to the wasteland groundwater was discharged via dry drainage, maintaining a balance in the wasteland groundwater salinity. After the reduction, the migration rate of salt from the farmland groundwater to the wasteland ranged from 71% to 76%, while the migration rate from the wasteland groundwater to the lake dropped to 26–41%. Additionally, 1402 kg/hm² of salt from the farmland groundwater remained in the wasteland groundwater, while 597–861 kg/hm² of salt accumulated in the farmland groundwater. This salt accumulation in both the farmland and wasteland groundwater negatively impacts the groundwater environment and crop growth and exacerbates soil salinization in the irrigation district.