Next Article in Journal
Fatty Acid Profile and Oxidative Stability of Layers’ Egg Yolk as Affected by Dietary Supplementation with Fresh Purslane and Addition of Aromatic Plant Essential Oils to Drinking Water
Previous Article in Journal
Dynamic Evaluation of Ecological Environment Quality in Coastal Cities from the Perspective of Water Quality: The Case of Fuzhou City
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 15
Issue 15
10.3390/su151511538
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 thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Analyzing Driving Factors of Soil Alkalinization Based on Geodetector—A Case in Northeast China

by
Lin Bai
1,*,
Jia Zhou
2,
Jinming Luo
1,
Hongshuang Dou
1 and
Ye Zhang
1
1
Department of Geography, Qiqihar University, Qiqihar 161006, China
2
Department of Geography, Harbin Normal University, Harbin 150025, China
*
Author to whom correspondence should be addressed.
Sustainability 2023, 15(15), 11538; https://0-doi-org.brum.beds.ac.uk/10.3390/su151511538
Submission received: 8 June 2023 / Revised: 19 July 2023 / Accepted: 20 July 2023 / Published: 26 July 2023
(This article belongs to the Section Social Ecology and Sustainability)
Browse Figures
Review Reports Versions Notes

Abstract

:
The Songnen Plain of Northeast China is one of the three largest soda saline–alkaline regions in the world. To better understand soil alkalinization in this important agricultural region of China, it is vital to explore the driving factors of soil alkalinity. Combined with prior research on the Wuyu’er–Shuangyang River Basin, this study examined the driving factors of soil alkalinity using the Geodetector method. First, the analysis results of the risk detector, the factor detector, and the ecological detector revealed the primary driving factors of soil alkalinity in the study area. Next, the analysis results of the interaction detector determined how combinations of driving factors impacted soil alkalinity in the study area. In general, the natural driving factors of altitude and spring temperature, especially altitude, played a key role in soil alkalinization. These results indicated that the closed terrain and warming trends were the main causes of soil alkalinization in the study area. In addition, there were significant enhance-nonlinear and enhance-bivariate relationships among the driving factors, which indicated that combined driving factors contributed more to soil alkalinization than individual driving factors.

1. Introduction

Soil salinization is a manifestation of soil degradation in arid and semiarid regions [1], which has a great effect on the ecological environment and agricultural productivity. Soil salinization is widely dispersed, and almost 20% of the world’s land is currently affected [2]. Therefore, soil salinization has attracted worldwide attention and has remained a popular issue in soil research [3]. Generally, soil salinization occurs when there is an increased concentration of soluble salts such as Na+, K+, Ca2+, Mg2+, and Cl−. These salts are abundantly redistributed to the soil surface of low-lying land in the horizontal and vertical directions. Soil alkalinization involves the enrichment of Na+ [4], HCO3−, and CO32− in the soil. When sodium ions are the main salts of the soil, the alkalinization process is also called sodification [5]. Saline and alkaline soils can be classified according to their electrical conductivity (EC), potential of hydrogen (pH), and exchangeable sodium percentage (ESP) of the soil saturation extract [6]. Saline soil refers to soil for which the EC is more than 4 mS/cm, ESP is less than 15, and pH is less than 8.5, while alkaline soil refers to soil for which the ESP is greater than 15, EC is less than 4 mS/cm, and pH is greater than 8.5; therefore, soil for which the EC is more than 4 mS/cm and ESP is greater than 15 is saline–alkaline soil [5]. Being a simple and convenient measurement of the acid or alkali levels in soil [6], soil pH can be considered a direct indicator of soil health [7,8].
The Songnen Plain in northeastern China is one of the three largest soda saline–alkaline [9] and black soil regions [10] in the world. This region is an important grain-producing area of China [11]. Soil productivity varies significantly by soil properties and fertility, which are influenced by soil salinization and alkalinization. Soil salinization can form white crusts of salts on the surface of soil and hinder the growth of plants by limiting their ability to take up water [5]. Soil acidity or alkalinity can directly affect plant growth, and most soil nutrients that plants need are readily available when the soil pH ranges from 6.0 to 7.5 [8]. Compared with salinity, the development of alkalinity and sodicity in soils can increase the soil pH and destroy the organic matter in black soil in a region. A decrease in the soil nutrient level further declines productivity in croplands and pasturelands [12]. Therefore, more attention should be paid to the Songnen Plain and other similar areas.
As soil pH can be used to describe soil acidity or alkalinity directly, how to acquire the spatial distribution of soil pH in a certain area is an important concern. Kumar et al. [8] used soil pH to describe the degrees of soil acidity using the digital image processing technique. Their findings provided a clue for the spectral signature capture of different pH values in acidic soil in the red, green, and blue wavelength regions. Zhang et al. [13] investigated laboratory-measured and field-measured spectra of alkaline soil in Xinjiang. The findings showed that there was a significant positive correlation between several field-measured spectral bands from 450 nm to 900 nm and the soil pH of alkaline soil. These studies suggested the probability of mapping the soil pH of alkaline soil and non-alkaline soil through the inversion of remote sensing images. Zabcic et al. [14] reported regional-scale soil pH maps generated from an airborne hyperspectral image in Spain. Bai et al. acquired a soil pH map of alkaline soil and non-alkaline soil via the inversion of multispectral images [15] and hyperspectral images [16] in the Wuyu’er–Shuangyang River Basin, China.
Natural and human factors are the main causes of salinization [17]. Climate is an important natural factor that can lead to salt accumulation on a soil surface [18,19,20]. Low rainfall [1], high temperature [5,17], and high evaporation [5,21] in arid regions result in these regions being the most prone to soil salinization and alkalinization. The freeze–thaw action, which is a particular mechanism, also has obvious control over soil salinization and alkalinization [5,22]. In the context of global warming and an increase in aridity, climate conditions may contribute to soil salinization within the time span of several years to several decades [23]. According to Wang et al. [5], soluble salts were accumulated on the soil surface and caused soil salinization and alkalinization, especially in spring and autumn, in the Songnen Plain, China, because of strong evaporation and rare precipitation. The water table plays a key role in soil salinization in some farmlands [24,25,26]. Soil parent materials, such as alkaline rocks, have direct impacts on soil salinization and alkalinization [27]. A low and flat terrain, which hinders the outflow of salts to the basin, can favor soil salinization and alkalization [5,28]. Tectonic movement in northeastern China, which shaped the soil parent materials and terrain that favor soil salinization and alkalinization, dates from the Jurassic period [27,29]. Across the world, human factors, including population growth [30], farming and planting [21,31], and irrigation in an oasis and along a coast [25,32,33], have caused soil salt accumulation to some extent in the past several decades. Therefore, soil salinization and alkalinization are involved in complex and interactive physical, chemical [34], and social processes. It is vital to explore and assess natural and human factors for saline–alkaline soil control and recovery.
Statistical methods have been used to assess factors influencing soil salinization [35]. For example, principal component analysis was used to develop three soil salinity models with five salinity indices and 11 environmental variables for the monitoring of salinity-affected areas in southwestern Bangladesh [36]. However, the factors considered in these models were relatively simple and could not quantify the interaction of the impact factors. According to Mohamed et al. [37], geostatistical analysis is a rapid and reliable method for obtaining information on the spatial distribution of salinity. However, the driving factors of such spatial distribution of salinity were not acquired in their study. Therefore, it is vital to explore the driving factors of soil alkalinity.
Geodetector is a spatial analysis method proposed by Wang et al. [38] in 2010 to quantitatively explore the relationship between an endemic disease or a potential health risk and its driving factors [39,40,41,42].
Over the ensuing decade, Geodetector was adopted to perform exploration of the spatial distribution of a certain environmental risk or a geographical phenomenon, along with identification of the driving factors. For example, Wei et al. [43] applied this method to data on different types of wetlands in the Yellow River delta acquired from Sentinel-2 images to examine the change process of wetlands and its driving factors. In the study by Zhang et al. [44], the Geodetector method was used to explore the spatiotemporal dynamics of vegetation and the interaction between environmental driving factors in the Qinba Mountains, China. Geodetector, assisted by a geographically weighted regression (GWR) model, was also used to characterize the spatial distribution of drought and identify the influencing factors of drought in the Inner Mongolia Autonomous Region, northern China [45]. In the study by Ma et al. [22], environmental factors influencing the spatial distribution of the ground surface freezing index (GFI) and the ground surface thawing index (GTI) were detected using the Geodetector method. Yang et al. [46] also applied Geodetector to detect the specific relationship between urban vitality and urban carbon emissions. He et al. [47] combined InVEST and Geodetector models to estimate Nr losses in the Taihu Lake Basin from 1990 to 2020, as well as explore the driving forces. Recent attempts have been made to quantitatively explore the driving factors of soil salinity. In the study by Su et al. [48], the salinization degree, spatial distribution, and drivers, as well as their quantitative relationship in Zhuanghe, were analyzed using Geodetector and other methods. The results of Geodetector suggested the primary driving factor was groundwater depth (EP = 0.90).
The aim of this study was to investigate the driving factors of soil alkalinization and their interaction using Geodetector. The Wuyu’er–Shuangyang River Basin in northeast China, which is a typical representative saline–alkaline soil area in western Songnen Plain, was chosen as a case study. On the basis of prior research on alkaline soil distribution in this area [15], driving factors such as meteorological data, groundwater, altitude, and population were all considered in this study. The distribution of alkaline soil and driving factors were analyzed using Geodetector. The sorting order of the driving factors, as well as the interaction between these driving factors, was determined.

2. Materials and Methods

2.1. Study Area

The Wuyu’er–Shuangyang River Basin is located in the north part of western Songnen Plain, Northeast China, covering an area of 23,000 km2, with a spatial range of 124°–127° E and 46°–48° N (Figure 1). In this basin, the climate is temperate continental monsoon, with annual average precipitation and temperature of approximately 415 mm and 3.2 °C, respectively. The terrain is low and flat, decreasing from the northeast to the southwest, with a mean altitude of about 207 m. The Wuyu’er River and Shuangyang River, which are two inner rivers in the closed area, flow into the Zhalong wetland in the south. Alkalinization occurs in the lower landscape positions between the margins of the Zhalong wetland and downstream rivers, where groundwater is relatively low, and the soil is rich in Na+ [4], HCO3−, and CO32− [15]. Grasslands and paddy fields are distributed near the rivers and the wetland. The soil types in the study area are mainly Phaeozems and Chernozems, which are beneficial for agricultural cultivation. There are several towns—Fuyu, Yi’an, Baiquan, Keshan, and Bei’an—in the administrative area.

2.2. Datasets and Preprocessing

2.2.1. Driving Factors

The driving factors were divided into two categories in this study: natural factors and social factors. Natural factors had three types and six driving factors (Table 1). The four climate driving factors were mean temperature, mean 0 cm ground temperature, total amount of precipitation, and total amount of evaporation in spring. The groundwater driving factor was groundwater depth in spring. The topographic driving factor was altitude, which was represented by a digital elevation model (DEM). Social factors had two types and two driving factors. The population driving factor was population density, and the land-use driving factor was farmland percentage (Table 1).
Climate data were obtained from 14 meteorological observation stations belonging to the Qiqihar Meteorological Bureau and Daqing Meteorological Bureau (Figure 2a). Groundwater data were obtained from 74 hydrological observation stations belonging to the Qiqihar Hydrological Bureau and included in situ sample points across all stations (Figure 2b). The acquisition dates of the climate driving factors and groundwater depth were also close to those of the Operational Land Imager (OLI) image acquisition, which provided the data on soil alkalinity used in this study. Thus, the mean temperature and the mean 0 cm ground temperature in spring were the average values of the temperature and the 0 cm ground temperature in March and April 2014, respectively. The total amount of precipitation and the total amount of evaporation in spring were the sum values of precipitation and evaporation, respectively, in March and April 2014. The altitude data (DEM) with a spatial resolution of 90 m were obtained from the Resource and Environment Science and Data Center (https://www.resdc.cn/).
The population and farmland quantity in the townships of the study area were obtained from the China Statistical Yearbook 2014. There are 89 townships in the study area (Figure 3). In order to compare the population and farmland quantity of townships of different sizes, the data were recalculated as the population density and farmland percentage of every township.

2.2.2. Soil pH

Using data collected in a prior research study, the physicochemical measurements of soil samples collected in the period from October 2013 to May 2015 were used to examine soil alkalinization in the study area [15]. A stepwise regression model was built with soil pH and the corresponding lab spectra to estimate soil pH [15]. Using the stepwise regression model, the spatial distribution of soil pH in spring was acquired from the OLI images (17 April, 30 April, and 3 May 2014), and the results are shown in Figure 4. The estimated soil pH values shown in Figure 4 were used to represent different levels of soil alkalinization in this study. As described in Song [49], the levels of soil alkalinity and pH can be divided into strongly alkaline (pH > 9.5), moderately alkaline (9.5 > pH > 8.5), slightly alkaline (8.5 > pH > 7.5), and non-alkaline (pH < 7.5). As shown in Figure 4, alkaline soils (slightly, moderately, and strongly) accounted for 44.46% of the soils in the basin and in the lower margins of the Zhalong wetland and downstream rivers. Therefore, soil alkalinization is a common geographical phenomenon in the study area.

2.2.3. Discretization

The climate data and groundwater depth were observed values obtained from scattered observation stations. The continuous spatial distributions of the climate data and groundwater depth across the basin were acquired using spatial interpolation (Kriging). Such continuous data need to be converted into discrete features which can be detected by Geodetector [50].
Discretization is a process of converting continuous variables into intervals, which are then handled as categorical data. Continuous variables should be discretized before classification [51]. Overall, discretization simplifies continuous data, accelerates the learning process, and acquires more accurate results. The first step of discretization is identifying the thresholds that define the intervals into which the continuous variables under investigation can be discretized [52]. There are various discretization methods, each with advantages and disadvantages [51,52]. The discretization methods based on equal interval, quantile, natural breakpoint, and geometrical interval can be applied in ArcGIS 10.1 software.

2.3. Methodological Approach

2.3.1. Exploratory Spatial Data Analysis

Exploratory spatial data analysis (ESDA) is a method of describing and visualizing spatial distributions of geographic data [53,54]. It can identify typical locations or spatial outliers using a Voronoi map. It discovers patterns of spatial association using a semi-variogram or covariance function graph. It explores if there is a trend in a certain direction using a trend distribution chart. ESDA is a prerequisite for applying Kriging interpolation.

2.3.2. Kriging Interpolation

Kriging interpolation is an important geostatistical method that generates an estimated surface from a scattered set of points with a certain feature [55,56]. Unlike other interpolation methods, Kriging interpolation is an effective tool that can acquire linear optimal and unbiased interpolation by estimating the spatial distribution of a geographical variable. The estimation of Kriging interpolation particularly focuses on the spatial autocorrelation characteristics of geographical variables and takes the values of the same variable in the neighborhood into account.

2.3.3. Geodetector

Geodetector is a spatial analysis method that explores the spatial differences of an environmental risk or a geographical phenomenon and detects the interaction between the driving factors of the risk or phenomenon according to their spatial autocorrelation. This software includes a risk detector, a factor detector, an ecological detector, and an interactive detector [39,41,57]:
(1) The risk detector uses t-test to identify whether the difference between the mean value of the environmental risk or geographical phenomenon in subregions is significant. The total region is classified into different subregions by using different classes and methods according to a driving factor. A significant difference means the driving factor has a great impact on the risk or phenomenon. The estimation is as follows:
t ij = R i − R j σ i 2 n i − σ j 2 n j ,
where tij is the inspection value of t, Ri or Rj is the mean value of the environmental risk or geographical phenomenon in the subregion i or j, respectively, σi2 or σj2 is the variance of the environmental risk or geographical phenomenon in the subregion i or j, respectively, ni or nj is the number of samples in the subregion i or j, respectively, and i or j is the number of subregions.
(2) The factor detector uses potential determinants (PD) to quantitatively identify the contribution of a driving factor to the environmental risk or geographical phenomenon. A higher PD denotes that a driving factor has a greater contribution to the occurrence of the environmental risk or geographical phenomenon. Its estimation is as follows:
P D = 1 − 1 n σ 2 ∑ i = 1 L n i σ i 2 ,
where PD is the explanatory power of a driving factor on a risk or phenomenon, σi2 and σ2 are, respectively, the variance of a risk or phenomenon in subregion i and the variance in the total region, where i = 1, …, L is the number of classes of a driving factor, and ni and n are the number of subregions in class i and the total region, respectively. The value of PD ranges from 0 to 1. A value of 1 indicates a driving factor can explain 100% of the environmental risk or geographical phenomenon, while a value of 0 indicates the opposite [39].
(3) The ecological detector uses the F-test to compare the impact between two driving factors on the environmental risk or geographical phenomenon. The estimation is as follows:
F = n C , P n C , P − 1 σ C , P 2 n D , P n D , P − 1 σ D , P 2 ,
where F is the value of the F-test, nC.P and nD.P are the numbers of driving factors C and D in subregion P, and σ C , P 2 and σ D , P 2 are the variances of driving factors C and D in subregion P.
(4) The interaction detector compares the combined contribution of two individual driving factors to the environmental risk or geographical phenomenon and their independent contribution. The combined contribution of two individual driving factors can be described as whether the driving factors C and D weaken or enhance each another when they are paired together, or whether they are independent in their influence on an environmental risk or geographical phenomenon. The interaction detector consists of five parts: enhance-bivariate, enhance-nonlinear, weaken-univariate, weaken-nonlinear, and independent [57]. The formulae and trends of the relationships between two driving factors are shown in Table 2. An independent relationship between two driving factors indicates that they have no additional impact on an environmental risk or geographical phenomenon when they are combined. An enhance-nonlinear or a weaken-nonlinear relationship between two driving factors indicates that they have a strong influence on the environmental risk or geographical phenomenon. An enhance-bivariate or a weaken-univariate relationship indicates that the two combined driving factors have a moderate effect on the environmental risk or geographical phenomenon.

3. Results

3.1. Driving Factors

3.1.1. Spatial Autocorrelation of Driving Factors

The meteorological data and groundwater depth in the study area were the observed values obtained from a scattered set of observation stations. ESDA was used to explore the spatial autocorrelation characteristics of every driving factor in order to acquire the interpolation-estimated spatial distribution of every driving factor. The results of the ESDA are shown in Table 3.
The mean values of temperature, precipitation, ground temperature, and groundwater depth were close to their median values (Table 3). The kurtosis of evaporation was close to 3 (Table 3). These results indicate that the distributions of meteorological data and groundwater depth were close to a normal distribution. There were direction trends and spatial autocorrelations in the scattered distribution of every driving factor (Table 3). There were also no outliers in the scattered distribution of every driving factor (Table 3). Therefore, the method of Kriging interpolation could be applied to estimate the spatial distribution of these driving factors.

3.1.2. Kriging Interpolation and Discretization of Driving Factors

Kriging interpolation was applied using the ArcGIS 10.1 software to interpolate the scattered values of meteorological data and groundwater depth in order to gain the estimated spatial distributions of these driving factors. Figure 5 displays the interpolated spatial distributions of the driving factors.
However, the spatial distributions of the natural driving factors were all continuous values. In order to further process the data using the Geodetector software and to be consistent with the social driving factors, the continuous values of every natural driving factor were recalculated for 89 townships.

3.2. Detection of Driving Factors

3.2.1. Risk Detector

The analysis of the risk detector was based on the results obtained after discretizing the driving factors. According to Cao et al. [50], the total region (89 townships) was divided into fewer subregions following the 2–8 breakpoints (3–9 classes) using the discretizing methods of equal interval, quantile, natural breakpoint, and geometrical interval in the ArcGIS 10.1 software. The average and variance of the geographical phenomenon (soil pH) in different subregions of a discretizing class and method were calculated. Then, the differences in driving factors in different sub-regions of all the discretizing classes and methods were calculated and compared using a t-test. The highest ratio of the number of significant t-test values to the number of all t-test values indicated the optimal classification method and class (the number of breakpoints and the number of subregions). The results of the t-test at the 0.05 probability level are shown in Table 4 for every driving factor.
As shown in Table 4, the discretizing methods of quantile and natural breakpoint were better than the methods of equal interval and geometrical interval for the driving factors. There were differences among the results of the t-tests for the driving factors classified using the methods of quantile and natural breakpoint. These results showed that the driving factors had a different impact on soil pH. As shown in Table 4, DEM (N6), spring temperature (N2), and 0 cm ground temperature in spring (N4) had the highest ratios (21/21, 20/21, and 6/6 > 80%). These results indicated that the three driving factors had the highest contribution to soil alkalinization. Similarly, farmland percentage (S2), population density (S1), and groundwater depth (N5) had higher ratios (16/28, 13/21, and 17/28 ≈ 60%). These results indicated that these three driving factors had a moderate contribution to soil alkalinization. Spring precipitation (N3) and spring evaporation (N1) had low ratios (7/21 and 10/21 < 50%). These results indicated that these two driving factors had a low contribution to soil alkalinization. According to the optimal discretizing method and class, the spatial distributions of the driving factors are displayed in Figure 6.

3.2.2. Factor Detector

The PD value was calculated to investigate the relative impact of various driving factors on soil pH. A higher PD value indicates a higher contribution to soil pH. The PD values of the driving factors were ranked, and the results are shown in Figure 7.
The rankings in order of importance based on the PD values of the driving factors with regard to their effect on soil pH were as follows: N6 (DEM) at 0.92 > N2 (spring temperature) at 0.80 > N4 (0 cm ground temperature in spring) at 0.74 > S2 (farmland percentage) at 0.48 > N1 (spring evaporation) at 0.42 > N5 (groundwater level) at 0.32 > S1 (population density) at 0.31 > N3 (spring precipitation) at 0.21. The PD values of DEM, spring temperature, and 0 cm ground temperature in spring were greater than 0.7. These high PD values indicated that these three driving factors were the primary driving factors for soil alkalinization. The PD values of farmland percentage, spring evaporation, groundwater level, and population density were lower than 0.5. These PD values indicated that these four driving factors were important driving factors for soil alkalinization. In particular, the PD value of farmland percentage was a primary social driving factor for soil alkalinization. Spring precipitation had the minimum impact (0.21) on soil pH.

3.2.3. Ecological Detector

According to the PD values, the ecological detector was used to explore whether a driving factor had more significant impact on soil alkalinization than other factors. The results are shown in Table 5.
The results suggested that the differences between the PD values of DEM and those of other factors were all statistically significant at the 0.05 probability level. The differences between the PD values of spring temperature and those of population density, farmland percentage, and spring evaporation were statistically significant at the 0.05 probability level. Therefore, DEM and spring temperature, especially DEM, had a strong impact on soil pH. The other six driving factors had a weak influence.

3.2.4. Interaction Detector

The interaction detector was used to explore whether two driving factors weakened or enhanced one another when they were paired together, or whether they were independent in their influence an environmental risk or geographical phenomenon (soil pH in this study). It compared the impact of two individual driving factors on soil alkalinization to the sum impact of the two driving factors when they were combined. The results are shown in Table 6.
The interaction between paired driving factors was greater than the impact of individual factors. These results indicated that the interaction between driving factors increased the probability or the degree of soil alkalinization.
Table 6 shows two pairs of natural driving factors: spring precipitation (N3) ∩ spring evaporation (N1) = 0.79 > spring precipitation (N3) + spring evaporation (N1) = 0.21 + 0.42 = 0.63, and groundwater depth (N5) ∩ spring precipitation (N3) = 0.71 > groundwater depth (N5) + spring precipitation (N3) = 0.32 + 0.21 = 0.52. These two pairs exhibited the most significant enhance-nonlinear interaction in which the combined impact was greater than the additive impact of individual driving factors. The remaining pairs of natural driving factors exhibited an enhance-bivariate interaction, in which the combined impact was greater than either of the two individual driving factors. The interaction of the two social driving factors was also enhance-bivariate. Three natural–social pairs exhibited an enhance-nonlinear interaction: spring evaporation (N1) ∩ population density (S1) = 0.78 > spring evaporation (N1) + population density (S1) = 0.42 + 0.31 = 0.73; spring precipitation (N3) ∩ population density (S1) = 0.63 > spring precipitation (N3) + population density (S1) = 0.21 + 0.31 = 0.52; spring precipitation (N3) ∩ farmland percentage (S2) = 0.74 > spring precipitation (N3) + farmland percentage (S2) = 0.21 + 0.48 = 0.69. The remaining natural–social pairs exhibited an enhance-bivariate interaction. The greatest contribution (0.96) of the combined natural–social pairs was the combination of DEM (N6) and population density (S1). The greatest contribution (0.69) of the combined social pairs was the combination of farmland percentage (S2) and population density (S1). The greatest contribution (0.96) of the combined natural pairs was the combination of DEM (N6) and spring precipitation (N3).

4. Discussion

4.1. Primary Driving Factors of Soil Alkalinization

The analysis of the risk detector, factor detector, and ecological detector revealed the primary driving factors of soil pH in the study area. DEM, spring temperature, and 0 cm ground temperature in spring had the strongest impact on soil pH. Their higher PD values also indicated that these three driving factors were the primary driving factors of soil pH. Farmland percentage, population density, and groundwater depth had a moderate impact on soil pH. Spring precipitation and spring evaporation had a low impact on soil pH. The lower PD values of the remaining five driving factors indicated that they had even less impact on soil pH. These results suggested that natural factors played a significant role in soil alkalinization in the study area. DEM (PD value = 0.92) and spring temperature (PD value = 0.80), especially DEM, had strong effects on soil alkalinization. Moreover, the differences between the PD values of DEM and those of the other factors were all statistically significant at the 0.05 probability level. These results suggested that topography macroscopically controlled the spatial occurrence and development of soil alkalinization on the soil surface in the study area. DEM, which was the dominant topographic driving factor and whose values were representative of closed terrain and low-lying land, hindered alkali salts in rivers from outflowing and trapped them in the depression of the basin. Alkali salts ultimately accumulated in the low-lying area of the basin. Similarly, Su et al. [48] found that elevation (PD = 0.64) was an important factor for soil salinization in a coastal area using Geodetector. Although elevation was not the primary factor, the contributions of a flat coastal plain and poor drainage conditions in the coastal area to soil salinization were relatively high. These results suggest that topography was a key factor for soil salinization and alkalinization.
The impact of spring temperature (PD value = 0.80) was also relatively high. The differences between the PD values of spring temperature and those of population density, farmland percentage, and spring evaporation were statistically significant at the 0.05 probability level. These results might be due to the warming trend of spring temperature greatly promoting the concentration of alkali salts in the soil surface of the low-lying area. The 0 cm ground temperature in spring (PD value = 0.74) was highly related to spring temperature and had a high impact on soil salinization and alkalinization. Similarly, temperature (PD = 0.61) was also an important factor for soil salinization of coastal areas [48].
Farmland percentage, which involves human activities, might also be a social factor influencing soil salinization and alkalinization. Grassland reclamation near wetlands and rivers can degrade soil vegetation, increasing the proneness of the study area to alkalinization.

4.2. The Interaction of Driving Factors

The quantitative analysis of the interaction detector (Table 6) suggested that the interaction between almost any pairs of driving factors had a greater impact on soil alkalinization than the impact of individual factors in this study. In particular, the findings suggested that the interaction between moderate driving factors could greatly increase the probability or degree of soil alkalinization, although DEM and spring temperature had a strong effect on soil alkalinization. There were significant enhance-nonlinear relationships (paired PD value = 0.71) between groundwater depth (PD value = 0.32) and spring precipitation (PD value = 0.21). Combining precipitation with groundwater depth accelerated the dissolution and transfer of alkali salts and raised the groundwater level and salt content. The shallow groundwater and high salt content likely caused alkali salts to rise to the soil surface through soil capillary. There was also a significant enhance-nonlinear relationship (paired PD value = 0.79) between spring precipitation (PD value = 0.21) and spring evaporation (PD value = 0.42). Combining precipitation with evaporation could accelerate the dissolution, aggregation, and recrystallization of alkali salts. In addition, social factors could enhance their influence when combined with a natural factor.
Overall, due to the factors related to topography and warming temperature, the degree of soil alkalinization in the study area is increasing substantively. Moreover, human activities, such as farming, over-grazing, and irrigation, accelerate soil alkalinization to some extent when these activities are combined with natural factors. However, soil alkalinization can be prevented or decelerated by some measures. Grasslands in the low-lying area near the rivers and the wetland should not be cultivated because vegetation can reduce soil evaporation and prevent soil from alkalinizing. Under global warming condition, an inter-basin water transfer project may be considered to decelerate soil alkalinizing.

5. Conclusions

This study used Geodetector to investigate the driving factors of soil alkalinization in the Wuyu’er–Shuangyang River Basin and detect the interaction between the driving factors on the basis of soil pH data acquired from image inversion. The results were as follows: (1) DEM, spring temperature, and 0 cm ground temperature in spring had the highest ratios of the number of significant t-test values to the number of all t-test values. These three driving factors were the primary driving factors of soil pH. (2) The higher PD values also indicated that DEM, spring temperature, and 0 cm ground temperature in spring were the primary driving factors of soil pH. DEM was the dominant driving factor among all driving factors. (3) The differences between the PD values of DEM and those of the remaining factors were statistically significant. Again, these significant differences suggested that DEM was the dominant driving factor among all driving factors. (4) The interactions between paired driving factors were all enhanced interactions. There interactions were significant and showed a strong enhance-nonlinear relationship between spring precipitation and spring evaporation, as well as between groundwater depth and spring precipitation. There were enhance-nonlinear relationships between the natural–social pairs and the social pairs, even though the impact of individual social factors was low. Overall, the natural driving factors of DEM and spring temperature played a key role in soil alkalinization. The social driving factors also contributed significantly to soil alkalinization when combined with the natural driving factors.
These findings explained the mechanisms that led to soil alkalinization in the Wuyu’er–Shuangyang River Basin. Furthermore, the Geodetector method can be applied to provide references and suggestions for other environmental problems or geographical phenomena.

Author Contributions

Conceptualization, L.B.; methodology, L.B. and J.Z.; software, L.B.; formal analysis, L.B. and J.L; investigation, L.B. and J.L.; data curation, L.B.; writing—original draft preparation, L.B.; writing—review and editing, H.D. and Y.Z.; funding acquisition, L.B. and J.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by the Fundamental Research Funds from the Heilongjiang Provincial Universities, under grant no. 135509120, and the Heilongjiang Provincial Nature Funds, under grant no. LH2020C109.

Data Availability Statement

Not applicable.

Acknowledgments

The authors thank the anonymous reviewers and editors for improving this paper. The authors also thank Yuexiang Wu and Jiaqi Wu for collecting data and writing this paper.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Gebremeskel, G.; Gebremicael, T.G.; Kifle, M.; Meresa, E.; Gebremedhin, T.; Girmay, A. Salinization pattern and its spatial distribution in the irrigated agriculture of nor thern Ethiopia: An integrated approach of quantitative and spatial analysis. Agric. Water Manag. 2018, 206, 147–157. [Google Scholar] [CrossRef]
  2. Singh, A. Soil salinization and waterlogging: A threat to environment and agricultural sustainability. Ecol. Ind. 2015, 57, 128–130. [Google Scholar] [CrossRef]
  3. Li, Q.F.; Xi, M.; Wang, Q.G.; Kong, F.; Li, Y. Characterization of soil salinization in typical estuarine area of the jiaozhou bay, china. Phys. Chem. Earth 2018, 103, 51–61. [Google Scholar] [CrossRef]
  4. Metternicht, G. Assessing temporal and spatial changes of salinity using fuzzy logic, remote sensing and GIS: Foundations of expert system. Ecol. Model. 2001, 144, 163–179. [Google Scholar] [CrossRef]
  5. Wang, L.; Seki, K.; Miyazaki, T.; Ishihama, Y. The causes of soil alkalinization in the Songnen Plain of Northeast China. Paddy Water Environ. 2009, 7, 259–270. [Google Scholar] [CrossRef]
  6. Csillag, F.; Pásztor, L.; Biehl, L.L. Spectral band selection for the characterization of salinity status of soils. Remote Sens. Environ. 1993, 43, 231–242. [Google Scholar] [CrossRef]
  7. Buckman, H.O.; Brady, N.C.; Weil, R.R. The Nature and Properties of Soils; Prentice Hall: Englewod Cliffs, NJ, USA, 2002. [Google Scholar]
  8. Kumar, V.; Vimal, B.K.; Kumar, R.; Kumar, R.; Kumar, M. Determination of soil pH by using digital image processing technique. J. Appl. Nat. Sci. 2014, 6, 14–18. [Google Scholar] [CrossRef]
  9. Li, X.J. The alkali-saline land and agricultural sustainable development of the Western Songnen Plain in China. Sci. Geogr. Sin. 2000, 20, 51–55. [Google Scholar]
  10. Ou, Y.; Alain, N.R.; Wang, L.X.; Yan, B. Spatio-temporal patterns of soil organic carbon and pH in relation to environmental factors-A case study of the Black Soil Region of Northeastern China. Agric. Ecosyst. Environ. 2017, 245, 22–31. [Google Scholar] [CrossRef]
  11. Liu, Z.J.; Yang, X.G.; Hubbard, K.G.; Lin, X. Maize potential yields and yield gaps in the changing climate of northeast China. Glob. Chang. Biol. 2012, 18, 3441–3454. [Google Scholar] [CrossRef]
  12. Wang, R.Z.; Li, J.D. Dynamic population models of the ecological dominance during the deterioration of Leymuschinensis grassland. Acta Phytoecol. Sin. 1995, 19, 170–174. (In Chinese) [Google Scholar]
  13. Zhang, F.; Xiong, H.G.; Ding, J.L.; Xia, Q. Characteristics of laboratory-field measured spectra responding to alkalinized soil and conversion. Trans. Chin. Soc. Agric. Eng. 2012, 28, 101–107. (In Chinese) [Google Scholar]
  14. Zabcic, N.; Rivard, B.; Ong, C.; Mueller, A. Using airborne hyperspectral data to characterize the surface pH and mineralogy of pyrite mine tailings. Int. J. Appl. Earth Obs. Geoinf. 2014, 32, 152–162. [Google Scholar] [CrossRef]
  15. Bai, L.; Wang, C.; Zang, S.; Zhang, Y.; Hao, Q.; Wu, Y. Remote Sensing of Soil Alkalinity and Salinity in the Wuyu’er-Shuangyang River Basin, Northeast China. Remote Sens. 2016, 8, 163. [Google Scholar] [CrossRef] [Green Version]
  16. Bai, L.; Wang, C.Z.; Zang, S.Y.; Wu, C.; Luo, J.; Wu, Y. Mapping Soil Alkalinity and Salinity in Northern Songnen Plain, China with the HJ-1 Hyperspectral Imager Data and Partial Least Squares Regression. Sensors 2018, 18, 3855. [Google Scholar] [CrossRef] [Green Version]
  17. Niñerola, V.B.; Navarro-Pedreño, J.; Lucas, I.G.; Pastor, I.M.; Vidal, M.M.J. Geostatistical assessment of soil salinity and cropping systems used as soil phytoremediation strategy. J. Geochem. Explor. 2017, 174, 53–58. [Google Scholar] [CrossRef]
  18. Wang, R.Z.; Gao, Q.; Chen, Q. Effects of climatic change on biomass and biomass allocation in Leymuschinensis (Poaceae) along the North-east China Transect (NECT). J. Arid Environ. 2003, 54, 653–665. [Google Scholar]
  19. Wang, F.S.; Tian, Z.C. The groundwater effect in the process of soil salinization of the Songnen Plain, Jilin province. Jilin Geol. 2002, 21, 79–87. (In Chinese) [Google Scholar]
  20. Zhang, D.F.; Wang, S.J. Mechanism of freeze-thaw action in the process of soil salinization in northeast China. Environ. Geol. 2001, 41, 96–100. [Google Scholar] [CrossRef]
  21. Li, J.G.; Pu, L.J.; Han, M.F.; Hu, M.; Zhang, R.; Xiang, Y. Soil salinization research in china: Advances and prospects. J. Geogr. Sci. 2014, 24, 943–960. [Google Scholar] [CrossRef]
  22. Ma, S.; Zhao, J.; Chen, J.; Zhang, S.; Dong, T.; Mei, Q.; Hou, X.; Liu, G. Ground Surface Freezing and Thawing Index Distribution in the Qinghai-Tibet Engineering Corridor and Factors Analysis Based on GeoDetector Technique. Remote Sens. 2023, 15, 208. [Google Scholar] [CrossRef]
  23. Okur, B.; Ren, N. Soil salinization and climate change. In Climate Change and Soil Interactions; Elsevier: Amsterdam, The Netherlands, 2020; pp. 331–350. [Google Scholar]
  24. Jin, Z.; Guo, L.; Wang, Y.Q.; Yu, Y.; Lin, H.; Chen, Y.; Chu, G.; Zhang, J.; Zhang, N. Valley reshaping and damming induce water table rise and soil salinization on the Chinese Loess Plateau. Geoderma 2019, 339, 115–125. [Google Scholar] [CrossRef]
  25. Tomaz, A.; Palma, P.; Fialho, S.; Lima, A.; Alvarenga, P.; Potes, M.; Costa, M.J.; Salgado, R. Risk Assessment of Irrigation-Related Soil Salinization and Sodification in Mediterranean Areas. Water 2020, 12, 3569. [Google Scholar] [CrossRef]
  26. Qiu, S.W.; Zhang, B.; Wang, Z.C. Analyses on current situation, causes of formation, and way of management of desertification in western Northeast Plain of China. Quat. Sci. 2005, 25, 63–73. (In Chinese) [Google Scholar]
  27. Lin, N.F.; Bounlom, V.; Tang, J.; Bian, J.M. Study on the relation between the formation of saline-alkali soil and the neotectonic movement. Glob. Geol. 2005, 24, 282–288. (In Chinese) [Google Scholar]
  28. Wu, L.Z.; Li, Q.S. Research of mechanism of saline desertification in Western Songnen Plain. Soil Water Conserv. 2003, 17, 79–81. (In Chinese) [Google Scholar]
  29. Lin, N.F.; Tang, J. Study on the environment evolution and the analysis of cause to land salinization and desertification in Songnen Plain. Quat. Sci. 2005, 25, 474–483. (In Chinese) [Google Scholar]
  30. Wang, X.F. Analysis on population growth of the Northeast China before 1949. Northeast Asia Forum. 2000, 4, 86–88. (In Chinese) [Google Scholar]
  31. Zhao, Y.L.; Ye, Z.S.; Cui, G.W. Sustainable use and development of grassland in the Songnen Plain. Heilongjiang Ani Sci. Vet. Med. 2004, 8, 56–57. (In Chinese) [Google Scholar]
  32. Louati, D.; Majdoub, R.; Rigane, H.; Abida, H. Effects of Irrigating with Saline Water on Soil Salinization (Eastern Tunisia). Arab. J. Sci. Eng. 2018, 43, 3793–3805. [Google Scholar] [CrossRef]
  33. Haj-Amor, Z.; Tóth, T.; Ibrahimi, M.-K.; Bouri, S. Effects of excessive irrigation of date palm on soil salinization, shallow groundwater properties, and water use in a Saharan oasis. Environ. Earth Sci. 2017, 76, 590. [Google Scholar] [CrossRef]
  34. Fauck, R. Influences of Agriculture Practices on Soil Degradation; FAO Soil Bulletion: Rome, Italy, 1977; Volume 34, pp. 83–84. [Google Scholar]
  35. Abdullah, A.Y.M.; Biswas, R.K.; Chowdhury, A.I.; Billah, S.M. Modeling soil salinity using direct and indirect measurement techniques: A comparative analysis. Environ. Dev. 2019, 29, 67–80. [Google Scholar] [CrossRef]
  36. Penov, I.; Manolov, I.; Alexiev, A.; Kavardziev, Y. Salinisation in Bulgaria: Institutional settings for soil conservation (a case study of Belozem village). Land Degrad. Dev. 2011, 22, 134–143. [Google Scholar] [CrossRef]
  37. AbdelRahman, M.A.E.; Metwaly, M.M.; Shalaby, A. Quantitative assessment of soil saline degradation using remote sensing indices in Siwa Oasis. Remote Sens. Appl. Soc. Environ. 2019, 13, 53–60. [Google Scholar] [CrossRef]
  38. Wang, J.F.; Li, X.H.; Chirstakos, G.; Liao, Y.-L.; Zhang, T.; Gu, X.; Zheng, X.-Y. Geographical detectors-based health risk assessment and its application in the neural tube defects study of the Heshun region, China. Int. J. Geogr. Inf. Sci. 2010, 24, 107–127. [Google Scholar] [CrossRef]
  39. Wang, J.F.; Hu, Y. Environmental health risk detection with GeogDetector. Environ. Model. Softw. 2012, 33, 114–115. [Google Scholar] [CrossRef]
  40. Wen, Q.; Yang, L.; Yu, J.; Wei, B.; Yin, S. Sources and Risk Characteristics of Heavy Metals in Plateau Soils Predicted by Geo-Detectors. Remote Sens. 2023, 15, 1588. [Google Scholar] [CrossRef]
  41. Hu, Y.; Wang, J.F.; Li, X.H.; Ren, D.; Zhu, J. Geodetector: Principle and prospective Geographical Detector-based Risk Assessment of the under-five mortality in the 2008 Wenchuan earthquake, China. PLoS ONE 2011, 6, e21427. [Google Scholar]
  42. Luo, L.L.; Mei, K.; Qu, L.Y.; Zhang, C.; Chen, H.; Wang, S.; Di, D.; Huang, H.; Wang, Z.; Xia, F.; et al. Assessment of the Geographical Detector Method for investigating heavy metal source apportionment in an urban watershed of Eastern China. Sci. Total Environ. 2019, 653, 714–722. [Google Scholar] [CrossRef] [Green Version]
  43. Wei, C.X.; Guo, B.; Fan, Y.W.; Zang, W.; Ji, J. The change pattern and Its dominant driving factors of wetlands in the Yellow River delta based on Sentinel-2 images. Remote Sens. 2022, 14, 4388. [Google Scholar] [CrossRef]
  44. Zhang, S.; Zhou, Y.; Yu, Y.; Li, F.; Zhang, R.; Li, W. Using the Geodetector method to characterize the spatiotemporal dynamics of vegetation and Its interaction with environmental factors in the Qinba Mountains, China. Remote Sens. 2022, 14, 5794. [Google Scholar] [CrossRef]
  45. Ji, B.; Qin, Y.; Zhang, T.; Zhou, X.; Yi, G.; Zhang, M.; Li, M. Analyzing Driving Factors of Drought in Growing Season in the Inner Mongolia Based on Geodetector and GWR Models. Remote Sens. 2022, 14, 6007. [Google Scholar] [CrossRef]
  46. Yang, H.; He, Q.; Cui, L.; Mohamed Taha, A.M. Exploring the Spatial Relationship between Urban Vitality and Urban Carbon Emissions. Remote Sens. 2023, 15, 2173. [Google Scholar] [CrossRef]
  47. He, X.H.; Tian, J.M.; Zhang, Y.Q.; Zhao, Z.; Cai, Z.; Wang, Y. Attribution and driving force of nitrogen losses from theTaihu Lake Basin by the InVEST andGeoDetector models. Sci. Rep. 2023, 13, 7440. [Google Scholar] [CrossRef]
  48. Su, Y.; Li, T.X.; Cheng, S.K.; Wang, X. Spatial distribution exploration and driving factor identification for soil salinisation based on geodetector models in coastal area. Ecol. Eng. 2020, 156, 105961. [Google Scholar] [CrossRef]
  49. Song, H.Y. Detection in Near-Infrared Spectroscopy of Soils; Chemistry Industry Press: Beijing, China, 2013; pp. 97–98. [Google Scholar]
  50. Cao, F.; Ge, Y.; Wang, J.F. Optimal discretization for geographical detectors-based risk assessment. GISci. Remote Sens. 2013, 50, 78–92. [Google Scholar] [CrossRef]
  51. Dougherty, J.; Kohavi, R.; Sahami, M. Supervised and Unsupervised Discretization of Continuous Features. In Proceedings of the 12th International Conference on Machine Learning, Tahoe City, CA, USA, 9–12 July 1995; pp. 194–202. [Google Scholar]
  52. Ge, Y.; Cao, F.; Duan, R.F. Impact of discretization methods on the rough set-based classification of remotely sensed images. Int. J. Digit. Earth 2011, 4, 330–346. [Google Scholar] [CrossRef]
  53. Haining, R.; Wise, S.; Ma, J. Exploratory spatial data analysis in a geographic information system environment. Statistician 1998, 47, 457–469. [Google Scholar] [CrossRef]
  54. Pisati, M. Exploratory spatial data analysis using Stata. In Proceedings of the German Stata Users Group Meetings: Stata Users Group, Berlin, Germany, 1 June 2012. [Google Scholar]
  55. Bargaoui, Z.K.; Chebbi, A. Comparison of two kriging interpolation methods applied to spatiotemporal rainfall. J. Hydrol. 2009, 365, 56–73. [Google Scholar] [CrossRef]
  56. Ha, H.; Olson, J.R.; Bian, L.; Rogerson, P.A. Analysis of Heavy Metal Sources in Soil Using Kriging Interpolation on Principal Components. Environ. Sci. Technol. 2014, 48, 4999–5007. [Google Scholar] [CrossRef]
  57. Wang, J.F.; Xu, C.D. Geodetector: Principle and prospective. Acta Geogr. Sin. 2017, 73, 219–231. (In Chinese) [Google Scholar]
Sustainability 15 11538 g001 550
Figure 1. The location of the Wuyu’er–Shuangyang River Basin.
Figure 1. The location of the Wuyu’er–Shuangyang River Basin.
Sustainability 15 11538 g001
Sustainability 15 11538 g002 550
Figure 2. Spatial distribution of observation points: (a) meteorological data; (b) groundwater depth.
Figure 2. Spatial distribution of observation points: (a) meteorological data; (b) groundwater depth.
Sustainability 15 11538 g002
Sustainability 15 11538 g003 550
Figure 3. Townships in the study area.
Figure 3. Townships in the study area.
Sustainability 15 11538 g003
Sustainability 15 11538 g004 550
Figure 4. Spatial distribution of different soil pH values in the study area.
Figure 4. Spatial distribution of different soil pH values in the study area.
Sustainability 15 11538 g004
Sustainability 15 11538 g005 550
Figure 5. The Kriging interpolation of scattered values: (a) spring evaporation; (b) spring temperature; (c) spring precipitation; (d) 0 cm ground temperature in spring; (e) groundwater depth.
Figure 5. The Kriging interpolation of scattered values: (a) spring evaporation; (b) spring temperature; (c) spring precipitation; (d) 0 cm ground temperature in spring; (e) groundwater depth.
Sustainability 15 11538 g005
Sustainability 15 11538 g006a 550Sustainability 15 11538 g006b 550
Figure 6. The discretization of driving factors: (a) population density; (b) farmland percentage; (c) spring evaporation; (d) spring temperature; (e) spring precipitation; (f) 0 cm ground temperature in spring; (g) groundwater depth; (h) altitude.
Figure 6. The discretization of driving factors: (a) population density; (b) farmland percentage; (c) spring evaporation; (d) spring temperature; (e) spring precipitation; (f) 0 cm ground temperature in spring; (g) groundwater depth; (h) altitude.
Sustainability 15 11538 g006aSustainability 15 11538 g006b
Sustainability 15 11538 g007 550
Figure 7. The rankings in order of importance based on the PD values of the driving factors.
Figure 7. The rankings in order of importance based on the PD values of the driving factors.
Sustainability 15 11538 g007
Table
Table 1. The category, type, unit, and code of driving factors.
Table 1. The category, type, unit, and code of driving factors.
CategoryTypeDriving FactorUnitCode
Social
factors		PopulationPopulation densitypersons/km2S1
Land useFarmland percentage%S2
Natural factorsClimateSpring evaporationmmN1
Spring temperature°CN2
Spring precipitationmmN3
0 cm ground temperature in spring°CN4
GroundwaterGroundwater depthmN5
TopographyAltitudemN6
Table
Table 2. The type, formula, and trend of interaction between driving factors C and D.
Table 2. The type, formula, and trend of interaction between driving factors C and D.
Interaction TypeFormulaTrend
Enhance-nonlinearPD(C∩D) > PD(C) + PD(D)High
Low
Null
Low
High		Sustainability 15 11538 i001
Enhance-bivariatePD(C) + PD(D) > PD(C∩D) > Max(PD(C), PD(D))
IndependentPD(C∩D) = PD(C) + PD(D)
Weaken-univariateMin(PD(C), PD(D)) < PD(C∩D) < Max(PD(C), PD(D))
Weaken-nonlinearPD(C∩D) < Min(PD(C), PD(D))
Table
Table 3. The results of ESDA for meteorological data and groundwater depth.
Table 3. The results of ESDA for meteorological data and groundwater depth.
Diving
Factor		Characteristics of Normal Distribution TrendOutlierSpatial-
Autocorrelation
MeanMedianKurtosis
N162.1542.263.31East–westNullExist
N23.533.781.80East–westNullExist
N33.643.004.54South–northNullExist
N47.006.71.90East–westNullExist
N53.383.072.81South–northNullExist
Table
Table 4. Optimal discretization method, number of classes, and results of the risk detector.
Table 4. Optimal discretization method, number of classes, and results of the risk detector.
Driving FactorResults of
t-Test
(Significant/
All Statistics)		Discretizing
Method		Number of BreakpointsNumber of Subregions
S113/21Quantile67
S216/28Natural breakpoint78
N110/21Quantile67
N220/21Natural breakpoint67
N37/21Natural breakpoint67
N46/6Quantile34
N517/28Natural breakpoint78
N621/21Quantile67
Table
Table 5. Statistically significant difference between two PD values.
Table 5. Statistically significant difference between two PD values.
S1 S2 N1N2 N3N4N5N6
S1
S2N
N1NN
N2YYY
N3NNNN
N4YYYNY
N5NNNNNN
N6YYYYYYY
Note: Y indicates that the contribution of one driving factor to soil alkalinization is more significant than that of another driving factor at the 0.05 probability level. N indicates that the difference in contribution is not significant.
Table
Table 6. Impact on soil pH of individual and paired driving factors.
Table 6. Impact on soil pH of individual and paired driving factors.
Driving FactorS1S2N1N2N3N4N5N6
Driving FactorPD0.310.480.420.800.210.740.320.92
S10.31
S20.480.69
N10.420.780.87
N20.800.860.890.93
N30.210.630.740.790.93
N40.740.820.840.820.900.80
N50.320.600.750.720.890.710.80
N60.920.960.950.950.950.960.920.95
Note: Sustainability 15 11538 i002 indicates an enhance-nonlinear relationship; Sustainability 15 11538 i003 indicates an enhance-bivariate relationship.
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

Bai, L.; Zhou, J.; Luo, J.; Dou, H.; Zhang, Y. Analyzing Driving Factors of Soil Alkalinization Based on Geodetector—A Case in Northeast China. Sustainability 2023, 15, 11538. https://0-doi-org.brum.beds.ac.uk/10.3390/su151511538

AMA Style

Bai L, Zhou J, Luo J, Dou H, Zhang Y. Analyzing Driving Factors of Soil Alkalinization Based on Geodetector—A Case in Northeast China. Sustainability. 2023; 15(15):11538. https://0-doi-org.brum.beds.ac.uk/10.3390/su151511538

Chicago/Turabian Style

Bai, Lin, Jia Zhou, Jinming Luo, Hongshuang Dou, and Ye Zhang. 2023. "Analyzing Driving Factors of Soil Alkalinization Based on Geodetector—A Case in Northeast China" Sustainability 15, no. 15: 11538. https://0-doi-org.brum.beds.ac.uk/10.3390/su151511538

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