Assessing field-scale salt leaching during saline soil remediation with electrical resistivity tomography and electromagnetic induction methods

During the remediation of saline wasteland, it is crucial to rapidly and accurately characterize the spatial distribution and temporal dynamics of soil salinity for evaluating the efficacy of leaching measures and guiding management strategies. In this study, electrical resistivity tomography (ERT) and electromagnetic induction (EMI) were deployed in a saline wasteland in the Jingtai Yellow River Irrigation District, Gansu Province, China, to monitor field-scale changes in salinity before and after salt leaching. A calibrated model linking soil bulk electrical conductivity and ground temperature to total dissolved solids (TDS) was applied to convert geophysical results into quantitative TDS values, enabling quantitative assessment of field-scale salt removal rates. Monitoring indicated that salinization was mainly caused by a rising groundwater table resulting from inadequate drainage, which led to evaporation-driven salt accumulation in low-lying areas and eventual land abandonment. Although remediation measures—including dredging drainage ditches and conducting salt-leaching irrigation—were implemented, their effectiveness was spatially heterogeneous due to the presence of low-permeability layers such as clay pans or caliche. For instance, along a transect 5 m from a newly dredged ditch, the desalination rate reached 87.8% after one leaching event. In contrast, along a transect 85 m from the ditch, the rate was only 16.8% after dredging and leaching. These findings underscore the need for targeted salt control strategies and optimized winter irrigation protocols to improve regional salinity management.

1 Introduction

The decline in land productivity, loss of arable land, and degradation of vegetation due to increased salinity in water and soil, driven by natural factors or human irrigation activities, have become pressing global challenges. Influenced by unsustainable agricultural practices, approximately half of the world’s cropland is projected to be at risk of salinization by 2050 (Nachshon, 2018; Cui et al., 2025). Against the backdrop of limited arable land and food shortages, preventing salinization of existing cropland and effectively rehabilitating already saline land are imperative for fostering socio-economic development.

Salinized soils predominantly occur in arid and semi-arid environments characterized by low precipitation and high evaporation. Salt originates primarily from the soil itself and parent geological materials, shallow mineralized groundwater, or saline irrigation water (Nachshon, 2018). Typically, groundwater serves as the main geological agent for salt transport, accumulation, and discharge. Salinity tends to increase along groundwater flow paths, from recharge areas towards valley bottoms and groundwater discharge zones of catchment (Salama et al., 1999). The freeze-thaw process of soil (Gao et al., 2023), along with meteorological factors such as wind speed (Guo et al., 2025), also influences the accumulation and migration of salts. Soil leaching via subsurface drainage systems represents a traditional strategy for removing excess salts from saline soils (Zhang et al., 2024). During the leaching process, both the irrigation method and the design of the drainage system are critical factors influencing leaching efficiency. Simulation results indicate that compared to full-area ponding method of leaching, partial ponding may achieve efficient salt leaching under more water-efficient conditions (Siyal et al., 2010).

Mapping and monitoring the spatial distribution and dynamics of soil salinity are crucial for ecological and agricultural management (Liu et al., 2016). They form the basis for understanding the formation mechanisms of saline soils, developing remediation strategies, and evaluating the effectiveness of various measures. Conventional methods involve mixing field-collected soil samples with water at specific ratios (e.g., 1:1, 1:2, 1:5) and determining soil salinity by measuring the electrical conductivity of the extract (Selim et al., 2025). However, this sampling approach is not only time-consuming and labor-intensive but also provides only point-specific information. Particularly over large areas, limited sampling points struggle to capture the spatial heterogeneity of water and salt distribution. Leveraging the strong correlation between soil salt content and electrical conductivity, electrical resistivity tomography (ERT) and electromagnetic induction (EMI) have been successfully applied in numerous studies to investigate salt distribution and dynamics in saline soils (Nguyen et al., 2024; Paz et al., 2024). For instance, Nguyen et al. (2024) have used 3D ERT to obtain three-dimensional electrical conductivity characteristics at depths up to 40 m, elucidating the accumulation of saline water in paddy fields. ERT has also been employed to reveal geo-electrical features under different hydrogeological conditions in an oasis, finding significant spatial variability of salinity both vertically and horizontally due to differences in soil type, texture, and water content (Mohammed et al., 2022).

Using apparent conductivity data from EMI and calibrated equations relating apparent conductivity to the electrical conductivity of saturated soil paste extracts, studies have derived spatiotemporal characteristics of soil salinity in research areas (Jadoon et al., 2015; Paz et al., 2024; Selim et al., 2025). The spatial distribution of soil salinity in the Yellow River Delta has been investigated based on EMI-derived apparent conductivity combined with geostatistical analysis (Liu et al., 2016). Integrating hydrological modeling to simulate water distribution after multiple irrigation events, and converting water content into electrical conductivity characteristics, has provided prior information for the inversion processes. This approach guides the selection of inversion parameters and interpretation of results, thereby enhancing the capability of electromagnetic induction techniques for rapidly surveying large-scale spatiotemporal variations in soil salinity (Farzamian et al., 2021).

In the Jingdian Irrigation District of northwestern China, cropland is primarily irrigated with water pumped from the Yellow River via hydro-engineered projects. Analyses based on Sentinel-1 remote sensing data and field sampling indicate a gradual increasing trend in soil salinity within this region, alongside a slow rise in groundwater mineralization and surface irrigation volume, and a continuous increase in groundwater table (Lian et al., 2022; Dong et al., 2025). Over the long term, this has led to the formation of extensive moderately and severely saline soils, with some areas abandoned due to excessively high salinity levels rendering them uncultivable. Remediation efforts for saline land in this study area primarily focus on gradually lowering the regional groundwater table by improving irrigation and drainage systems, combined with leaching irrigation and soil amendments to enhance soil quality. Unlike studies utilizing geophysical methods for one-time spatial distribution mapping of electrical conductivity, the research site is located in a seasonally frozen soil region where shallow soil temperatures exhibit significant seasonal variations. Therefore, when analyzing soil total dissolved solids (TDS) based on multiple electrical conductivity surveys, it is necessary to correct the data to a consistent temperature condition. Furthermore, the relationship between soil electrical conductivity and TDS is closely tied to site-specific factors such as soil properties, groundwater, and salt characteristics, making it inherently site-specific. Consequently, relationship models established in other study areas cannot be directly applied. To assess the spatial distribution of soil salinity and monitor changes in water and salt content during remediation, surveys using both ERT and EMI were conducted in the study area—once prior to implementing leaching measures and again following the initial leaching event. Through in situ ground temperature monitoring and laboratory measurements of soil sample TDS, a site-specific model was developed that incorporates ground temperature to describe the relationship between TDS and electrical conductivity. The objectives of this study are to investigate: 1) the characteristics of spatial salinity distribution in the area; 2) the diagnosis of primary causes of salinization in the study area; and 3) changes in TDS before and after salt leaching, and the efficacy of the salt removal measures.

2 Materials and methods

2.1 Study area

The study area is located in Jingtai County, Gansu Province, China (Figure 1a), a transitional zone between monsoon and non-monsoon regions characterized by a temperate continental arid to semi-arid climate. The region experiences consistently low annual precipitation and high evaporation, with a mean annual precipitation of 185.7 mm and a mean annual evaporation reaching 2,433.7 mm. Most rainfall occurs between June and September (Lian et al., 2022). Due to the specific geological structure and relatively enclosed topography of the irrigation district, a significant amount of irrigation water infiltrates into the subsurface but cannot drain effectively (Figure 1b).

Figure 1

Figure 1. Study site and layout of geophysical survey profiles. (a) Geographic location of the study site; (b) Topography and geomorphology of the surrounding study area; (c) UAV aerial photograph of the study area and survey line locations (November 2024); (d) UAV aerial photograph of the study area and survey line locations (June 2025).

To remediate the extensive moderately and severely saline soils in the study area, an open-ditch drainage-based strategy was implemented. This involved establishing a comprehensive irrigation and drainage system across the salt-affected area, combined with the application of soil amendments, microbial agents, and bio-organic fertilizers. Additionally, salt-tolerant crop varieties, cultivation techniques, and supporting facilities were promoted for saline soil management.

2.2 ERT and EMI data acquisition

The ERT and the EMI methods were both employed to conduct two geophysical surveys: the first before the implementation of remediation measures (November 2024) and the second after the construction of drainage ditches and the initial salt leaching event (June 2025). This approach allowed for a comparison of soil moisture and salinity characteristics before and after the measures.

Survey lines were primarily established based on surface crop growth conditions and topographic features. Their locations in May 2024 were shown in Figure 1c. The ERT profile P1 was oriented north-south across an area of abandoned saline soil, exhibiting surface salt crusts and powdery salt crystals. Profile P2, running approximately east-west, traversed abandoned saline land, a small patch of normally cultivated land, a previously deteriorated drainage ditch, and an extensive normally farmed area. The EMI profile P3 was laid out parallel to an old drainage ditch, with normally cultivated land to the north and abandoned wasteland to the south. A borehole was drilled at the southern end of profile P1 to a depth of 10 m. Soil lithology was logged during drilling, and a slotted PVC pipe was installed with sensors for groundwater level monitoring.

The survey line layout in June 2025 is shown in Figure 1d. The ERT profile P4 completely overlapped the original profile P1 to compare soil electrical conductivity before and after the drainage measures. EMI data were also collected along P4 to compare the results from the two methods. The east-west ERT profile P5 in the southern part of the study area crossed normally cultivated land, abandoned saline soil, and two newly excavated drainage ditches, to analyze the impact of the new ditches on soil moisture and salinity dynamics. The EMI profile P6 coincided with profile P3, running parallel to the drainage ditch, to compare the electrical conductivity changes in both cultivated and abandoned plots before and after salt leaching.

The ERT operates on the principle of resistivity/conductivity contrast between a target body and the surrounding medium. Current is injected into the ground via two electrodes inserted into the earth’s surface, while two other electrodes measure the resulting potential distribution (Figure 2a, b). The measured apparent resistivity is then inverted to determine the resistivity/conductivity distribution of the subsurface medium, allowing for inferences about geological structures, lithology, water content, or contamination. Given the strong correlation between salt content and electrical conductivity in saline soils, this method is effective for detecting the spatial distribution of salinity. Data acquisition utilized a Wenner array configuration with an electrode spacing of 2 m. Survey lines P1 and P4 each had a total length of 358 m, line P2 was 718 m long, and line P5 was 598 m long. The position and elevation of each electrode were surveyed using Real-Time Kinematic (RTK) GPS for topographic correction during the inversion process. The acquired apparent resistivity data were inverted using Res2dINV software (Loke, 2015).

Figure 2

Figure 2. (a) Field data acquisition of ERT method, (b) Schematic diagram of data acquisition mode.

The EMI is an efficient, non-contact technique for soil salinity investigation. It determines the conductivity characteristics of the subsurface medium by measuring the ratio of the primary magnetic field transmitted by the system to the secondary magnetic field component induced in the ground. As shown in Figure 3, Data were collected using a GEM-2 electromagnetic induction sensor manufactured by Geophex Ltd. (Won and Huang, 2012). Higher-frequency signals primarily probe shallower soil layers, whereas lower-frequency signals investigate deeper strata, thereby enabling rapid characterization of the subsurface electrical structure. However, field tests revealed that within the instrument’s operational frequency band, excessively low frequencies suffer from a low signal-to-noise ratio and severe noise interference. After balancing the requirements for depth of investigation and signal quality, five frequencies—ranging from high to low (1,525 Hz, 5,325 Hz, 18,325 Hz, 63,025 Hz, and 90,025 Hz)—were ultimately employed. After processing the EMI data by removing local anomalous noise and averaging values within a 1 m range, the data were inverted using the open-source software EmagPy (McLachlan et al., 2021) to obtain the conductivity distribution with depth along the survey lines.

Figure 3

Figure 3. Field data acquisition using the EMI method.

2.3 Inversion of ERT and EMI data

The apparent conductivity measured by ERT and EMI methods represented a composite response of subsurface geological bodies, and only equaled the true conductivity under homogeneous infinite half-space conditions. In practical scenarios, formation conductivity exhibited complex spatial distribution patterns due to variations in lithology, moisture content, and salinity. The discrepancy between apparent conductivity and true resistivity necessitated inversion procedures to accurately determine both conductivity and depth parameters.

The inversion processes for both ERT methods and EMI methods were based on minimizing the following equation:

${\mathrm{\Phi }=\mathrm{\Phi }}_d+\alpha {\mathrm{\Phi }}_m(1)$

where ${\mathrm{\Phi }}_d$ denoted the data misfit between the observed and forward-modeled responses. It is calculated as

${\mathrm{\Phi }}_d=d-F\left(m\right)(2)$

where d denoted the observed data, m represented the model parameters, and F(m) signified the forward modeling process for parameter model m. Φ_m constituted the model penalty term incorporated to solve the ill-posed minimization problem, typically implemented as a smoothing matrix of model parameters. The $\alpha$ served as the coefficient that balances the influence of the model smoothing term on the overall inversion objective function. The inversion for ERT methods constituted a two-dimensional inversion problem, whose optimization process with smoothness constraints was based on the following equation (Loke, 2012):

$\left(J^{T}J+\alpha R\right)\Delta m=J^T(d-F\left(m_k\right)-\alpha R{m}_k(3)$

where J represented the Jacobian matrix, and R denoted the model roughness matrix. The forward modeling process for induced electromagnetic methods employed a computational approach based on solving Maxwell’s equations, while the inversion optimization procedure utilized the L-BFGS-B algorithm (McLachlan et al., 2021) from the scipy open-source toolkit (Virtanen et al., 2020). The above Equations 1–3 was used to estimate the soil bulk conductivities from ERT and EMI measurements in this study.

2.4 Borehole core lithological information

The stratigraphy of the study area was revealed by borehole lithological logging. The shallow subsurface consisted of a 0.5 m thick layer of loam, underlain by clay loam from 0.5 to 1.0 m depth. From 1.0 to 4.1 m, the material was loamy sand, followed by another clay layer from 4.1 to 6.0 m. A gravelly sand layer was present from 6.0 to 8.0 m, with weathered bedrock extending from 8.0 m to at least 10.0 m depth. The groundwater depth measured within the borehole was 0.8 m in November 2024. Following the excavation of the new drainage salt ditch, the groundwater depth dropped to 2.3 m by June 2025.

2.5 Calibrated relationship between soil electrical conductivity, temperature, and salinity

To investigate the relationship between electrical conductivity and salinity derived from ERT and EMI methods at the study site, sensors were deployed within an 80 cm depth range at the borehole location to measure soil bulk electrical conductivity (EC) and temperature (T) at 10 cm intervals. Core samples from various depths at the borehole were collected and analyzed. The electrical conductivity (EC_1:5) was determined using clarified suspensions with a 1:5 soil-to-water ratio at 25 °C, while the total dissolved solids (TDS) of soil extracts were obtained through the oven-drying method. A relationship model between EC_1:5 and TDS was established. Based on the sample measurements, EC_1:5 and TDS exhibited a strong linear correlation, yielding the following fitting equation:

$TDS=0.0269\times {EC}_{1:5}+0.1942(4)$

Since field-measured soil electrical conductivity was not only strongly correlated with salinity but also influenced by temperature variations, particularly in shallow soil layers where significant temperature gradients occur along the depth profile. Therefore, based on depth-specific soil temperature monitoring data, the following formula was employed to normalize the EC measurements at different depths to standardized values at 25 °C (Ko et al., 2023):

${EC}_{25}=\frac{EC}{1+\beta \left(T-25\right)}(5)$

where $\beta$ represented a soil-dependent coefficient, set at 2% for non-clay conditions and 1% for clay conditions. Given that the shallow subsurface of the study area primarily consisted of sandy loam, $\beta$ was assigned a value of 2% in this investigation.

After converting the field-measured EC to temperature-normalized EC₂₅, a strong linear correlation was also observed with laboratory-measured EC_1:5 at the same reference temperature. The regression analysis yielded the following relationship:

${EC}_{1:5}=0.9715\times {EC}_{25}-99.4299(6)$

Based on Equations 4–6, the relationship model between soil electrical conductivity and TDS in the study area could be derived as follows:

$TDS=0.0261\frac{EC}{1+0.02\left(T-25\right)}-2.4805(7)$

During data acquisition using ERT and EMI methods, the daily average ground temperature profile along the depth at the borehole location was shown in the Figure 4. Given the relatively flat terrain and minimal surface heterogeneity at the study site, spatial variations in ground temperature at identical depths across different locations were negligible. By applying the normalization Equation 7 to the inversion results from both ERT and EMI methods, the TDS distribution across various locations was determined. ERT and EMI surveys indicated that high salinity content was primarily concentrated within 2 m below the surface. The average TDS within this 2-m depth range was calculated. During temperature correction, ground temperature monitoring was only available at a depth of 80 cm. Considering the variation in ground temperature below 80 cm may be insignificant, the temperature value recorded at the 80 cm depth was uniformly adopted for the 80–200 cm depth interval.

Figure 4

Figure 4. Ground temperature profiles during two data acquisition campaigns.

3 Results

3.1 Characteristics of EC before salt leaching

Along survey line P1, the surface was characterized by abundant salt accumulation, presenting a white appearance across the entire area (Figure 5a). The inverted electrical conductivity profile was generally characterized by high values in the shallow subsurface and lower values at depth (Figure 5b). Within the upper 2 m, the conductivity predominantly exceeded 200 mS/m, with localized, discontinuous zones exhibiting very high conductivity over 500 mS/m and a maximum value of approximately 740 mS/m. This indicated significant shallow salt accumulation in the abandoned saline soil. The thickness of the high-conductivity layer increased to approximately 5–8 m within the segments of 220–240 m and 260–270 m along the survey line. At depths between 2 m and 12 m, conductivity ranged from 70 to 200 mS/m, below which the profile transitions into a low-conductivity zone.

Figure 5

Figure 5. (a) Surface characteristics and the location of the ERT survey line (P1, orange line) and the borehole (red triangle), and (b) ERT inversion results.

A comparison between the soil texture and lithology from borehole data and the ERT inversion profile revealed a notable conductivity contrast at around 2 m depth. While the groundwater depth was at 80 cm during data acquisition, the observed variation in shallow conductivity with depth did not show a clear correspondence with either the soil texture boundaries or the water table interface. This suggested that the shallow subsurface conductivity was primarily influenced by salt content, with major salt accumulation concentrated within the upper ∼2 m. The bedrock interface identified from the borehole correlated well with the deep low-conductivity layer (below 25 mS/m) in the ERT inversion results, and the interpreted bedrock surface gradually deepened from south to north along the survey line.

Survey line P2 ran west-to-east across a topographic gradient, with the western end approximately 1.4 m higher than the eastern end. A clear correlation was observed between crop growth status and the inverted shallow soil electrical conductivity (Figure 6a,b). In the abandoned saline wasteland, intense evaporation drove salts upward and leads to accumulation at the surface, forming areas of high conductivity. In normal cultivated plots, the shallow-layer electrical conductivity was below 50–100 mS/m. In areas where crop growth was impeded, located at lateral distances of 80–120 m and 625–645 m, the corresponding shallow-layer conductivity was higher, measuring 100–160 mS/m. Within uncultivated wasteland areas, the conductivity reached 200–400 mS/m in the 45–70 m and 195–250 m sections, and 310–800 mS/m in the 120–170 m section. In the normally cultivated section between 270 and 430 m, perennial irrigation and leaching effectively flushed salts from the shallow layer, forming a low-conductivity zone with minimal downward salt migration. However, a localized high-conductivity anomaly (160–320 mS/m) was present at greater depths (3–13 m).

Figure 6

Figure 6. (a) Surface conditions, and (b) ERT inversion results at survey line P2.

The observed spatial variation in deep conductivity—differing by a factor of 3–4 over a 150 m length—suggested that groundwater flow was obstructed, preventing the efficient outflow of saline water and resulting in localized zones of high salinity at depth. A localized low-conductivity zone (below 100 mS/m) between 500 and 520 m at depths of 3–13 m might be associated with variations in local stratigraphy or lithology that impede groundwater movement. Integration with borehole data from line P1 indicated that the bedrock interface beneath this line deepened significantly, reaching an estimated depth of 20–25 m.

The surface characteristics and the inversion results from the EMI for survey line P3 prior to the implementation of remediation measures were shown in Figure 7. This line was situated approximately 5 m from an existing drainage ditch. The first half of the survey line (0–125 m) traversed normally cultivated land, while the second half (125–320 m) crossed abandoned saline wasteland (Figure 7a). The EMI results revealed a distinct contrast in shallow electrical conductivity between the two land use types. The electrical conductivity ranged from 200 to 240 mS/m in normal cultivated plots, and a significantly higher conductivity of 420–1,220 mS/m was observed in abandoned wasteland (Figure 7b).

Figure 7

Figure 7. (a) Surface characteristics, and (b) EMI inversion results at survey line P3.

3.2 Characteristics of EC after salt leaching

Following the excavation of drainage ditches and the implementation of salt leaching measures, the inverted electrical conductivity profile for survey line P4 in June 2025 was shown in Figure 8. Within the shallow 2 m depth, the conductivity generally remained above 200 mS/m, with localized, discontinuous zones exceeding 500 mS/m and a maximum recorded value of 692 mS/m (Figure 8b). This indicated that after the initial leaching process, the peak conductivity decreased only slightly, and the extent of high-conductivity zones in the shallow subsurface showed no significant reduction overall. Integrating the stratigraphic information from the nearby borehole, the presence of a clay loam layer within the 0.5–1.0 m depth range was identified as a likely primary reason for the relatively low leaching efficiency.

Figure 8

Figure 8. (a) Surface characteristics, (b) ERT inversion results, and (c) EMI inversion results at survey line P4.

The inversion results from the EMI data, collected concurrently along the same P4 survey line, were presented in Figure 8c. Overall, a strong consistency was observed between the conductivity profiles derived from the EMI and ERT inversions. The sharp conductivity transition boundary at approximately 2 m depth can be identified from the EMI results. Laterally, the positions of localized shallow high-conductivity zones and thicker high-conductivity layers (e.g., within the 220–240 m and 260–270 m intervals) also aligned well with the ERT results. Furthermore, the range of conductivity within the upper 2 m was consistent between the two methods. In comparison, while the EMI exhibits a shallower investigation depth, it offered significantly faster data acquisition rates, providing a distinct advantage for rapid, large-area surveys.

The surface characteristics of survey line P5 before the implementation of remediation measures (November 2024) were shown in Figure 9a, while the post-remediation surface conditions (April 2025) and the corresponding ERT inversion results were presented in Figures 9b,c, respectively. Consistent with the other survey lines, a strong correlation existed between shallow electrical conductivity and land cultivation status. The normal cultivated plot, located between 10 and 110 m along the transect, was characterized by a shallow-layer electrical conductivity of 50–100 mS/m. The segment from 110 to 580 m, originally abandoned saline wasteland, exhibited significantly higher shallow conductivity of 160–500 mS/m, with a peak value of approximately 700 mS/m, resulting in a low emergence rate for newly planted crops. A new drainage ditch was excavated at the 220–240 m section, based on the course of a pre-existing, deteriorated ditch. The shallow conductivity on both sides of this newly dredged ditch was slightly lower than in other abandoned areas, suggesting that the original ditch still provided some residual functionality. In contrast, around the newly constructed drainage ditch located at 515–535 m (with a depth of 2.7 m), the shallow conductivity values were comparable to those in areas farther from drainage infrastructure. The presence of a shallow clay layer here was a plausible explanation for the relatively slow salt leaching progress. Additionally, the groundwater flow direction intersected the ditch at an angle which might be another reason.

Figure 9

Figure 9. (a) Surface characteristics at survey line P5 in November 2024; (b) Surface characteristics in June 2025, and (c) ERT inversion results.

Compared to the north-south ERT survey line (P2), the conductivity along line P5 decreased rapidly with depth, and no secondary, elevated conductivity layer—similar to the one observed beneath the shallow high-conductivity layer in P2—was detected. This difference was likely attributable to the shallower bedrock interface (approximately 10 m depth) and the slightly higher topographic relief at the P5 line location. This topographic resulted in a locally higher water table and promoted groundwater discharge towards nearby drainage ditches and the northern part of the region.

Following the re-excavation of the drainage ditch and the implementation of salt leaching, the surface characteristics and the inverted electrical conductivity distribution from the EMI survey along line P6 (June 2025) were presented in Figure 10. The conductivity in the first half of the survey line (0–125 m) ranged from 50 to 80 mS/m, while the shallow conductivity in the second half (125–320 m) ranged from 160 to 290 mS/m (Figure 10b).

Figure 10

Figure 10. (a) Surface characteristics, and (b) EMI inversion results at survey line P6.

3.3 TDS calculated for different field plots

The east-west oriented ERT survey lines P2 and P5 traversed normally cultivated areas, zones with impaired crop growth, and abandoned wasteland. The TDS values for different plots along these lines were shown in Figure 11. In Line P2, the TDS in normally cultivated plots was below 3.2 g/kg, while in the impaired growth zones (80–120 m and 625–645 m), TDS ranged from 1.9 to 5.0 g/kg. The highest TDS in abandoned wasteland was observed between 120 and 170 m, reaching 14.0–30.2 g/kg (Figure 11a). In Line P5, TDS in normally cultivated plots was below 0.8 g/kg (Figure 11b). This line was surveyed after the construction of salt drainage ditches and spring irrigation, resulting in greater spatial variability in near-surface electrical conductivity. TDS in the abandoned wasteland along this line varied between 4 and 23 g/kg.

Figure 11

Figure 11. TDS derived from the inversion of ERT data along survey lines P2 (a) and P5 (b).

The spatial variability of TDS along survey line P1 was significant. Prior to salt leaching, TDS values along the line primarily ranged from 15.5 to 23.2 g/kg (Figure 12a). After leaching, the values were mainly between 12.4 and 20.5 g/kg (Figure 12b). The average desalination rate for the entire line was 16.8%, with a maximum removal rate of 55.3% (Figure 12c).

Figure 12

Figure 12. Change in TDS and the desalination rate along survey line P1 (P4). (a) before salt leaching, (b) after salt leaching, (c) desalination rate.

Along survey line P3, TDS exhibited a general pattern of being lower in the normally cultivated plots in the front section and higher in the abandoned wasteland in the rear section. In the normally cultivated areas, TDS was concentrated in the range of 3.4–5.2 g/kg, with slightly higher values (approximately 5.6–8.3 g/kg) within a 2-m zone from the plot edges. In contrast, TDS in the abandoned wasteland ranged from 14.9 to 33.9 g/kg (Figure 13a). After spring irrigation for salt leaching, TDS in the normally cultivated plots dropped below 0.46 g/kg, while in the abandoned wasteland, it ranged from 1.48 to 6.10 g/kg (Figure 13b). The salt leaching effect following the construction of the salt drainage ditch along this line was remarkable. The average desalination rate for the entire line was approximately 87.8%. The desalination rate reached 85.1%–97.9% in the normally cultivated plots and 67.2%–93.0% in the abandoned wasteland.

Figure 13

Figure 13. Change in TDS and the desalination rate along survey line P3 (P6). (a) before salt leaching, (b) after salt leaching, (c) desalination rate.

4 Discussion

The integrated results from the ERT and the EMI methods revealed significant spatial heterogeneity in soil electrical conductivity within the study area, indicating distinct mechanisms of salt accumulation and transport at different locations. Shallow electrical conductivity in normally cultivated plots was generally below, gradually increased in plots exhibiting impaired crop growth, and reached the highest in the abandoned saline wasteland. The high conductivity soil was predominantly concentrated within the upper 2 m. In normally cultivated crop fields, the TDS were generally below 5.0 g/kg, with most areas recording values below 3.2 g/kg. In plots where crops exhibited impaired growth, TDS ranged from 1.9 to 5.2 g/kg, whereas abandoned wasteland areas showed significantly higher concentrations, ranging from 14.0 to 33.9 g/kg.

The primary cause for this pattern is the retardation effect of the clay loam layer in shallow depth, large amount of irrigation resulting in the low conductivity in cultivated plots while the concentrated evaporation in abandoned saline wasteland with high groundwater table. In the normally cultivated plots on the northeastern side of the study area, shallow conductivity remained low (50–100 mS/m). However, a localized high-conductivity anomaly (160–320 mS/m) was detected at greater depths between 3 and 13 m. Based on the local topography, the general groundwater direction was inferred to be from west to east. Under conditions of unimpeded groundwater discharge, salts would be expected to distribute more uniformly with the flow. However, the groundwater flow in this area was sluggish, with localized stagnation constrained by the stratum and geological conditions (Figure 6). This suggested that the high salinity at depth did not primarily originate from downward leaching of surface salts but rather from the lateral transport of salts by groundwater flow from upslope areas, accumulating where groundwater movement is obstructed, thus forming localized zones of deep salt accumulation. If these drainage pathways were not promptly cleared, these salts could migrate upwards via capillary rise during evaporation, ultimately leading to surface salinization. Most existing studies about the spatial heterogeneity focus on normally cultivated fields. Although variations in conductivity or salinity are observed between different plots due to factors such as groundwater and soil properties, the differences are generally minor (Liu et al., 2016). At the watershed scale, different mechanisms of salt accumulation associated with various stages of moisture movement are common (Salama et al., 1999). However, at our study site, the electrical conductivity differences between plots ranged from several-fold to over tenfold. Moreover, due to the impeded groundwater flow, distinct phenomena of shallow versus deeper salt accumulation were observed within a distance of just a few hundred meters.

Remediation measures in the study area, including dredging or re-excavating drainage ditches to lower the regional water table and conducting leaching irrigation in winter/spring, yielded significantly different outcomes depending on location. At survey line P3 (P6), located about 5 m from a newly dredged ditch, the desalination rate reached 87.8% after the initial leaching event, indicating highly effective salt removal. The desalination rate in the normal cultivated plots in the first half was higher than that in the abandoned wasteland in the second half (Figure 13). At survey line P1, approximately 85 m from a drainage ditch, the desalination rate was only 16.8% after the initial leaching (Figure 12). Following the leaching process, localized increases in TDS were observed along some sections of the line. This was attributed to the land leveling process, during which topsoil with high salt content was piled near the field ridges. During the spring irrigation for salt leaching, the water did not submerge these ridges, leaving the accumulated salt effectively unflushed (Figure 12). Additionally, no significant reduction in shallow conductivity was observed adjacent to the newly excavated ditch on the southern side of the study area as the major direction of groundwater flow from west to east was same with the ditch direction. Borehole core data near this line revealed a clay loam layer at about 0.5–1.0 m depth. Its low permeability increased the resistance to salt transport, resulting in the low vertical leaching efficiency during spring irrigation.

The combined use of ERT and EMI proved effective for rapidly assessing the spatial distribution of soil conductivity and evaluating the efficacy of the salt removal measures. ERT offered a greater depth of investigation and, when combined with localized borehole information, could provide insights into deep stratigraphy, albeit with relatively lower data acquisition efficiency. EMI, while having a shallower penetration depth, enabled much faster data collection. In this study, the shallow soil conductivity distributions derived from both methods along the same survey lines showed good consistency, confirming the reliability of the approaches.

A comparison of the two surveys at line P1 (P4) with the observed water table drop in the borehole (from 0.8 m to 2.3 m after ditch excavation) revealed no significant change in the inverted resistivity profiles. This was attributed to the differing factors controlling electrical conductivity: in non-saline soils, conductivity was influenced by soil texture, water content, and pore water conductivity (Archie, 1942), whereas in saline soils, soil salinity typically dominates (Corwin and Lesch, 2005; Jadoon et al., 2015). Furthermore, the shallow soil, composed mainly of clay and sandy loam, had high water retention capacity. Consequently, the soil likely remained near-saturated even after the water table dropped, leading to the insignificant change in conductivity. The lagging effect of the salt transport in groundwater also contributed to this stability.

For the EMI, its application typically assumes a low induction number (LIN) condition. This assumption can be violated in environments with high frequency and high electrical conductivity. Research indicates that the LIN assumption is generally valid for earth conductivities below 12 mS/m (Beamish, 2011), a threshold vastly exceeded by the saline soils in this area (200–800 mS/m). Under such conditions, EMI inversion values may underestimate the true soil conductivity (McLachlan et al., 2021). This is consistent with other studies noting good agreement between EMI and ERT results in low-conductivity zones, but significant underestimation by EMI in high-conductivity areas (Paz et al., 2024). Nevertheless, for identifying highly conductive saline layers and comparing relative conductivity changes before and after leaching, the deviation from the LIN assumption has a relatively minor impact.

5 Conclusion

To efficiently assessing field scale salt leaching, the combined application of two geophysical tools including ERT and EMI were demonstrated in a saline wasteland in the Jingtai Yellow River Irrigation District, Gansu Province, China. Three major findings of this study are summarized as follows.

1. Comparative analysis of groundwater levels and soil profile conductivity before and after drainage ditch construction confirmed that regional salinization was primarily driven by intense evaporation coupled with a progressively rising water table due to inefficient drainage. This process led to continuous salt enrichment in shallow soils of low-lying areas, ultimately resulting in widespread land abandonment.

2. The calibrated EC-TDS relationship enabled efficient quantification of the field-scale desalination rate via ERT/EMI surveys. However, the effectiveness of remediation measures—such as dredging drainage ditches to lower the water table—was highly spatially variable. For instance, one leaching event led to a high desalination rate of 87.8% at profile P3 (5 m from a dredged ditch), compared to only 16.8% at profile P1 (85 m from the ditch).

3. The combined use of ERT and EMI, which balances ERT’s detection depth and EMI’s detection efficiency, not only enables quantitative assessment of desalination rates but also helps reveal the differential mechanisms of winter irrigation for salt leaching at the field scale. The subsurface architecture, characterized by an undulating bedrock topography and shallow, confining layers (e.g., clay and caliche), plays a critical role in controlling salt leaching efficiency.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

YY: Conceptualization, Investigation, Writing – original draft. WF: Funding acquisition, Resources, Writing – review and editing. HL: Project administration, Writing – review and editing. GL: Formal Analysis, Investigation, Writing – review and editing. SW: Data curation, Validation, Writing – review and editing. XP: Conceptualization, Formal Analysis, Funding acquisition, Writing – review and editing. XW: Investigation, Methodology, Writing – review and editing. JY: Investigation, Visualization, Writing – review and editing. CW: Investigation, Visualization, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the Pilot Project on Comprehensive Utilisation of Saline and Alkaline Land in Jingtai County, Baiyin City, Gansu Province - (Section II) Core Pilot Demonstration Project (JTYJDSDXM-SG-02), Science and technology research project, CCSHCC (Grant No. KJFZ-2019-041), and Youth Science and Technology Innovation Project, CCCC (2021-ZJKJ-QNCX16).

Conflict of interest

Author WF was employed by CCCC Second Highway Consultants Co., Ltd.

The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Archie, Ge. (1942). The electrical resistivity log as an aid in determining some reservoir characteristics. Trans. AIMe 146, 54–62. doi:10.2118/942054-g

CrossRef Full Text | Google Scholar

Beamish, D. (2011). Low induction number, ground conductivity meters: a correction procedure in the absence of magnetic effects. J. Appl. Geophys. 75, 244–253. doi:10.1016/j.jappgeo.2011.07.005

CrossRef Full Text | Google Scholar

Corwin, D. L., and Lesch, S. M. (2005). Apparent soil electrical conductivity measurements in agriculture. Comput. Electron. Agric. 46, 11–43. doi:10.1016/j.compag.2004.10.005

CrossRef Full Text | Google Scholar

Cui, G., Liu, Y., Li, X., Wang, S., Qu, X., Wang, L., et al. (2025). Impacts of groundwater storage variability on soil salinization in a semi-arid agricultural plain. Geoderma 454, 117162. doi:10.1016/j.geoderma.2024.117162

CrossRef Full Text | Google Scholar

Dong, L., Wang, W., Che, T., Wang, Y., Huang, X., Zhang, S., et al. (2025). Simultaneous retrieval of soil moisture and salinity in arid and semiarid regions using sentinel-1 data and a revised dielectric model for salty soil. Agric. Water Manag. 312, 109410. doi:10.1016/j.agwat.2025.109410

CrossRef Full Text | Google Scholar

Farzamian, M., Autovino, D., Basile, A., De Mascellis, R., Dragonetti, G., Monteiro Santos, F., et al. (2021). Assessing the dynamics of soil salinity with time-lapse inversion of electromagnetic data guided by hydrological modelling. Hydrol. Earth Syst. Sci. 25, 1509–1527. doi:10.5194/hess-25-1509-2021

CrossRef Full Text | Google Scholar

Gao, G., Guo, W., and Ren, Y. (2023). Numerical study on coupled modelling of artificial freezing marine soil. Comput. Geotech. 159, 105409. doi:10.1016/j.compgeo.2023.105409

CrossRef Full Text | Google Scholar

Guo, W., Dong, Y., and Gao, G. (2025). Coupled modelling on seasonal unsaturated saline soil subgrade with sunny-shady slope. Int. Commun. Heat. Mass Transf. 169, 109579. doi:10.1016/j.icheatmasstransfer.2025.109579

CrossRef Full Text | Google Scholar

Jadoon, K. Z., Moghadas, D., Jadoon, A., Missimer, T. M., Al-Mashharawi, S. K., and McCabe, M. F. (2015). Estimation of soil salinity in a drip irrigation system by using joint inversion of multicoil electromagnetic induction measurements. Water Resour. Res. 51, 3490–3504. doi:10.1002/2014WR016245

CrossRef Full Text | Google Scholar

Ko, H., Choo, H., and Ji, K. (2023). Effect of temperature on electrical conductivity of soils – role of surface conduction. Eng. Geol. 321, 107147. doi:10.1016/j.enggeo.2023.107147

CrossRef Full Text | Google Scholar

Lian, H., Sun, Z., Xu, C., and Gu, F. (2022). The relationship between the distribution of water and salt elements in arid irrigation areas and soil salination evolution. Front. Earth Sci. 10, 852485. doi:10.3389/feart.2022.852485

CrossRef Full Text | Google Scholar

Liu, G., Li, J., Zhang, X., Wang, X., Lv, Z., Yang, J., et al. (2016). GIS-mapping spatial distribution of soil salinity for eco-restoring the Yellow River Delta in combination with electromagnetic induction. Ecol. Eng. 94, 306–314. doi:10.1016/j.ecoleng.2016.05.037

CrossRef Full Text | Google Scholar

Loke, M. H. (2012). Tutorial: 2-D and 3-D electrical imaging surveys. Geotomo Softw. Malays.

Google Scholar

Loke, M. H. (2015). RES2DINV ver. 4.04. Rapid 2- Resist. IP Inversion Using Least Sq. Method.

Google Scholar

McLachlan, P., Blanchy, G., and Binley, A. (2021). EMagPy: open-source standalone software for processing, forward modeling and inversion of electromagnetic induction data. Comput. Geosci. 146, 104561. doi:10.1016/j.cageo.2020.104561

CrossRef Full Text | Google Scholar

Mohammed, M., El Mahmoudi, A., and Almolhem, Y. (2022). Applications of electromagnetic induction and electrical resistivity tomography for digital monitoring and assessment of the soil: a case study of Al-ahsa oasis, Saudi Arabia. Appl. Sci. 12, 2067. doi:10.3390/app12042067

CrossRef Full Text | Google Scholar

Nachshon, U. (2018). Cropland soil salinization and associated hydrology: trends, processes and examples. Water 10, 1030. doi:10.3390/w10081030

CrossRef Full Text | Google Scholar

Nguyen, V. H., Germer, J., and Asch, F. (2024). Evaluating topsoil salinity via geophysical methods in rice production systems in the Vietnam mekong Delta. J. Agron. Crop Sci. 210, e12676. doi:10.1111/jac.12676

CrossRef Full Text | Google Scholar

Paz, M. C., Castanheira, N. L., Paz, A. M., Gonçalves, M. C., Monteiro Santos, F., and Farzamian, M. (2024). Comparison of electromagnetic induction and electrical resistivity tomography in assessing soil salinity: insights from four plots with distinct soil salinity levels. Land 13, 295. doi:10.3390/land13030295

CrossRef Full Text | Google Scholar

Salama, R. B., Otto, C. J., and Fitzpatrick, R. W. (1999). Contributions of groundwater conditions to soil and water salinization. Hydrogeol. J. 7, 46–64. doi:10.1007/s100400050179

CrossRef Full Text | Google Scholar

Selim, T., Amer, A., Farzamian, M., Bouksila, F., Elkiki, M., and Eltarabily, M. G. (2025). Prediction of temporal and spatial soil salinity distributions using electromagnetic conductivity imaging and regional calibration. Irrig. Sci. 43, 157–175. doi:10.1007/s00271-024-00998-4

CrossRef Full Text | Google Scholar

Siyal, A. A., Skaggs, T. H., and Genuchten, V. (2010). Reclamation of saline soils by partial ponding: simulations for different soils. Vadose Zone J. 9, 486–495. doi:10.2136/vzj2009.0129

CrossRef Full Text | Google Scholar

Virtanen, P., Gommers, R., Oliphant, T. E., Haberland, M., Reddy, T., Cournapeau, D., et al. (2020). SciPy 1.0: fundamental algorithms for scientific computing in python. Nat. Methods 17, 261–272. doi:10.1038/s41592-019-0686-2

PubMed Abstract | CrossRef Full Text | Google Scholar

Won, I. J., and Huang, H. (2012). Magnetometers and electro-magnetomenters. Lead. Edge 23, 448–451. doi:10.1190/1.1756834

CrossRef Full Text | Google Scholar

Zhang, J., Ye, Y., and Lu, C. (2024). Salt leaching efficiency of subsurface drainage systems with mass transfer limitations. J. Hydrol. 632, 130940. doi:10.1016/j.jhydrol.2024.130940

CrossRef Full Text | Google Scholar