Next Article in Journal
Renewable Energy Communities: Frameworks and Implementation of Regulatory, Technical, and Social Aspects Across EU Member States
Next Article in Special Issue
The Impact of Organic Fertilizer Substitution on Microbial Community Structure, Greenhouse Gas Emissions, and Enzyme Activity in Soils with Different Cultivation Durations
Previous Article in Journal
Integrating Cross-Regional Ecological Networks in Blue–Green Spaces: A Spatial Planning Approach for the Yangtze River Delta Demonstration Area
Previous Article in Special Issue
Detecting Soil Tillage in Portugal: Challenges and Insights from Rules-Based and Machine Learning Approaches Using Sentinel-1 and Sentinel-2 Data
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 17
Issue 9
10.3390/su17094194
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Significant Changes in Soil Properties in Arid Regions Due to Semicentennial Tillage—A Case Study of Tarim River Oasis, China

by
Ying Xiao
1,2,
Mingliang Ye
1,2,
Jing Zhang
1,2,
Yamin Chen
3,
Xinxin Sun
4,
Xiaoyan Li
1,5 and
Xiaodong Song
1,2,*
1
State Key Laboratory of Soil and Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 211135, China
2
University of the Chinese Academy of Sciences, Beijing 100049, China
3
School of Ecology, Northeast Forestry University, Harbin 150040, China
4
College of Civil Engineering, Nanjing Forestry University, Nanjing 210037, China
5
School of Environmental Science and Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, China
*
Author to whom correspondence should be addressed.
Sustainability 2025, 17(9), 4194; https://doi.org/10.3390/su17094194
Submission received: 10 April 2025 / Revised: 24 April 2025 / Accepted: 30 April 2025 / Published: 6 May 2025
(This article belongs to the Special Issue Sustainable Soil Management: Conservation Strategies for Agricultural Resilience)
Browse Figures
Versions Notes

Abstract

:
Quantifying changes in soil properties greatly benefits our understanding of soil management and sustainable land use, especially in the context of strong anthropogenic activities and climate change. This study investigated the effects of long-term reclamation on soil properties in an artificial oasis region with a cultivation history of more than 50 years. Critical soil properties were measured at 77 sites, and a total of 462 soil samples were collected down to a depth of 1 m, which captures both surface and subsurface processes that are critical for long-term cultivation effects. Thirteen critical soil properties were analyzed, among which four properties—soil organic carbon (SOC), total phosphorus (TP), pH, and ammonium nitrogen (NH4⁺)—were selected for detailed analysis due to their ecological significance and low intercorrelation. By comparing cultivated soils with nearby desert soils, this study found that semicentennial cultivation led to significant improvements in soil properties, including increased concentrations of SOC, NH4⁺, and TP, as well as reduced pH throughout the soil profile, indicating improved fertility and reduced alkalinity. Further analysis suggested that environmental factors—including temperature, clay content, evaporation differences between surface and subsurface layers, sparse vegetation cover, cotton root distribution, as well as prolonged irrigation and fertilization—collectively contributed to the enhancement of SOC decomposition and the reduction of soil alkalinity. Furthermore, three-dimensional digital soil mapping was performed to investigate the effects of long-term cultivation on the distributions of soil properties at unvisited sites. The soil depth functions were separately fitted to model the vertical variation in the soil properties, including the exponential function, power function, logarithmic function, and cubic polynomial function, and the parameters were extrapolated to unvisited sites via the quantile regression forest (QRF), boosted regression tree, and multiple linear regression techniques. The QRF technique yielded the best performance for SOC (R2 = 0.78 and RMSE = 0.62), TP (R2 = 0.79 and RMSE = 0.12), pH (R2 = 0.78 and RMSE = 0.10), and NH4+ (R2 = 0.71 and RMSE = 0.38). The results showed that depth function coupled with machine learning methods can predict the spatial distribution of soil properties in arid areas efficiently and accurately. These research conclusions will lead to more effective targeted measures and guarantees for local agricultural development and food security.

1. Introduction

Arid regions constitute approximately 20% of the Earth’s terrestrial surface and are among the ecosystems most vulnerable to climate change [1]. As the largest country under threat of desertification [2], China features decertified land spanning over 2.6 million km2, accounting for 27% of the whole territory [3]. These areas are predominantly located in extremely arid, arid, and semiarid zones, where water resources are critically scarce [4].
Arid and semiarid regions are particularly vulnerable to desertification due to low and highly variable precipitation, high evaporation rates, nutrient-poor soils, sparse vegetation, and fragile ecosystems [5]. These regions also feature saline soils with low moisture content and extremely high electrical conductivity [6,7]. Such physicochemical constraints limit the transformation rate of organic matter, leading to the accumulation of untransformed materials. This phenomenon negatively affects soil fertility, impeding the recycling of essential nutrients required for plant growth [8,9,10]. Therefore, most desert areas are difficult for developing agriculture, sparsely populated, and economically backward. The objective of desert reclamation is precisely to promote agricultural and economic development by cultivating crops to sustain the increasing population [11].
The scientific reclamation of desert areas, defined as the systematic conversion of barren or degraded arid land into agriculturally productive land through irrigation, fertilization, and cultivation, is a primary strategy to mitigate desertification and promote sustainable land use in arid regions [11]. Reclamation practices typically involve long-term irrigation schemes, land leveling, soil amendment, and continuous cropping, all of which significantly alter soil physical, chemical, and biological properties. While natural soil development is a slow process, reclamation accelerates these changes by enhancing moisture retention, modifying soil structure, and increasing organic matter inputs [12,13]. Studies have shown that reclamation improves the physical properties of desert soils by reducing bulk density and increasing porosity, clay, and silt content, thereby improving water-holding capacity and resistance to wind erosion [14,15,16]. Chemically, reclamation enhances nutrient availability by boosting levels of soil organic carbon (SOC), total nitrogen (N), and phosphorus (P), often due to the combined effects of fertilization and stimulated microbial activity in response to crop root exudates and organic amendments [17,18]. Moreover, continuous agricultural activity promotes the cycling of nutrients and the formation of aggregates, further contributing to improved soil fertility. Overall, these transformations suggest that reclamation leads to substantial improvements in critical soil quality indicators and plays a vital role in restoring soil function in arid environments.
However, soil quality improvements from reclamation are not indefinite and may potentially impact local ecological, agricultural, and economic sustainability. According to Hua, desert soils stabilize 6–8 years after reclamation, and soil quality often begins to decline after 10 years [19]. What is more, due to the low amount of precipitation, high rate of evaporation, and unreasonable irrigation measures, salinization will appear in the reclaimed areas at a later stage, and there is a risk of them turning back to deserts [20]. Most existing studies focus on semi-shrubby deserts, grasslands, and oasis forests in regions with limited agricultural histories, and less on the pure desert reclamation. At the same time, studies have shown that at least 50 years may be required for desert soils to transform into highly productive agricultural land [21]. Short-term observations may not adequately reflect the gradual and multifaceted changes in soil properties that occur over time after reclamation [11,22]. the long-term effects of reclamation, particularly for pure desert soils, remain uncertain. Therefore, investigating the change trend of critical soil properties following the long-term reclamation of pure desert holds great significance for ensuring both local environmental sustainability and economic development.
Few studies have focused on subsoil in arid regions due to its limited development and low particle cohesion, which complicate the application of conventional investigative methods. In this context, digital soil mapping (DSM) offers a promising alternative for predicting the three-dimensional distribution of various soil properties. Over the past two decades, DSM techniques have gained significant traction for identifying spatial soil patterns and their associated characteristics [23,24,25]. Among these techniques, machine learning algorithms have become increasingly popular due to their capacity to model complex interactions between soil properties and environmental variables with high accuracy, particularly when applied to large soil datasets [26,27]. Environmental covariates can profoundly affect soil properties. Topography significantly influences soil formation and evolution through various factors, including moisture, erosion and deposition, temperature, and vegetation distribution [28]. This influence not only impacts surface soil characteristics but also alters the properties of deeper soil layers over extended periods [29]. Therefore, in DSM, topographic factors such as slope, aspect, elevation, and curvature and vegetation characteristics such as NDVI are commonly used as crucial predictive variables to characterize soil spatial distribution patterns and improve the prediction accuracy. Several studies have shown that machine learning techniques can accurately predict SOC at various scales and in a variety of ecosystems, including semiarid and arid regions [30,31]. These DSM products are essential for enhancing soil management and conservation strategies, particularly given the vulnerability of semiarid regions to climate change and anthropogenic pressures [32]. Moreover, soils exhibit significant temporal and spatial variability, necessitating an investigation of vertical distributions of soil properties [33]. Each soil property follows a unique vertical distribution pattern, varying continuously with depth within the profile. These patterns can be modeled using depth functions, which have evolved from the freehand curves pioneered by Jenny et al. to more sophisticated mathematical models [34,35]. Examples include polynomial functions, power functions, equal-area quadratic spline functions (EAQSFs) [35], exponential decay functions (EDFs) [36], and sigmoid functions [37].
When combined with machine learning algorithms, these depth functions have proven effective in modeling the three-dimensional distribution of soil properties, as demonstrated by studies on SOC [38,39]. However, research on other soil properties, particularly in arid and semiarid regions, remains limited. Therefore, further investigation is urgently needed to evaluate and optimize the performance of existing modeling approaches for predicting a broader range of soil properties in these challenging environments.
To determine the best approach for predicting soil properties in the study area in Xinjiang Province, Northwest China, three machine learning models—the multiple linear regression (MLR), quantile regression forest (QRF), and boosted regression tree (BRT) methods—combined with various depth functions were investigated. The aims of this study were to (1) predict and compare the spatial distributions of soil properties (SOC, TP, NH4+, and pH) via various DSM approaches and (2) verify the impact of semicentennial desert reclamation on local soil properties and soil evolution and verify whether it is sustainable. The selection of these properties, rather than physical indicators such as texture or structure, was based on their relatively low intercorrelation and the clarity of their variation trends with soil depth, which made them more suitable for depth function modeling and spatial prediction in this context.
The innovation of this study lies in its comparative analysis of critical soil properties along soil profiles in reclaimed arid lands, which provides a scientific basis for understanding the current status of sustainable agricultural development in desert regions. Furthermore, this study verifies the most effective method for predicting the three-dimensional spatial distribution of critical soil properties in arid environments, offering methodological insights for future digital soil mapping in similar contexts.

2. Materials and Methods

2.1. Description of the Study Area

The study area is situated in southern Xinjiang, China (81°17′–81°22′ E, 40°28′–40°31′ N) (Figure 1), at the confluence of the Hetian, Aksu, and Yerqiang rivers. It lies south of the Tarim River and north of the Taklimakan Desert. The region experiences high annual sunshine exposure, with a mean duration of 2303.1 h. Due to its arid climate, characterized by low precipitation (mean annual precipitation of 60 mm) and high evapotranspiration (1800 mm), substantial amounts of water from the Tarim River have historically been utilized to reclaim desert land and establish oasis agriculture, beginning with the founding of Aral City. The 12th Regiment Farm was born from this background. According to the local land management records and farmer descriptions, the farm has been cultivated in natural sandy lands for more than 50 years. The main crops cultivated in the study area are cotton and jujube. Information on irrigation methods, irrigation amounts, fertilizer application, and crop yields is summarized in Table 1.
However, due to variations in farmers’ management practices, which are influenced by individual experience and resource availability, it is difficult to obtain consistent and precise quantitative data. The values provided are based on field visits, expert estimations, and selected references from previous studies and are therefore intended for reference only.
In recent years, an increasing number of farms—similar to the 12th Regiment Farm—have expanded into desert areas, gradually pushing the southern boundary of the desert outward. However, the long-term effects of such land reclamation on soil properties and the sustainability of agricultural production remain unclear.

2.2. Experimental Design and Soil Sampling

In August 2019, 71 sampling sites were distributed via a regular grid layout on the farm, with intervals of 1 km from east to west and 500 m from north to south. At each sampling site, two soil drills were drilled to the depth of 1 m. Soil samples were taken from six depth intervals: 0–10 cm, 10–20 cm, 20–40 cm, 40–60 cm, 60–80 cm, and 80–100 cm. The surface layer (0–20 cm) was further divided into 0–10 cm and 10–20 cm to capture the potential effects of long-term tillage, which typically alters physical and chemical properties more strongly in the upper soil profile [40]. The soil samples from the same level at different drilling points were mixed, and visible roots and litter debris were removed from each soil sample. Finally, the samples were air-dried and sieved through a 2 mm soil sieve to prepare for further physical and chemical analysis. Previous studies in this area revealed that the soil properties in the cropland areas before cultivation were similar to those in the adjacent uncultivated sandy areas. So, in May 2024, 6 adjacent sampling sites from uncultivated virgin sandy land served as a reference. A new round of control sampling was conducted using the 2019 sampling scheme. Although there was a five-year interval between the two sampling campaigns, this design is justified given the extremely slow pedogenic processes in desert soils, where physicochemical properties remain largely stable over such timeframes [41].

2.3. Analysis of Soil Physical and Chemical Properties

A total of 13 soil properties (comprising 15 measurement indicators) were measured in this study, including soil organic carbon (SOC), total phosphorus (TP), pH, ammonium nitrogen (NH4⁺), total nitrogen (TN), total potassium (TK), cation exchange capacity (CEC), electrical conductivity (Ec), soil texture (clay, silt, and sand content), dissolvable total nitrogen (DTN), available potassium (AK), calcium carbonate (CaCO3), and nitrate nitrogen (NO3⁻). SOC was determined using the potassium dichromate volumetric method with external heating. TP and TK were measured by the alkali fusion method. Soil pH was determined by the potentiometric method. CEC was analyzed using the hexamminecobalt trichloride spectrophotometric method, and Ec was measured using the electrode method. Soil texture was determined using the hydrometer method. TN, DTN, and NO3⁻ were measured using the Kjeldahl digestion method. AK was determined by flame photometry, and CaCO3 was measured by acid–base titration. All these data were measured according to Soil and Agricultural Chemistry Analysis [42].
Eventually, we selected four soil properties from the set, which exhibited low correlations with each other and showed a clear trend of variation with depth in the profile.

2.4. Environmental Covariates

One of the primary elements influencing the formation of soil in arid and semiarid areas is topography [31]. In addition to topography, there are many other environmental variables that influence the prediction of soil properties. Because the study area is located in the desert, where the interaction between irrigation-induced groundwater and surface water with the soil often leads to the accumulation of carbonates in desert oasis regions [43]. A higher concentration of carbonates indicates soil salinization and alkalization, which severely limits plant growth and agricultural productivity [44]. The carbonate index (CAEX) and salinity index (SI) derived from remote sensing are not only efficient and convenient but also provide a clear and direct representation of the carbonate and salt content in the local soil.
Therefore, we downloaded a digital elevation model (DEM) with a cell size of 12.5 m × 12.5 m from ALOS PALSAR in order to obtain the topographic attributes [45]. Elevation (El), topographic wetness index (TWI), slope (SL), slope position (Slpposi), profile curvature (PrCu), plan curvature (PlCu), and distance to an irrigation canal (DIC) were among the terrain variables derived from the DEM.
The NDVI [46], CAEX [47], and salinity index (SI) [48] were among the auxiliary variables used in remote sensing. To estimate these indexes of the study area, data from Sentinel-2 were collected. After the transformation of the following calculation formulations for each band, the final values are obtained.
N D V I = ( N I R − R ) / ( N I R + R )
C A E X = G / B
S I = ( B − R ) / ( B + R )
where R, G, B, and NIR stand for the bands of red, green, blue, and near-infrared, respectively.
The environmental covariates were obtained using the SAGA GIS (System for Automated Geoscientific Analysis), and predictors with varying resolutions were resampled to 10 m using the bilinear interpolation approach.
Given the limited size of the 12th Regiment Farm, variations in climate and soil texture within the region are not pronounced. Consequently, this study did not consider soil texture and climate factors as environmental covariates [49]. However, the exclusion of these factors may introduce a certain degree of uncertainty or error into the modeling results. In larger or more heterogeneous study areas, the inclusion of such variables is essential to enhance model accuracy and reliability.

2.5. Modeling Approaches

Two nonlinear machine learning models, namely, quantile regression forest (QRF), boosted regression tree (BRT), and multiple linear regression (MLR), were combined with different depth functions, such as exponential functions, logarithmic functions, power functions and cubic polynomial functions, to facilitate the digital mapping of soil properties (SOC, TP, NH4⁺, and pH).
BRT analysis is a parametric data mining method capable of managing both linear and nonlinear relationships [50]. This approach has been extensively applied in digital soil mapping (DSM) [47].
A regression forest (RF) is an ensemble model consisting of multiple decision or regression trees, each built using a randomly selected vector sampled independently but following the same distribution as the others in the dataset [47]. Meinshausen suggested altering the RF procedure’s outputs by permitting the calculation of the targeted variables’ prediction intervals [51]. The QRF technique takes into account the spread of the response variable from which prediction intervals are formed, whereas the RF technique merely preserves the mean of the observations that fall into each node and ignores all other information for each node and each tree.
Spatial prediction has long employed linear regression, one of the first statistical methods [52]. Multiple linear regression models the relationship between variables by fitting a straight line or, in higher-dimensional spaces, a plane determined by the number of independent variables [53]. As a result, each soil property has been given a quantitative estimate using the MLR model.
In this study, hyperparameter tuning and overfitting control were carefully considered to ensure the robustness and generalizability of the machine learning models. For the QRF model, hyperparameters were optimized using a grid search approach combined with leave-one-out cross-validation (LOOCV). The final model was configured with 500 trees (n_estimators = 500), a maximum depth of 10 (max_depth = 10), and a minimum of 5 samples per leaf (min_samples_leaf = 5).
For the BRT model, a similar optimization process was applied. The model was tuned using grid search with LOOCV and early stopping based on validation loss. The best-performing configuration included a learning rate of 0.01, a maximum depth of 6, and 1000 boosting iterations, with an early stopping threshold of 50 rounds. LOOCV helped assess model generalizability, and early stopping further mitigated overfitting.
Pearson correlation coefficients between the environmental covariates and soil parameters were computed in order to identify significant environmental covariates. Each soil attribute was then predicted using only relevant factors [54]. Based on variable relevance, the most significant variables were determined for each model. Quantifying variable importance is a fundamental aspect of many applied models [55] as it aids in interpreting the influence of variables on model accuracy and identifying the most relevant environmental covariates used in the analysis [56]. The “caret” package in R 4.3.3 and RStudio 2024.12.1 was used for all modeling and variable significance computations.

2.6. Statistical and Validation Strategy

Leave-one-out cross-validation (LOOCV) was performed to evaluate the prediction methods. In order to ensure that every data point appears in the test set at least once, cross-validation offers a framework for generating multiple training/test splits in the dataset. This approach has the benefit of impartiality and consistent performance on smaller datasets.
The model performance was assessed using three validation criteria: the coefficient of determination (R2), the mean absolute error (MAE), and the root mean square error (RMSE).
R M S E = ∑ i = 1 n P i − O i 2 n
M A E = ∑ i = 1 n P i − O i n
R 2 = ∑ i = 1 n O i − O a v g × P i − P a v g ∑ i = 1 n O i − O a v g 2 × P i − P a v g 2 2
where Pi, Oi, Oavg, Pavg, n, and p stand for the model’s total number of explanatory variables, number of data points, average of the observed and predicted soil property values at the ith point, and predicted and observed values, respectively.

3. Results and Discussion

3.1. Correlation Analysis and Descriptive Statistics

SOC and pH are among the most critical soil properties due to their fundamental roles in regulating nutrient availability, microbial activity, and overall soil health [57], and they were therefore selected as foundational parameters in this study. The selection of other soil properties was based on their correlation and the clarity of their variation trends with soil depth. We aim to select properties that exhibit low correlation but high interpretability, so that when fitting with depth functions, they are more diverse and possess clear ecological significance. We measured a variety of properties and selected four soil properties, SOC, pH, NH4+, and TP, which are related to other soil properties, but their correlations with each other are low (Figure 2). The soil properties not selected, such as TN and Ec, were due to their high correlations with selected indicators—TN was highly correlated with SOC (r = 0.94), and EC showed a substantial correlation with pH (r = 0.63), indicating potential multicollinearity. In this way, we ensured the maximum representativeness and heterogeneity of the selected attributes. Although CaCO3 exhibited relatively low correlations with all other measured properties, its variation with depth was minimal, rendering it unsuitable for depth function modeling. Nonetheless, CaCO3 was retained as a supplementary variable to support the interpretation of soil formation and salinization processes in the discussion.
The summary statistics of the soil properties in the study area are presented in Table S1, while those of the CK are presented in Table S2. Scatter diagrams of SOC, pH, NH4+, and TP are shown in Figure 3. To enhance the visual clarity of depth function plots and reduce overlapping, the ten maximum and ten minimum values from each depth were excluded only for figure rendering purposes. All original data were used in the model fitting and statistical analyses.
In the study area, the mean values and standard deviations of the SOC content decreased with depth, and the mean values were 4.49, 4.19, 3.61, 2.69, 2.31, and 2.19 g/kg, respectively. The coefficients of variation (COVs) of the SOC content at the 0–20 cm depth were less than 35%, indicating moderate variation, and those at the 20–100 cm depth ranged from 35% to 43%, indicating high variation. The mean values of pH increased with increasing soil depth and were 8.83, 8.86, 8.91, 8.99, 9.02, and 9.07. The COVs of the pH values at the six depths ranged from 2.54% to 3.11%, indicating low variation. The mean values of NH4+ increased with depth from 0 to 50 cm and then decreased with depth from 50 to 100 cm, with values of 3.06, 2.80, 2.43, 2.07, 2.33, and 2.48 mg/kg, respectively. The COVs of the NH4+ content at the 0–20 cm depth were less than 35%, indicating moderate variation, and those at the 20–100 cm depth ranged from 35% to 62%, indicating high variation. The mean values and standard deviations of the TP content decreased with depth, and the mean values were 1.04, 1.05, 0.90, 0.78, 0.71, and 0.70 g/kg, respectively. The COVs of the TP content at depths of 0–20 cm were less than 35%, indicating moderate variation, and those at depths of 20–100 cm were less than 25%, indicating little variation.
In the CK treatment, there was no significant trend in the values of SOC, pH, NH4+, and TP with increasing depth. We plotted the changes in the values of the four soil properties at different depths after semicentennial tillage, as shown in Figure 3, which revealed the vertical changes in the soil properties following semicentennial tillage in desert areas. The observations revealed that the accumulation of SOC is higher than that of the CK at 0–60 cm but lower in the 60–100 cm layer. Additionally, the pH of the reclaimed soils was lower than that of the CK soils, while the NH4+ and TP contents were greater than those of the CK soils in both the topsoil and subsoil.
Meanwhile, significant changes occurred in CaCO3, soil texture, and EC after 50 years of cultivation (TG) compared with the uncultivated desert soil (CK), as shown in Table S3. As Ec is highly positively correlated with soil salinity, it was used in this study as an indicator of soil salt content [58]. Ec increased substantially in surface (0–20 cm) and deep layers (80–100 cm), indicating strong salt accumulation trends. CaCO3 contents increased consistently across all depths, with the most pronounced increases in subsoil layers (532.54% at 80–100 cm). Clay and silt contents rose sharply (clay increased by over 1800% in all layers), while sand content decreased significantly, reflecting a major shift in soil texture from sandy to finer-grained material.
In arid and semiarid regions, nutrient accumulation occurs primarily through artificial means. The unique natural characteristics of these areas, particularly the ratio of evaporation to precipitation, pose challenges for the accumulation of essential nutrients such as C, N, and P, which are necessary for the growth of plants [59].
The contents of SOC are determined primarily by the balance between exogenous crop carbon inputs and the decomposition of existing SOC [60]. In arid and semiarid regions, differences in temperature between the topsoil and subsoil affect microbial activity, which subsequently influences SOC mineralization at different depths [61,62,63]. After reclamation, topsoil has a relatively high wet aggregate stability that facilitates the formation of mineral-associated organic carbon compounds [64,65]. The minimal variation observed in SOC at different depths in the CK can be attributed to the high and uniform sand content.
The reduced levels of SOC in the 60–100 cm layer relative to those in the CK soil can be explained by the evaporation rate exceeding the precipitation rate, which limited the downwards migration of soluble organic carbon. The study area is located on the northern edge of the Taklimakan Desert and is characterized by scarce precipitation and rapid evaporation. This environment has led to a gradual accumulation of alkaline ions and soluble salts [66]. The low vegetation cover results in minimal root activity, which adversely affects the decomposition of organic matter. Cotton, a tap-rooted crop, possesses a main root that can extend to depths of 1–2 m; however, its numerous lateral roots are primarily concentrated within the 0.2–0.6 m range [67,68]. These lateral roots play crucial roles in promoting SOC decomposition and reducing soil alkalinity, thereby contributing to differences in pH between the 20–60 cm and 60–100 cm soil depths.
NH4+ and TP are both essential nutrients; however, natural processes in desert soils are insufficient for their effective accumulation. Under natural conditions (CK), the concentrations of NH4⁺ and TP were generally low and exhibited minimal variation with depth. This pattern can be attributed to both the low nutrient inputs and weak vertical redistribution in the absence of anthropogenic disturbance [69]. After reclamation, the distribution pattern of NH4⁺ became more stratified, showing a gradual decrease from the surface to 50 cm, followed by a slight increase at deeper depths. This trend likely reflects the combined effects of fertilization, irrigation-driven leaching, and subsoil retention. At the depth of 0–50 cm, surface-applied NH4⁺ may have been mobilized downward by irrigation water, particularly in these predominantly sandy soils with low cation exchange capacity [70,71]. At the depth of 50–100 cm, the slight increase in NH4⁺ could be attributed to reduced microbial activity and oxygen availability, which may inhibit nitrification and lead to ammonium persistence [72]. Additionally, occasional textural transitions or moisture barriers may retard further leaching, allowing for localized accumulation at depth.
In contrast, TP exhibited surface enrichment across the reclaimed profiles due to its low solubility and strong adsorption to soil particles. Phosphorus tends to remain near the surface unless subjected to significant physical disturbance. Moreover, under alkaline conditions, phosphate ions readily react with calcium to form insoluble calcium phosphates, further limiting its mobility in the soil profile [73,74,75].
The observed changes in soil properties (SOC, pH, NH4⁺, TP, Ec, CaCO3, and soil texture) reflect the complex interplay of factors resulting from long-term tillage and irrigation in desert soils. While the cultivation has improved soil fertility and water retention in the surface layers, it has also led to the accumulation of salts (indicated by increased Ec) and changes in pH, which may pose challenges for sustainable soil management in the long term. The increase in clay and silt content indicates a shift toward a finer soil texture, which can enhance soil fertility but may also require management to avoid issues like soil compaction or surface crusting. Thus, while there are improvements in soil quality, there are also emerging risks associated with salinization and pH changes that need careful attention.

3.2. Optimal Depth Functions Selection

Under normal circumstances, there is little difference in the vertical distribution of desert soil properties, which has been verified in Table S2. In the cultivation process, watering and fertilization will produce a series of effects, resulting in different degrees of soil property changes in depth, which has also been verified in Figure 4. In order to further analysis and demonstrate the variation trend of desert soil properties with depth after semicentennial tillage, we used depth functions to simulate them. SOC, pH, NH4+, and TP within 1 m were fitted by different depth functions, including the exponential function, power function, logarithmic function, and cubic polynomial function. The results revealed that different function types had significantly different fitting effects for each soil property (Table 2 and Figure 4). The binomial exponential function had the best fitting effect for SOC (R2 = 0.90) and TP (R2 = 0.77), and the cubic equation had the best fitting effect for pH (R2 = 0.92) and NH4+ (R2 = 0.89).
This showed that after semicentennial tillage, the contents of SOC and TP decreased with the increase in depth, pH showed an “S”-shaped increasing trend at depth, and NH4+ decreased at the 0–50 cm depth and increased at the 50–100 cm depth. These are similar to the trends in most farmland soil properties but completely different from the trends in desert soil. For example, Murphy used an exponential function to fit SOC at 1 m depth of cultivated land on the Darling Riverine Plain in Australia and obtained a high R2 [76]. Emiru and Gebrekidan analyzed the content of TP in 0–40 cm soil of arable land on the Ethiopian plateau, which showed that TP had a decreasing trend with the increase in depth [77]. And Zhang proposed a sigmoid depth function to represent the distribution of soil pH with depth [37].
In fact, depth functions can reflect the evolution of soil. For instance, Minasny and McBratney proposed that the depth function of soil pH can reflect the degree of soil development [24]. In this study, the depth function of soil pH exhibited a decline in the tilled layer, with only minor fluctuations observed in the deeper layers. This pattern is attributed to prolonged farming practices, including fertilization and irrigation, which facilitate the continuous decomposition of organic matter and the subsequent production of organic acids [78]. Furthermore, long-term cultivation fosters increased microbial activity within the soil, with microorganisms generating acidic by-products during organic matter decomposition, thereby exacerbating soil acidification [79]. In contrast, the deeper soil layers are less influenced by acidification and water evaporation, benefiting from the soil’s inherent buffering capacity, which helps maintain a relatively high pH in these deeper zones [80].
These results showed that semicentennial tillage strongly changed the vertical distribution of critical soil properties and transformed the uniform distribution of desert soil properties from topsoil to subsoil, which may accelerate soil evolution and makes it more suitable for agriculture.

3.3. Model Performance

Semicentennial tillage tremendously changed the soil properties in the study area, not only in the soil profiles but also in different regions, and these changes might have different rates with increasing depth. Two nonlinear machine learning models, namely QRF and BRT, and a linear model, namely MLR, were used to predict the distributions of the four soil properties. The performances of the models were evaluated via leave-one-out cross-validation (Table 3). On the basis of the validation criteria, the nonlinear models (QRF and BRT) were more accurate than MLR. The findings demonstrated that the QRF and BRT techniques performed the best in predicting every property among the models under study. This result aligns well with the conclusions of Youssef et al. (2016), who found that ensemble machine learning methods such as RF and BRT outperformed linear models in accurately mapping landslide susceptibility [81].
The superior performance of QRF in predicting soil properties arises from its ability to capture complex, nonlinear relationships between environmental covariates and soil attributes while providing robust uncertainty estimates. QRF leverages ensemble learning with bootstrap aggregation, reducing overfitting and enhancing predictive accuracy. BRT, while also capable of modeling nonlinear interactions, iteratively refines predictions but is more sensitive to hyperparameters and prone to overfitting in smaller datasets [82]. In contrast, MLR assumes linearity, failing to represent intricate soil–environment interactions, leading to the lowest accuracy. The findings underscore the necessity of machine learning approaches over parametric models in digital soil mapping. This observation is further corroborated by previous studies, such as those by da Silva Chagas et al. (2016) and Pahlavan-Rad et al. (2020), who demonstrated that RF consistently outperforms MLR in predicting the spatial variability of soil properties, particularly under semiarid and dryland conditions [83,84].
The overall trends of the SOC distribution at the six depths were significantly different, with a spatial pattern of low values in the southeast and west and high values in the middle and north at depths of 0–40 cm. At depths of 40–100 cm, the spatial distribution pattern is characterized by high values in the southeast and west and low values in the middle and north, which was completely opposite to the predicted results for the surface layer (Figure 5). The inversion in the spatial distribution of SOC between surface and subsurface layers may also be explained by variations in soil texture along the soil profile. As shown in Table S3, the contents of silt and clay increased markedly across the entire profile following cultivation, with a slight upward trend with depth. This vertical trend is indicative of selective particle translocation driven by rainfall infiltration and wind erosion processes. Specifically, fine particles (silt and clay) tend to settle at deeper depths due to eluviation following precipitation events, while surface layers become increasingly dominated by sand fractions owing to wind-driven deflation and the removal of finer materials [85]. Spatially, areas with higher surface SOC levels were associated with finer surface textures (greater silt content), which are more retentive of SOC and reduce leaching potential. Conversely, in areas with sandier surface textures and lower surface SOC, enhanced vertical permeability likely facilitated downward SOC transport, resulting in subsurface enrichment. These contrasting textural regimes modulate the SOC retention and mobility capacity, even under uniform management practices.
The overall variation trends of the spatial distributions of pH at the six depths were obviously different, and the pH was generally low at depths of 0–20 cm, roughly following the distribution characteristics of farmland. At depths of 20–100 cm, the spatial distribution pattern is characterized by low values in the southeast and west and high values in the middle and north (Figure 6). The spatial inversion in pH distribution between the surface and subsurface layers can also be explained by mechanisms similar to those driving the SOC distribution, both of which reflect the influence of soil texture on the vertical redistribution of critical soil properties. Areas with higher sand content and lower SOC content are more susceptible to acidification due to limited buffering capacity, greater leaching losses of base cations, and the accumulation of acidifying substances such as fertilizers and organic acids. In contrast, subsurface layers with higher clay and silt contents exhibit greater cation exchange capacity (CEC), which enhances their ability to retain alkaline ions (e.g., Ca2⁺, Mg2⁺, K⁺), thus maintaining a higher pH.
These transformations in soil properties collectively confirm that, over semicentennial tillage, the soils in this formerly desert area have evolved from unstructured sand into weakly developed arid soils with high base saturation. This trajectory of soil development under long-term tillage reflects a progressive shift toward conditions more suitable for cultivation.
There was no significant difference in the overall trend in the NH4+ spatial distribution at the six depths (Figure 7). The overall change trend of the spatial distribution of TP at the six depths was consistent, with low values in the southeast and west and high values in the center and north. The distribution characteristics at 0–40 cm were roughly consistent with those of farmland (Figure 8).

3.4. Confidence Interval of Prediction

The QRF model was used to map the predicted uncertainties. QRF can be used to estimate the value of a target variable for any required quantile. The upper (95%) and lower (5%) quantile maps can then be used to compute prediction intervals (e.g., 90%). The 5% quantile value indicates that 5% of all possible predictions are lower than this value. This means that under the same conditions, one can predict with 95% confidence that the result will be higher than this value. The 95% quantile value means that 95% of all possible predictions are lower than this value. This means that under the same conditions, one can predict with 5% confidence that the result will be higher than this value. In this study, we used 5% and 95% to represent the confidence interval of prediction (Supplementary Figures S1–S8).

4. Conclusions

In this study, the four most important soil properties (SOC, pH, NH4+, and TP) with the lowest correlation were selected as the assessment objects to measure the sustainability of reclamation in the study area. The results showed that reclamation has greatly changed the soil properties in the study area, and the vertical distribution of soil properties was similar to that of cultivated soil in non-arid areas but completely different from that of desert soil. While these findings suggest positive transformation, it is equally important to recognize emerging risks. Substantial increases in Ec and CaCO3 content signal a potential long-term risk of soil salinization. Ec is widely recognized as an indicator of soluble salt accumulation, and its elevated values point to ongoing salt buildup. Similarly, CaCO3 concentrations increased steadily with depth, reaching more than a fivefold increase in deep subsoils. These changes, combined with the pronounced shift in soil texture—from sandy to silty—indicate enhanced salt retention capacity, especially under irrigation-driven leaching. Therefore, despite overall improvements, the region remains ecologically fragile, and the potential for secondary salinization cannot be overlooked.
To ensure sustainable land use, it is imperative to implement scientifically grounded and adaptive irrigation strategies. These should aim not only to support crop productivity but also to mitigate salinity risk and prevent land degradation, which are essential for the long-term ecological resilience of reclaimed arid lands.
In addition, depth functions were used to assess the variation trends with depth, and nonlinear and linear techniques were used to predict the spatial distributions of SOC, pH, NH4+, and TP. The results showed that the binomial exponential function has the best fitting effect for SOC and TP, and the cubic equation has the best fitting effect for pH and NH4+. The QRF model has the best performance in predicting the SOC, pH, NH4+, and TP contents. While the QRF model demonstrated high accuracy within the relatively homogeneous and small-scale study area, its applicability may be constrained in larger, more heterogeneous landscapes due to increased model complexity and interpretability challenges. Nevertheless, this study provides valuable insights into digital soil mapping in arid regions, particularly for enhancing the spatial characterization of critical soil attributes under limited data conditions.
Despite the insights provided in this study, we acknowledge several limitations that highlight opportunities for future research. The spatial extent of our study area is relatively limited, which may restrict the representativeness and applicability of the findings to broader desert ecosystems. Furthermore, the current analysis does not fully capture the mechanistic processes underlying the observed changes in soil properties, such as salinization, carbon cycling, and carbonate accumulation.
To better understand and predict the long-term impacts of agricultural practices in arid regions, future studies should incorporate more mechanistic, process-based approaches. This includes the development and application of scientifically grounded and data-driven models that can simulate soil evolution under prolonged cultivation. Such models should account for complex interactions among soil texture, moisture dynamics, nutrient fluxes, and anthropogenic inputs over time. Expanding the spatial and temporal scale of investigations, integrating multidisciplinary methods, and validating findings across multiple desert environments will be essential to provide a more comprehensive and generalizable understanding of soil transformation in response to long-term reclamation.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/su17094194/s1, Figure S1: The uncertainty maps of SOC predicted by the QRF model at different soil depths for 5% quantile; Figure S2: The uncertainty maps of pH predicted by the QRF model at different soil depths for 5% quantile; Figure S3: The uncertainty maps of NH4+ predicted by the QRF model at different soil depths for 5% quantile; Figure S4: The uncertainty maps of TP predicted by the QRF model at different soil depths for 5% quantile; Figure S5: The uncertainty maps of SOC predicted by the QRF model at different soil depths for 95% quantile; Figure S6: The uncertainty maps of pH predicted by the QRF model at different soil depths for 95% quantile; Figure S7: The uncertainty maps of NH4+ predicted by the QRF model at different soil depths for 95% quantile; Figure S8: The uncertainty maps of TP predicted by the QRF model at different soil depths for 95% quantile; Table S1: Statistics of SOC, pH, NH4+, TP, CEC, TN, TK, Ec, DTN, AK, CaCO3, NO3−, Clay, Silt and Sand at soil depths of 0–10 cm, 10–20 cm, 20–40 cm, 40–60 cm, 60–80 cm and 80–100 cm; Table S2: Statistics of SOC, pH, NH4+, TP, CEC, TN, TK, Ec, DTN, AK, CaCO3, and NO3− at soil depths of 0–10 cm, 10–20 cm, 20–40 cm, 40–60 cm, 60–80 cm and 80–100 cm in the CK; Table S3: Comparison of Ec, CaCO3, Clay, Silt and Sand at soil depths of 0–10 cm, 10–20 cm, 20–40 cm, 40–60 cm, 60–80 cm and 80–100 cm between TG (values in the experimental group) and CK (minus the control group).

Author Contributions

Conceptualization, Y.X. and M.Y.; methodology, X.S. (Xiaodong Song); software, Y.X. and Y.C.; validation, Y.X.; formal analysis, J.Z. and Y.C.; investigation, X.L. and Y.X.; resources, X.S. (Xiaodong Song); writing—original draft preparation, Y.X. and X.S. (Xinxin Sun); writing—review and editing, Y.X. and X.S. (Xiaodong Song); visualization, X.S. (Xinxin Sun) and M.Y.; supervision, X.S. (Xiaodong Song); project administration, X.S. (Xiaodong Song); funding acquisition, X.S. (Xiaodong Song). All authors have read and agreed to the published version of the manuscript.

Funding

The research was financially supported by the National Natural Science Foundation of China (Nos. 42322102 and 42271058), the Natural Science Foundation of Jiangsu Province (No. BK20220093), the Carbon Peak and Carbon Neutral Science and Technology Innovation Project of Jiangsu Province (No. BE2023398), and the Youth Innovation Promotion Association of the Chinese Academy of Sciences (No. 2021310).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Lian, X.; Piao, S.; Chen, A.; Huntingford, C.; Fu, B.; Li, L.Z.X.; Huang, J.; Sheffield, J.; Berg, A.M.; Keenan, T.F.; et al. Multifaceted Characteristics of Dryland Aridity Changes in a Warming World. Nat. Rev. Earth Environ. 2021, 2, 232–250. [Google Scholar] [CrossRef]
  2. Guo, B.; Wei, C.; Yu, Y.; Liu, Y.; Li, J.; Meng, C.; Cai, Y. The Dominant Influencing Factors of Desertification Changes in the Source Region of Yellow River: Climate Change or Human Activity? Sci. Total Environ. 2022, 813, 152512. [Google Scholar] [CrossRef] [PubMed]
  3. Ren, Y.; Zhang, B.; Chen, X.; Liu, X. Analysis of Spatial-Temporal Patterns and Driving Mechanisms of Land Desertification in China. Sci. Total Environ. 2024, 909, 168429. [Google Scholar] [CrossRef] [PubMed]
  4. Wei, W.; Guo, Z.; Shi, P.; Zhou, L.; Wang, X.; Li, Z.; Pang, S.; Xie, B. Spatiotemporal Changes of Land Desertification Sensitivity in Northwest China from 2000 to 2017. J. Geogr. Sci. 2021, 31, 46–68. [Google Scholar] [CrossRef]
  5. Maestre, F.T.; Benito, B.M.; Berdugo, M.; Concostrina-Zubiri, L.; Delgado-Baquerizo, M.; Eldridge, D.J.; Guirado, E.; Gross, N.; Kéfi, S.; Le Bagousse-Pinguet, Y.; et al. Biogeography of Global Drylands. New Phytol. 2021, 231, 540–558. [Google Scholar] [CrossRef] [PubMed]
  6. Hachmi, A.; Andich, K.; Alaoui-Faris, F.E.E.; Mahyou, H. Amélioration de l’état de la végétation et de la fertilité des sols des parcours arides du Maroc par les techniques de restauration et de réhabilitation. Revue d’Écologie 2018, 73, 401. [Google Scholar] [CrossRef]
  7. Mihi, A.; Tarai, N.; Chenchouni, H. Can Palm Date Plantations and Oasification Be Used as a Proxy to Fight Sustainably against Desertification and Sand Encroachment in Hot Drylands? Ecol. Indic. 2019, 105, 365–375. [Google Scholar] [CrossRef]
  8. Hall, E.K.; Maixner, F.; Franklin, O.; Daims, H.; Richter, A.; Battin, T. Linking Microbial and Ecosystem Ecology Using Ecological Stoichiometry: A Synthesis of Conceptual and Empirical Approaches. Ecosystems 2011, 14, 261–273. [Google Scholar] [CrossRef]
  9. Sinsabaugh, R.L.; Hill, B.H.; Follstad Shah, J.J. Ecoenzymatic Stoichiometry of Microbial Organic Nutrient Acquisition in Soil and Sediment. Nature 2009, 462, 795–798. [Google Scholar] [CrossRef]
  10. Wang, Y.P.; Law, R.M.; Pak, B. A Global Model of Carbon, Nitrogen and Phosphorus Cycles for the Terrestrial Biosphere. Biogeosciences 2010, 7, 2261–2282. [Google Scholar] [CrossRef]
  11. Ma, D.; He, Z.; Ju, W.; Zhao, W.; Zhao, P.; Wang, W.; Lin, P. Long-Term Conventional Cultivation after Desert Reclamation Is Not Conducive to the Improvement of Soil Carbon Pool and Nutrient Stocks, a Case Study from Northwest China. Geoderma 2024, 445, 116893. [Google Scholar] [CrossRef]
  12. Liu, Z.; Cao, S.; Sun, Z.; Wang, H.; Qu, S.; Lei, N.; He, J.; Dong, Q. Tillage Effects on Soil Properties and Crop Yield after Land Reclamation. Sci. Rep. 2021, 11, 4611. [Google Scholar] [CrossRef] [PubMed]
  13. Man, M.; Wagner-Riddle, C.; Dunfield, K.E.; Deen, B.; Simpson, M.J. Long-Term Crop Rotation and Different Tillage Practices Alter Soil Organic Matter Composition and Degradation. Soil Tillage Res. 2021, 209, 104960. [Google Scholar] [CrossRef]
  14. Fallahzade, J.; Karimi, A.; Naderi, M.; Shirani, H. Soil Mechanical Properties and Wind Erosion Following Conversion of Desert to Irrigated Croplands in Central Iran. Soil Tillage Res. 2020, 204, 104665. [Google Scholar] [CrossRef]
  15. Shang, Z.H.; Cao, J.J.; Degen, A.A.; Zhang, D.W.; Long, R.J. A Four Year Study in a Desert Land Area on the Effect of Irrigated, Cultivated Land and Abandoned Cropland on Soil Biological, Chemical and Physical Properties. CATENA 2019, 175, 1–8. [Google Scholar] [CrossRef]
  16. Cao, Q.; Li, J.; Wang, G.; Wang, D.; Xin, Z.; Xiao, H.; Zhang, K. On the Spatial Variability and Influencing Factors of Soil Organic Carbon and Total Nitrogen Stocks in a Desert Oasis Ecotone of Northwestern China. CATENA 2021, 206, 105533. [Google Scholar] [CrossRef]
  17. Chen, L.; He, Z.; Zhao, W.; Liu, J.; Zhou, H.; Li, J.; Meng, Y.; Wang, L. Soil Structure and Nutrient Supply Drive Changes in Soil Microbial Communities during Conversion of Virgin Desert Soil to Irrigated Cropland. Eur. J. Soil Sci. 2020, 71, 768–781. [Google Scholar] [CrossRef]
  18. Li, F.-R.; Liu, J.-L.; Ren, W.; Liu, L.-L. Land-Use Change Alters Patterns of Soil Biodiversity in Arid Lands of Northwestern China. Plant Soil 2018, 428, 371–388. [Google Scholar] [CrossRef]
  19. Fan, H.; Pan, X.; Li, Y.; Chen, F.; Zhang, F. Evaluation of Soil Environment after Saline Soil Reclamation of Xinjiang Oasis, China. Agron. J. 2008, 100, 471–476. [Google Scholar] [CrossRef]
  20. Rani, J.; Paul, B. Challenges in Arid Region Reclamation with Special Reference to Indian Thar Desert—Its Conservation and Remediation Techniques. Int. J. Environ. Sci. Technol. 2023, 20, 12753–12774. [Google Scholar] [CrossRef]
  21. Su, Y.Z.; Yang, R.; Liu, W.J.; Wang, X.F. Evolution of Soil Structure and Fertility After Conversion of Native Sandy Desert Soil to Irrigated Cropland in Arid Region, China. Soil Sci. 2010, 175, 246. [Google Scholar] [CrossRef]
  22. Bacq-Labreuil, A.; Neal, A.L.; Crawford, J.; Mooney, S.J.; Akkari, E.; Zhang, X.; Clark, I.; Ritz, K. Significant Structural Evolution of a Long-term Fallow Soil in Response to Agricultural Management Practices Requires at Least 10 Years after Conversion. Eur. J. Soil Sci. 2021, 72, 829–841. [Google Scholar] [CrossRef]
  23. McBratney, A.B.; Mendonça Santos, M.L.; Minasny, B. On Digital Soil Mapping. Geoderma 2003, 117, 3–52. [Google Scholar] [CrossRef]
  24. Minasny, B.; McBratney, A.B. Digital Soil Mapping: A Brief History and Some Lessons. Geoderma 2016, 264, 301–311. [Google Scholar] [CrossRef]
  25. Poggio, L.; de Sousa, L.M.; Batjes, N.H.; Heuvelink, G.B.M.; Kempen, B.; Ribeiro, E.; Rossiter, D. SoilGrids 2.0: Producing Soil Information for the Globe with Quantified Spatial Uncertainty. SOIL 2021, 7, 217–240. [Google Scholar] [CrossRef]
  26. Arrouays, D.; McBratney, A.; Bouma, J.; Libohova, Z.; Richer-de-Forges, A.C.; Morgan, C.L.S.; Roudier, P.; Poggio, L.; Mulder, V.L. Impressions of Digital Soil Maps: The Good, the Not so Good, and Making Them Ever Better. Geoderma Reg. 2020, 20, e00255. [Google Scholar] [CrossRef]
  27. Heuvelink, G.B.M.; Webster, R. Spatial Statistics and Soil Mapping: A Blossoming Partnership under Pressure. Spat. Stat. 2022, 50, 100639. [Google Scholar] [CrossRef]
  28. Yimer, F.; Ledin, S.; Abdelkadir, A. Soil Property Variations in Relation to Topographic Aspect and Vegetation Community in the South-Eastern Highlands of Ethiopia. For. Ecol. Manag. 2006, 232, 90–99. [Google Scholar] [CrossRef]
  29. McGuire, K.J.; McDonnell, J.J.; Weiler, M.; Kendall, C.; McGlynn, B.L.; Welker, J.M.; Seibert, J. The Role of Topography on Catchment-Scale Water Residence Time. Water Resour. Res. 2005, 41, W05002. [Google Scholar] [CrossRef]
  30. Wiesmeier, M.; Barthold, F.; Blank, B.; Kögel-Knabner, I. Digital Mapping of Soil Organic Matter Stocks Using Random Forest Modeling in a Semi-Arid Steppe Ecosystem. Plant Soil 2011, 340, 7–24. [Google Scholar] [CrossRef]
  31. Zeraatpisheh, M.; Ayoubi, S.; Jafari, A.; Tajik, S.; Finke, P. Digital Mapping of Soil Properties Using Multiple Machine Learning in a Semi-Arid Region, Central Iran. Geoderma 2019, 338, 445–452. [Google Scholar] [CrossRef]
  32. Azamat, S.; Ilgiz, A.; Ruslan, S.; Mirsayapov, R.; Ilyusya, G.; Tuktarova, I.; Belan, L. Assessing and Mapping of Soil Organic Carbon at Multiple Depths in the Semi-Arid Trans-Ural Steppe Zone. Geoderma Reg. 2024, 38, e00855. [Google Scholar] [CrossRef]
  33. Poggio, L.; Gimona, A. National Scale 3D Modelling of Soil Organic Carbon Stocks with Uncertainty Propagation—An Example from Scotland. Geoderma 2014, 232–234, 284–299. [Google Scholar] [CrossRef]
  34. Jenny, H. Factors of Soil Formation: A System of Quantitative Pedology; McGraw-Hill Book Company, Inc.: New York, NY, USA, 1941. [Google Scholar]
  35. Bishop, T.F.A.; McBratney, A.B.; Laslett, G.M. Modelling Soil Attribute Depth Functions with Equal-Area Quadratic Smoothing Splines. Geoderma 1999, 91, 27–45. [Google Scholar] [CrossRef]
  36. Minasny, B. Prediction and Digital Mapping of Soil Carbon Storage in the Lower Namoi Valley. Available online: https://www.publish.csiro.au/sr/SR05136 (accessed on 19 September 2024).
  37. Zhang, Y.; Biswas, A.; Adamchuk, V.I. Implementation of a Sigmoid Depth Function to Describe Change of Soil pH with Depth. Geoderma 2017, 289, 1–10. [Google Scholar] [CrossRef]
  38. Allory, V.; Séré, G.; Ouvrard, S. A Meta-Analysis of Carbon Content and Stocks in Technosols and Identification of the Main Governing Factors. Eur. J. Soil Sci. 2022, 73, e13141. [Google Scholar] [CrossRef]
  39. Fu, P.; Clanton, C.; Demuth, K.M.; Goodman, V.; Griffith, L.; Khim-Young, M.; Maddalena, J.; LaMarca, K.; Wright, L.A.; Schurman, D.W.; et al. Accurate Quantification of 0–30 Cm Soil Organic Carbon in Croplands over the Continental United States Using Machine Learning. Remote Sens. 2024, 16, 2217. [Google Scholar] [CrossRef]
  40. Wang, X.; Qi, J.-Y.; Zhang, X.-Z.; Li, S.-S.; Latif Virk, A.; Zhao, X.; Xiao, X.-P.; Zhang, H.-L. Effects of Tillage and Residue Management on Soil Aggregates and Associated Carbon Storage in a Double Paddy Cropping System. Soil Tillage Res. 2019, 194, 104339. [Google Scholar] [CrossRef]
  41. Walk, J.; Schulte, P.; Bartz, M.; Binnie, A.; Kehl, M.; Mörchen, R.; Sun, X.; Stauch, G.; Tittmann, C.; Bol, R.; et al. Pedogenesis at the Coastal Arid-Hyperarid Transition Deduced from a Late Quaternary Chronosequence at Paposo, Atacama Desert. CATENA 2023, 228, 107171. [Google Scholar] [CrossRef]
  42. Bao, S. Soil and Agricultural Chemistry Analysis; China Agricultural Press: Beijing, China, 2000. [Google Scholar]
  43. Hassani, A.; Azapagic, A.; Shokri, N. Global Predictions of Primary Soil Salinization under Changing Climate in the 21st Century. Nat. Commun. 2021, 12, 6663. [Google Scholar] [CrossRef]
  44. Singh, A. Soil Salinization Management for Sustainable Development: A Review. J. Environ. Manag. 2021, 277, 111383. [Google Scholar] [CrossRef] [PubMed]
  45. ASF Data Search. Available online: https://search.asf.alaska.edu/#/ (accessed on 6 October 2024).
  46. Wang, F.; Shi, Z.; Biswas, A.; Yang, S.; Ding, J. Multi-Algorithm Comparison for Predicting Soil Salinity. Geoderma 2020, 365, 114211. [Google Scholar] [CrossRef]
  47. Taghizadeh-Mehrjardi, R.; Minasny, B.; Sarmadian, F.; Malone, B.P. Digital Mapping of Soil Salinity in Ardakan Region, Central Iran. Geoderma 2014, 213, 15–28. [Google Scholar] [CrossRef]
  48. Khan, N.M.; Rastoskuev, V.V.; Sato, Y.; Shiozawa, S. Assessment of Hydrosaline Land Degradation by Using a Simple Approach of Remote Sensing Indicators. Agric. Water Manag. 2005, 77, 96–109. [Google Scholar] [CrossRef]
  49. Barrena-González, J.; Gabourel-Landaverde, V.A.; Mora, J.; Contador, J.F.L.; Fernández, M.P. Exploring Soil Property Spatial Patterns in a Small Grazed Catchment Using Machine Learning. Earth Sci. Inform. 2023, 16, 3811–3838. [Google Scholar] [CrossRef]
  50. Myles, A.J.; Feudale, R.N.; Liu, Y.; Woody, N.A.; Brown, S.D. An Introduction to Decision Tree Modeling. J. Chemom. 2004, 18, 275–285. [Google Scholar] [CrossRef]
  51. Meinshausen, N. Quantile Regression Forests. J. Mach. Learn. Res. 2006, 7, 983–999. [Google Scholar]
  52. Draper, N.R.; Smith, H. Applied Regression Analysis; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 1998. [Google Scholar]
  53. Mbagwu, J.S.C.; Abeh, O.G. Prediction of engineering properties of tropical soils using intrinsic pedological parameters. Soil Sci. 1998, 163, 93. [Google Scholar] [CrossRef]
  54. Carré, F.; McBratney, A.B.; Mayr, T.; Montanarella, L. Digital soil assessments: Beyond DSM. Geoderma 2007, 142, 69–79. [Google Scholar] [CrossRef]
  55. Genuer, R.; Poggi, J.-M.; Tuleau-Malot, C. Variable Selection Using Random Forests. Pattern Recognit. Lett. 2010, 31, 2225–2236. [Google Scholar] [CrossRef]
  56. Nauman, T.W.; Thompson, J.A. Semi-Automated Disaggregation of Conventional Soil Maps Using Knowledge Driven Data Mining and Classification Trees. Geoderma 2014, 213, 385–399. [Google Scholar] [CrossRef]
  57. Philippot, L.; Chenu, C.; Kappler, A.; Rillig, M.C.; Fierer, N. The Interplay between Microbial Communities and Soil Properties. Nat. Rev. Microbiol. 2024, 22, 226–239. [Google Scholar] [CrossRef]
  58. Muhetaer, N.; Nurmemet, I.; Abulaiti, A.; Xiao, S.; Zhao, J. A Quantifying Approach to Soil Salinity Based on a Radar Feature Space Model Using ALOS PALSAR-2 Data. Remote Sens. 2022, 14, 363. [Google Scholar] [CrossRef]
  59. Gamalero, E.; Bona, E.; Todeschini, V.; Lingua, G. Saline and Arid Soils: Impact on Bacteria, Plants, and Their Interaction. Biology 2020, 9, 116. [Google Scholar] [CrossRef]
  60. Zhang, F.; Chen, X.; Yao, S.; Ye, Y.; Zhang, B. Responses of Soil Mineral-Associated and Particulate Organic Carbon to Carbon Input: A Meta-Analysis. Sci. Total Environ. 2022, 829, 154626. [Google Scholar] [CrossRef]
  61. Li, X.; Xie, J.; Zhang, Q.; Lyu, M.; Xiong, X.; Liu, X.; Lin, T.; Yang, Y. Substrate Availability and Soil Microbes Drive Temperature Sensitivity of Soil Organic Carbon Mineralization to Warming along an Elevation Gradient in Subtropical Asia. Geoderma 2020, 364, 114198. [Google Scholar] [CrossRef]
  62. Wang, C.; Morrissey, E.M.; Mau, R.L.; Hayer, M.; Piñeiro, J.; Mack, M.C.; Marks, J.C.; Bell, S.L.; Miller, S.N.; Schwartz, E.; et al. The Temperature Sensitivity of Soil: Microbial Biodiversity, Growth, and Carbon Mineralization. ISME J. 2021, 15, 2738–2747. [Google Scholar] [CrossRef]
  63. Yu, H.; Sui, Y.; Chen, Y.; Bao, T.; Jiao, X. Soil Organic Carbon Mineralization and Its Temperature Sensitivity under Different Substrate Levels in the Mollisols of Northeast China. Life 2022, 12, 712. [Google Scholar] [CrossRef] [PubMed]
  64. Blanco-Canqui, H.; Ruis, S.J. No-Tillage and Soil Physical Environment. Geoderma 2018, 326, 164–200. [Google Scholar] [CrossRef]
  65. Song, W.; Li, J.; Li, X.; Xu, D.; Min, X. Effects of Land Reclamation on Soil Organic Carbon and Its Components in Reclaimed Coal Mining Subsidence Areas. Sci. Total Environ. 2024, 908, 168523. [Google Scholar] [CrossRef]
  66. Peng, L.; Wan, Y.-B.; Li, H.; Du, M.-D.; Shi, Q.-D. Influence of Surface Water and Groundwater Gradient on Spatial Distribution of Typical Vegetation in the Hinterland of Taklamakan Desert. Sci. Total Environ. 2024, 953, 176060. [Google Scholar] [CrossRef] [PubMed]
  67. Zhang, Z.; Dong, X.; Wang, S.; Pu, X. Benefits of Organic Manure Combined with Biochar Amendments to Cotton Root Growth and Yield under Continuous Cropping Systems in Xinjiang, China. Sci. Rep. 2020, 10, 4718. [Google Scholar] [CrossRef] [PubMed]
  68. Kang, J.; Liu, L.; Zhang, F.; Shen, C.; Wang, N.; Shao, L. Semantic Segmentation Model of Cotton Roots In-Situ Image Based on Attention Mechanism. Comput. Electron. Agric. 2021, 189, 106370. [Google Scholar] [CrossRef]
  69. Shao, W.; Wang, Q.; Guan, Q.; Luo, H.; Ma, Y.; Zhang, J. Distribution of Soil Available Nutrients and Their Response to Environmental Factors Based on Path Analysis Model in Arid and Semi-Arid Area of Northwest China. Sci. Total Environ. 2022, 827, 154254. [Google Scholar] [CrossRef]
  70. Peng, X.; Maharjan, B.; Yu, C.; Su, A.; Jin, V.; Ferguson, R.B. A Laboratory Evaluation of Ammonia Volatilization and Nitrate Leaching Following Nitrogen Fertilizer Application on a Coarse-Textured Soil. Agron. J. 2015, 107, 871–879. [Google Scholar] [CrossRef]
  71. Vinten, A.J.A.; Vivian, B.J.; Wright, F.; Howard, R.S. A Comparative Study of Nitrate Leaching from Soils of Differing Textures under Similar Climatic and Cropping Conditions. J. Hydrol. 1994, 159, 197–213. [Google Scholar] [CrossRef]
  72. Shareef, M.; Gui, D.; Zeng, F.; Waqas, M.; Ahmed, Z.; Zhang, B.; Iqbal, H.; Xue, J. Nitrogen Leaching, Recovery Efficiency, and Cotton Productivity Assessments on Desert-Sandy Soil under Various Application Methods. Agric. Water Manag. 2019, 223, 105716. [Google Scholar] [CrossRef]
  73. Geng, Y.; Pan, S.; Zhang, L.; Qiu, J.; He, K.; Gao, H.; Li, Z.; Tian, D. Phosphorus Biogeochemistry Regulated by Carbonates in Soil. Environ. Res. 2022, 214, 113894. [Google Scholar] [CrossRef]
  74. Luo, D.; Wang, L.; Nan, H.; Cao, Y.; Wang, H.; Kumar, T.V.; Wang, C. Phosphorus Adsorption by Functionalized Biochar: A Review. Environ. Chem. Lett. 2023, 21, 497–524. [Google Scholar] [CrossRef]
  75. Yahaya, S.M.; Mahmud, A.A.; Abdullahi, M.; Haruna, A. Recent Advances in the Chemistry of Nitrogen, Phosphorus and Potassium as Fertilizers in Soil: A Review. Pedosphere 2023, 33, 385–406. [Google Scholar] [CrossRef]
  76. Murphy, B.W.; Wilson, B.R.; Koen, T. Mathematical Functions to Model the Depth Distribution of Soil Organic Carbon in a Range of Soils from New South Wales, Australia under Different Land Uses. Soil Syst. 2019, 3, 46. [Google Scholar] [CrossRef]
  77. Emiru, N.; Gebrekidan, H. Effect of Land Use Changes and Soil Depth on Soil Organic Matter, Total Nitrogen and Available Phosphorus Contents of Soils in Senbat Watershed, Western Ethiopia. Am. J. Agric. Biol. Sci. 2013, 8, 206–212. [Google Scholar]
  78. Tian, D.; Niu, S. A Global Analysis of Soil Acidification Caused by Nitrogen Addition. Environ. Res. Lett. 2015, 10, 024019. [Google Scholar] [CrossRef]
  79. Liang, F.; Li, B.; Vogt, R.D.; Mulder, J.; Song, H.; Chen, J.; Guo, J. Straw Return Exacerbates Soil Acidification in Major Chinese Croplands. Resour. Conserv. Recycl. 2023, 198, 107176. [Google Scholar] [CrossRef]
  80. Meng, C.; Tian, D.; Zeng, H.; Li, Z.; Yi, C.; Niu, S. Global Soil Acidification Impacts on Belowground Processes. Environ. Res. Lett. 2019, 14, 074003. [Google Scholar] [CrossRef]
  81. Youssef, A.M.; Pourghasemi, H.R.; Pourtaghi, Z.S.; Al-Katheeri, M.M. Landslide Susceptibility Mapping Using Random Forest, Boosted Regression Tree, Classification and Regression Tree, and General Linear Models and Comparison of Their Performance at Wadi Tayyah Basin, Asir Region, Saudi Arabia. Landslides 2016, 13, 839–856. [Google Scholar] [CrossRef]
  82. Yang, R.-M.; Zhang, G.-L.; Liu, F.; Lu, Y.-Y.; Yang, F.; Yang, F.; Yang, M.; Zhao, Y.-G.; Li, D.-C. Comparison of Boosted Regression Tree and Random Forest Models for Mapping Topsoil Organic Carbon Concentration in an Alpine Ecosystem. Ecol. Indic. 2016, 60, 870–878. [Google Scholar] [CrossRef]
  83. Pahlavan-Rad, M.R.; Dahmardeh, K.; Hadizadeh, M.; Keykha, G.; Mohammadnia, N.; Gangali, M.; Keikha, M.; Davatgar, N.; Brungard, C. Prediction of Soil Water Infiltration Using Multiple Linear Regression and Random Forest in a Dry Flood Plain, Eastern Iran. CATENA 2020, 194, 104715. [Google Scholar] [CrossRef]
  84. da Silva Chagas, C.; de Carvalho Junior, W.; Bhering, S.B.; Calderano Filho, B. Spatial Prediction of Soil Surface Texture in a Semiarid Region Using Random Forest and Multiple Linear Regressions. CATENA 2016, 139, 232–240. [Google Scholar] [CrossRef]
  85. Liu, X.; Du, H.; Li, S.; Wang, T.; Fan, Y. Effects of Different Cropland Reclamation Periods on Soil Particle Size and Nutrients From the Perspective of Wind Erosion in the Mu Us Sandy Land. Front. Environ. Sci. 2022, 10, 861273. [Google Scholar] [CrossRef]
Sustainability 17 04194 g001
Figure 1. Spatial distribution of borehole sampling sites in the study area.
Figure 1. Spatial distribution of borehole sampling sites in the study area.
Sustainability 17 04194 g001
Sustainability 17 04194 g002
Figure 2. Correlations among various soil properties.
Figure 2. Correlations among various soil properties.
Sustainability 17 04194 g002
Sustainability 17 04194 g003
Figure 3. Fitted vertical distributions of SOC, pH, NH4+, and TP. The red circles indicate the mean values of SOC, pH, NH4+, and TP in each layer. All the presented formulas are inverse functions.
Figure 3. Fitted vertical distributions of SOC, pH, NH4+, and TP. The red circles indicate the mean values of SOC, pH, NH4+, and TP in each layer. All the presented formulas are inverse functions.
Sustainability 17 04194 g003
Sustainability 17 04194 g004
Figure 4. Histograms of the SOC, pH, NH4+, and TP values in the experimental group (TG) minus the control group (CK).
Figure 4. Histograms of the SOC, pH, NH4+, and TP values in the experimental group (TG) minus the control group (CK).
Sustainability 17 04194 g004
Sustainability 17 04194 g005
Figure 5. Spatial distribution of SOC predicted by the QRF model at different soil depths.
Figure 5. Spatial distribution of SOC predicted by the QRF model at different soil depths.
Sustainability 17 04194 g005
Sustainability 17 04194 g006
Figure 6. Spatial distribution of pH predicted by the QRF model at different soil depths.
Figure 6. Spatial distribution of pH predicted by the QRF model at different soil depths.
Sustainability 17 04194 g006
Sustainability 17 04194 g007
Figure 7. Spatial distribution of NH4+ predicted by the QRF model at different soil depths.
Figure 7. Spatial distribution of NH4+ predicted by the QRF model at different soil depths.
Sustainability 17 04194 g007
Sustainability 17 04194 g008
Figure 8. Spatial distribution of TP predicted by the QRF model at different soil depths.
Figure 8. Spatial distribution of TP predicted by the QRF model at different soil depths.
Sustainability 17 04194 g008
Table 1. Overview of crop management practices in the study area.
Table 1. Overview of crop management practices in the study area.
Crop TypeIrrigation MethodIrrigation VolumeWater SourceFertilizer ApplicationYield
m3/hakg/hakg/ha
CottonSubsurface drip irrigation3900–4800Reservoir (glacial meltwater)750–1200 (N: 46.4%)6000–7800
Winter–Spring flooding2250–3000Groundwater450–600 (K2O: 34.2%)
JujubeDrip tape irrigation4800–6000Reservoir (glacial meltwater)384 (N: 46.4%)27,000–33,000
Winter–Spring flooding1050–1350Groundwater704 (P2O5: 60.5%)
Table 2. Coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE) values of various depth functions for different soil properties.
Table 2. Coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE) values of various depth functions for different soil properties.
Soil PropertyType of Depth FunctionR2RMSEMAE
SOC
(g/kg)		A x ek x depth0.870.540.45
B + A x ek x depth0.900.460.39
A + k x ln(-depth + 1)0.840.550.46
pHA + B x depth3 + C x depth2 + D x depth0.920.070.05
A x ek x depth0.790.100.07
A + k x ln(-depth + 1)0.770.100.08
NH4+
(mg/kg)		A x (-depth)k0.201.120.82
A + k x ln(-depth + 1)0.570.770.45
B + A x ek x depth0.640.650.40
A x ek x depth0.560.780.45
A + B x depth3 + C x depth2 + D x depth0.890.380.25
TP
(g/kg)		A + k x ln(-depth + 1)0.670.150.09
B + A x ek x depth0.770.120.08
A x ek x depth0.700.140.08
Table 3. Validation results for the prediction of SOC, pH, NH4+, and TP at depths of 0–10 cm via the quantile regression forest (QRF), boosting regression tree (BRT), and multiple linear regression (MLR) models.
Table 3. Validation results for the prediction of SOC, pH, NH4+, and TP at depths of 0–10 cm via the quantile regression forest (QRF), boosting regression tree (BRT), and multiple linear regression (MLR) models.
Soil PropertyType of ModelR2RMSEMAE
SOC
(g/kg)		QRF0.780.620.51
BRT0.720.610.47
MLR0.131.361.04
pHQRF0.790.100.07
BRT0.760.100.07
MLR0.210.210.17
NH4+
(mg/kg)		QRF0.780.380.28
BRT0.730.400.29
MLR0.190.800.64
TP
(g/kg)		QRF0.710.120.08
BRT0.620.140.09
MLR0.120.260.17
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Xiao, Y.; Ye, M.; Zhang, J.; Chen, Y.; Sun, X.; Li, X.; Song, X. Significant Changes in Soil Properties in Arid Regions Due to Semicentennial Tillage—A Case Study of Tarim River Oasis, China. Sustainability 2025, 17, 4194. https://doi.org/10.3390/su17094194

AMA Style

Xiao Y, Ye M, Zhang J, Chen Y, Sun X, Li X, Song X. Significant Changes in Soil Properties in Arid Regions Due to Semicentennial Tillage—A Case Study of Tarim River Oasis, China. Sustainability. 2025; 17(9):4194. https://doi.org/10.3390/su17094194

Chicago/Turabian Style

Xiao, Ying, Mingliang Ye, Jing Zhang, Yamin Chen, Xinxin Sun, Xiaoyan Li, and Xiaodong Song. 2025. "Significant Changes in Soil Properties in Arid Regions Due to Semicentennial Tillage—A Case Study of Tarim River Oasis, China" Sustainability 17, no. 9: 4194. https://doi.org/10.3390/su17094194

APA Style

Xiao, Y., Ye, M., Zhang, J., Chen, Y., Sun, X., Li, X., & Song, X. (2025). Significant Changes in Soil Properties in Arid Regions Due to Semicentennial Tillage—A Case Study of Tarim River Oasis, China. Sustainability, 17(9), 4194. https://doi.org/10.3390/su17094194

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop