Management of Al³⁺ Residue in the Soil by Mapping Soil Capability in Retaining and Transporting Al³⁺ in the Farmland of Trang Bom District, Vietnam

Abstract

The assessment of soil capability in retaining and transporting chemical substances is necessary, especially currently, with the overuse of chemical products for crop production. Depending on the soil properties, these chemicals may bound on soil particles or release and transport in the soil solution. In this study, we developed maps of the capability of soil to retain and transport Al³⁺, thereby evaluating the main soil factors affecting Al³⁺ fate in the agricultural land of Trang Bom District, Dong Nai Province, Vietnam. Information and data of the factors slope, soil texture, pH, organic matter, and ferrallitisation were processed and analyzed. The GIS tool was applied in combination with the analytical hierarchical process (AHP) to create the maps. Four hundred simulation runs were performed for criteria weight sensitivity analysis to explore the dependency of the resultant maps on the weights of the input factors. Sampling soil data were used to validate the accuracy of information given by the resultant maps. Results from the two maps show that the soils in the area have high capability in retaining and transporting Al³⁺. Ninety nine percent of the soils in the area have medium to high capability of Al³⁺ retention and about 65% of the soils have medium to high capability of transporting Al³⁺. For the agricultural land, about 65% of the land ranked as having a high to very high soil Al³⁺ retention capability and about 58% of the land ranked as having a medium to high capability of transporting Al³⁺. These maps can support the process of decision-making in identifying the appropriate dose and frequency of the chemical products that are applied on each soil capability zone; in this case study, the products contain aluminum. The accumulation of Al³⁺ in the soil, especially in the high Al³⁺ retention capability soil, can cause soil degradation and can cause negative effects on plant growth.

1. Introduction

One of the main factors limiting soil quality and agricultural production is related to excessive exchangeable aluminum (Al³⁺) presence in the soils. This is because the acid cations such as Al³⁺ induce hydrolysis and decrease soil pH, leading to phytotoxicity in the rhizosphere [1,2,3]. A high concentration of Al³⁺ can adversely affect many physiological processes of the plant, such as inhibiting root elongation, which subsequently impairs nutrient and water uptake [3], occupying and replacing the position of calcium in the cell membrane, inhibiting the growth of the membrane, leading to the abnormal growth of root cells, causing biomass retardation, depositing callose and lignin in the root tips, and many others [3,4]. Excessive Al³⁺ may alter nutrient levels such as N, K, Ca, Mg, and P, and can reduce the photosynthetic rate, stomatal conductance, and leaf transpiration rate in plants [5]. Apart from damaging the soil quality and plant health, Al³⁺ in the soil can deteriorate underground water.

In addition to the natural occurrence of Al³⁺, some plants protection products (pesticides, fungicides, herbicides, etc.) release Al³⁺ into the soil when being used, such as Aliette 800 WG, which is known as one of the popular phosphonate fungicides that helps to inhibit harmful fungi growth in soils and plants, e.g., the fungi Phytophthora, Pythium, and Plasmopara [6,7]. Alternatively, the application of acidifying fertilizer, including elemental sulfur (S) and ammonium (NH₄) salt, can accelerate the soil acidification process, leading to increasing Al³⁺ concentrations in the soil solution [8]. The fate and mobility of chemicals in the soil involve complex mechanisms that are influenced by many processes, which depend on the type of chemicals and soil characteristics [9].

A healthy agricultural environment is a basic requirement, not only for food safety but also for environmental quality and human health [10,11]. The continued accumulation of chemicals due to the overuse of agricultural chemical products severely threatens the quality of agricultural lands. Thus, in order to use crop protection products effectively and to mitigate the negative effects of their residues, it is necessary to evaluate the capability of retaining and transporting chemical residues (Al³⁺ in this study) in different soil ecotopes.

Land assessment on the potentials and limitations of the land for crop production has been increasing [12,13]; for example the studies on mapping land suitability for wheat farming in Southern Iran [14], and mapping soil organic matter and evaluating the related natural and anthropogenic influencing factors in Croatia [15]. There have also been studies that have focused on soil pollution and that have mapped their distribution. In those studies, the pollutants and nutrients were analyzed and then their pollution levels were determined and the related risks were assessed, such as the studies on the spatial distribution of copper concentration in agricultural areas from 2004 to 2017 [16], the spatial distribution of mercury [17], and the influences of lead (Pb), cadmium (Cd), chromium (Cr), and arsenic (As) concentrations on grain production in China [18,19,20,21,22,23].

These results did not give sufficient information on the capacity of the soil to retain or to transport chemicals in the environment. Typical characteristics of the soil may affect the fate of the pollutants. In the case of aluminum, the forms and amount of Al in the soil depend on the amount and type of Al compounds present in the soil and their reactions within the soil solution [2,24]. The bioavailability of Al and nutrients in the soil depends on their concentration [2,25]. The different forms of Al available in the soil involve the retention of anions and cations in the soil. The speciation of Al compounds is controlled by the hydrolysis of the Al³⁺ ion [2,26] and is affected by many factors such as soil pH, organic matter, clay mineralogy, soil slope, Al concentration, and concentrations of other cations and anions in the soil solution. Additionally, the spatial patterns of Al distributed in the soil revealed that the soil contamination is related to the types of land use, the types of anthropogenic input, and the soil properties. On agricultural areas, the complexity of the soil matrix as well as the unmanageable usage of plant protection chemicals make it more difficult to detect the behavior of chemical residues in the soil. For that reason, mapping soil capacities in retaining and transporting contaminants is important. It helps in guiding the policymakers to make reasonable decisions for the appropriate chemicals to be used to protect the crops but also to protect soil resources and prevent pollution.

The primary objectives of this research were to create maps of capabilities of soils that influence the fate of Al³⁺ in the agricultural area of Trang Bom District, Dong Nai Province, Vietnam. These maps show the capability of the soil, which depends on its characteristics, in adsorbing Al³⁺ (soil Al³⁺ retention capability), and the characteristics of the soil as a medium for Al³⁺ to transport or spread to adjacent areas (soil Al³⁺ transportation capability).

The method applied was the GIS-based analytic hierarchy process (AHP). AHP is one of the popular methods introduced by Saaty [27] to resolve multi-criteria decision-making (MCDM) problems. This method incorporates multi-factors that give different dimensions and scores for the assessment [14]. Geographic information systems (GIS) are best suited for processing spatial datasets to be included in spatial modeling [28]. Spatial analysis functions in GIS support in determining the spatially relative importance of interactive factors. The combined use of GIS and AHP in zoning land characteristics supports the decision-making process and enhances the evaluation accuracy on a regional to local scale [29]. The combination of GIS and AHP is currently an approach for land suitability evaluation [12,14,30].

2. Materials and Methods

2.1. Study Area

The study was conducted in Trang Bom District, which lies between a latitude from 10°51′38.85″ N to 11° 5′34.92″ N and a longitude from 106°54′3.56″ E to 107° 7′35.94″ E. This area belongs to Dong Nai Province in the south of Vietnam (Figure 1). The total area is 32,541.2 ha. It is a hilly area, and its elevation slightly descends from north to south, ranging from ~10 to 148 m above mean sea level. More than 85% of the area has the slope ranging from 0 to 8 degrees.

The climate is typical for a tropical monsoon region and has two seasons: the rainy season lasts from April to November and the dry season is from December to March. The annual average temperature is about 25–26 °C, maximum annual temperature is from 34 to 35 °C, and the minimum annual temperature is from 20 to 21 °C. The average rainfall is 1800–2000 mm year⁻¹, and the highest rainfall reaches up to 2550 mm year⁻¹ and the lowest rainfall is 1500 mm year⁻¹ (statistical data during the last 10 years).

The soil in Trang Bom is composed of 5 main groups based on the soil classification system from IUSS Working Group WRB [31]: Ferralsols (accounting for 39.76% of the area), Luvisols (46.36%), Acrisols (8.53%), Gleysols (2.04%), and Leptosols (3.31%). These soils have proper nutrients for crops to grow. Natural conditions are very suitable for agricultural development with a variety of crops. Due to favorable conditions, Trang Bom has a large area of agricultural land, accounting for about 70% (by 2020) of the total area. Agricultural areas cover most of the southeastern and northern parts of Trang Bom where the Luvisols and Ferralsols are distributed (Figure 1d).

2.2. GIS-Based Multicriteria Analysis

The GIS-based AHP method was applied to calculate the weights and importance levels of soil capabilities in adsorbing and transporting Al³⁺. The following steps were carried out: (i) Defining the objective, (ii) Identifying factors and constraints using different information sources (expert opinion, literature search, and analysis of field data), (iii) Defining the relationships of the soil properties and their capability for retaining and transporting Al³⁺, (iv) Weighting the criteria based on their relative importance to each soil capability (retention or transport), (v) Combining and aggregating all the layers/criteria to produce a final weighted estimate of soil capabilities, and (vi) Conducting sensitivity analyses and validating the results.

Procedure for conducting the study is shown in the workflow in Figure 2. Implementation of the methodology was conducted using ArcGIS 10.4.1 software (ESRI, Redlands, CA, USA). Coordinate reference system of spatial data was the Universal Transverse Mercator (UTM) zone 48 North. Data were converted into raster layers at 100 m² cell size.

2.2.1. Calculation Method for Soil Capability in Retaining and Transporting Al³⁺

Procedure of the multicriteria analysis based on AHP was applied for the calculation of soil Al³⁺ retention and transportation levels. This procedure includes the standardization of criteria values, their weights determination, and the aggregation of selected criteria. Standardization was performed in four classes (from 1 to 4) by stepwise standardization method; level 1 denotes the lowest and 4 denotes the highest soil capability. A comparison of each criterion to one another with a rating scale (1–9) was carried out using a pairwise comparison matrix [32,33]. Detailed procedure of pairwise comparison was described in the work of Saaty [32]. Assessment of the consistency was conducted through the consistency ratio (CR), and CR values < 0.10 (indicating a real estimation) were chosen. CR is calculated as follows,

$$CR=\frac{CI}{RI}$$

(1)

$$CI=\frac{\lambda -n}{n-1}$$

(2)

where CI is the consistency index, RI is the random consistency index representing an average CI from a random matrix, λ is the average value of consistency vector, and n is the number of criteria [34].

Soil capability classes are calculated as follows:

$$S={\displaystyle \sum }_{i=1}^n{W}_i{X}_i$$

(3)

where S is the soil capability class. W_i and X_i are the weights and importance levels of factor i, and n is the total number of factors. Soil capabilities range from 1 to 4, level 1 (S1, S < 1.5) shows no capability for Al³⁺ retention or transportation in the soil, level 2 (S2, S $\in$ (1.5, 2.5)) shows medium capability, level 3 (S3, S $\in$ (2.5, 3.5)) shows high capability, and level 4 (S4, S > 3.5) shows a strong capability. Local experts’ opinions (n = 53) and the related referenced works as well as sampling data were used as the basis to determine the factors and their ranks of capabilities.

2.2.2. Determination of Influencing Factors

The critical soil characteristics at which sufficient Al is toxic (in form of Al³⁺) depend on soil factors including soil texture (grain size), soil weathering characteristics, soil pH, organic matter (OM), soil slope as well as other cations and anions [35,36,37,38,39]. Thus, in this work, the five indicators of soil pH, organic matter, soil slope, soil texture, and ferrallitisation (which is typical in the study area) were considered.

The first factor is soil pH. Soil pH is an important variable of the soil, controlling solubility, bioavailability, mobility, ionic speciation, and ultimately toxicity of any metal in the soil [40]. Kryzevicius et al. [41] reported that Al³⁺ decreased at a soil pH above 5.9. When soil pH decreases to below 5.5, the solubility of Al increases exponentially [42]. Their result scoped well with our field sampling data (Figure 3a).

The second factor is soil organic matter (OM), which is a storehouse of nutrients for plants. The amount, composition, and properties of OM have a major influence on the soil formation and on the processes of physical, chemical, and biological properties occurring in the soil. Exchangeable aluminum can be complexed by soluble and solid forms of OM, causing the decrease in the acid dissociation constants of the soil organic fraction [43,44].

The concentration of Al³⁺ also depends on the clay content in the soil because this soil property determines the characteristics and the availability of soil colloids, and also the ability to exchange cations in the soil. When molecules exist in the form of cations, the adsorption process will be strong because soil colloids (clay minerals and humus) are mainly negatively charged. The adsorption capacity of clay rich soil is higher than that of loamy soil or sandy soil. Exchangeable aluminum is primarily bound to silicate clays [9,40].

Characteristics of the soil weathering process also affect Al content in soil, because aluminum-containing rocks that become weathered release more Al and gibbsite (Al(OH)³) [45]. In this study area, the typical and common soil forming processes are ferrallitisation and humification, which are the processes of rock weathering to form soils consisting of clay and sesquioxide, in the form of hydrated oxides of iron and aluminum [46].

Soil slope is another factor influencing Al³⁺ transportation in the soil because it involves geographical disparities which may affect the spatial differences of the amounts of chemicals observed. Higher slope degrees will increase the spread of Al³⁺ to the adjacent areas.

2.3. Data Sources and Data Processing

2.3.1. Collected Data

Data of the above identified factors were collected and processed. The digital elevation model (DEM, Figure 1b) was obtained from the United States Geological Survey—USGS (https://earthexplorer.usgs.gov/). Soil slope was calculated based on this DEM. Slope values less than 2% were considered flat areas according to guidelines of soil profile description [47].

Soil map (1:50,000, Figure 1c) and soil data (from 8 soil profiles) of the A and AB horizons with depth up to 50 cm were obtained at the Department of Agriculture and Rural Development of Dong Nai Province. Soil data used for analysis included soil type, soil texture, and soil organic matter.

Data of soil pH were estimated from the corrected spectral reflectance satellite image Sentinel-2 MSI on 27 February 2020, based on 31 field pH measurement positions. Sentinel-2 was obtained from the Copernicus Open Access Hub (https://scihub.copernicus.eu/). The multiple regression model for soil pH estimation is shown in Equation (4). This equation was based on the work of Ghazali et al. [48]. The correlation coefficient of this estimation was R² = 0.636 with a p-value = 0.05.

$$\mathrm{pH}=6.2062-47.1054\times B2-70.791\times B3+108.351\times B4+4.476\times B11-20.29\times B12$$

(4)

where B2, B3, B4, B11, and B12 are the spectral bands of the Sentinel-2 image.

2.3.2. Field Measurement Data

Interviews with farmers and agricultural experts were carried out on 7, 13, 20, 24 March, on 23 November 2019, and on 6 March 2020 using a directive survey protocol. There were 53 farmers and experts (30 in 2019 and 23 in 2020) interviewed. This work was conducted to obtain exhaustive information on the agriculture activities in this area.

Field measurement was conducted in March and May 2020. There were 31 positions for soil pH and soil moisture measurement using the instrument Takemura DM-15. The distribution of the measuring positions is shown in Figure 1d. The sampling locations were randomly selected, covering different types of crops, and depending on the accessibility of the sites. Among those positions, 21 soil samples (

Table 1. Soil characteristics at sampling positions. These data are used for maps validation.

Sample	Sampling Depth (cm)	Soil Type	Soil Grain Size				pH H2O	pH KCl	Al3+ (meq/100 g)	Total (%)
			0.2–2 mm	0.02–0.2 mm	0.002–0.02 mm	<0.002 mm				OM	Al
F1-10	10	Luvisols	1.514	17.49	13.576	67.419	5.2	3.26	4.72	3.945	0.473
F1-20	20	Luvisols						3.26	6.255	4.281	1.69
F1-50	50	Luvisols						3.65	1.217	2.887	9.587
F4-10	10	Luvisols	27.818	13.43	8.654	50.096	6.1	4.31	0.034	2.512	1.74
F4-20	20	Luvisols						4.42	0.016	2.407	1.119
F5-10	10	Luvisols	24.908	12.51	8.708	53.873	5.6	3.63	1.147	2.551	6.62
F5-20	20	Luvisols						3.6	1.019	1.558	7.193
F5-50	50	Luvisols						4.36	0.047	2.05	2.764
F6-10	10	Luvisols	56.628	1.82	1.509	40.039	5	4.73	0.005	5.911	2.192
F6-10-2	10	Luvisols						4.28	0.012	7.844	2.507
F6-20-2	20	Luvisols						4.75	0.001	6.528	1.182
F10-10	10	Luvisols	2.227	21.14	16.917	59.715	6.75	4.34	0.002	5.103	7.293
F10-20	20	Luvisols						4.23	0.005	5.036	3.756
F7-10	10	Ferralsols	41.394	18.82	13.737	26.017	5.6	3.9	0.951	1.785	9.276
F7-20	20	Ferralsols						3.9	0.786	2.593	12.724
F7-50	50	Ferralsols						3.86	1.426	3.317	12.384
F3-10	10	Ferralsols	1.322	34.56	11.499	52.618	6.25	4.46	0.001	2.947	0.613
F3-20	20	Ferralsols						3.09	5.49	2.102	5.701
F3-50	50	Ferralsols						4.17	0.185	2.261	10.462
R-S1-10	10	Ferralsols	12.520	15.150	0.760	71.570	6.7	4.880	0.160	4.490	4.320
R-S2-10	10	Ferralsols	14.250	10.380	2.070	73.300	6.5	4.600	0.290	5.350	6.930
R-S3-10	10	Ferralsols	14.510	18.500	0.300	66.690	6.4	4.630	0.160	4.280	6.910

) were taken at 0–10 cm, 10–20 cm, and 20–50 cm depth at 7 positions. The coordinates of each sample were recorded by the global positioning system (Garmin GPSMAP 64s) instrument. Soil samples were kept in polyethylene bags and transported to the laboratory at Ho Chi Minh City Institute of Resource Geography (Vietnam Academy of Science and Technology, Vietnam) for analysis. The particle size distribution was carried out using the standard pipette method [49]. The pH of each soil sample was measured in 1 M KCl solution (pHKCl) at ratio of 1:5 following the method described in Al-Busaidi et al. [50]. Soil organic matter (OM) was determined by the Walkley–Black procedure [51]. Availability of Al in soils was extracted using 1 M KCl solution. A 1:2 soil to solution ratio was shaken for 1 h on an end-over-end shaker at 50 rpm, then centrifuged at 2000 rpm for 15 min. The Al concentration in all of the extracts was then determined by inductively coupled plasma atomic emission spectrometry (ICP-OES) [52].

2.4. Sensitivity Analysis and Map Validation

The model performance was checked by two types of information: the first one for the AHP based on CR with the inclusion of sensitivity analysis. Weight values with low CR (CR < 0.1) were chosen. Sensitivity analysis was conducted to determine how changes of each factor’s weight affect the resultant maps [53]. Application of sensitivity analysis was performed to discover the range of the input parameters values of the model, which maintain the stability of the model outputs. In this work, attention was focused on the stability of the capability ranking levels to any changes in criteria weights. New weights were generated for each factor by adding and subtracting 1% from the original weight, proceeding up to 25%. The weights of the unadjusted factors were also changed in order to keep all weights in the original ratio proportion and the weights were summed to one. A series of evaluation runs was conducted. In total, there were 400 evaluation runs, where each run generated a single new soil capability classification map. In the second step, the resultant maps were validated using field sampling data (Table 1) to evaluate if the models explain the data.

3. Results

3.1. Calculation of Soil Capability in Retaining and Transporting Al³⁺

3.1.1. Factors Standardization and Criteria Weights

The standardization of soil factors capability is shown in

Table 2. Selected influencing factors associated with the significance level for the classification of soil capability in retaining or transporting Al³⁺.

Criteria	Factors	Standardized Value				References
		1	2	3	4
Soil Al3+ retention capability	Soil pH	pH ≥ 6	5 < pH < 6	4.5 < pH ≤ 5.0	3 < pH ≤ 4.5	Data from this study, [54,55,56,57]
	Ferrallitisation	Luvisols, Gleysols	Acrisols	Ferralsols	Ferralsols	Soil experts’ opinions
	Organic matter	OM < 2%	2% < OM < 3%	3% < OM < 5%	OM > 5%	Data from this study, [3,40,43]
	Soil texture	Sand	Loam	Clay loam	Clay	[9,40]
Soil Al3+ transportation capability	Soil pH	pH ≥ 6	5 < pH < 6	4.5 < pH ≤ 5.0	3 < pH ≤ 4.5	Soil experts’ opinions
	Soil texture	Clay	Clay loam	Loam	Sand	Soil experts’ opinions, [9,40]
	Soil slope	0°	0–10°	>10°		Soil experts’ opinions, [30,58,59,60,61]
	Organic matter	OM >5%	3% < OM < 5%	2% < OM < 3%	OM < 2%	Data from this study, [43]

. The qualitative factors (ferrallitisation and soil texture) were standardized by assigning a standardized value to each subclass. The standardization of the quantitative factors (OM, pH, and soil slope) was performed by the selection of the value interval to the standardized values. The range of values of the factors pH and OM was decided based on their relationships with Al³⁺. The solubility of Al increases exponentially with the decrease in pH and OM (Figure 3a,b). The amount of Al adsorbed in the rich OM soil is dependent on pH because the dissociation of functional groups in organic matter is pH dependent [43,44]. Low pH soils that have high OM content will increase the possibility of releasing Al³⁺.

The capability levels of each factor were standardized, ranging from one to four. Level one denotes the lowest and four denotes the highest capability. Importance levels of soil factors which influence the fate of Al³⁺ in the soil are shown in Table 2.

The pairwise comparison matrix of the five factors of slope, soil texture, pH, organic matter, and ferrallitisation and their weights on the soil capabilities to adsorb, retain, and to transport Al³⁺ are shown in

Table 3. Pairwise comparison matrix of influencing factors and their weights.

Retention Capability
	Soil pH	Ferrallitisation	Organic Matter	Soil Texture	Weight
Soil pH	1	3	4	1	0.4
Ferrallitisation	1/3	1	2	1	0.2
Organic matter	1/4	1/2	1	1/3	0.1
Soil texture	1	1	3	1	0.3
SUM	2.583	5.5	10	3.333	1
Transportation Capability
	Soil pH	Soil texture	Soil slope	Organic matter	Weight
Soil pH	1	1/5	1/2	2	0.12
Soil texture	5	1	3	3	0.52
Soil slope	2	1/3	1	6	0.28
Organic matter	1/2	1/3	1/6	1	0.08
SUM	8.5	1.867	4.667	12	1

.

For the Al³⁺ retention capability, soil pH and soil texture are the highest weighted influencing factors (their weight values are 0.4 and 0.3, respectively) while organic matter receives the lowest weight (weight value of 0.1). This means that pH and soil texture have a significant influence on the Al³⁺ retention capability of the soil.

The factors of soil texture and soil slope have the highest weight values for the soil capability in transporting Al³⁺ (0.52 and 0.28, respectively), while the soil pH and organic matter have the lowest importance levels (their weight values are 0.12 and 0.08, respectively). This result is expected because soil texture (proportion of sand, silt, and clay-sized particles) influences the soil infiltration rate, while slope determines the thickness of the soil layer and drives the movement of soil water and substances.

3.1.2. Sensitivity analysis

The results of the sensitivity analysis (50 runs for each criterion) are summarized in Table 4 and Figure 4 and Figure 5. Table 4 summarizes the range of criteria weights, which keep the models stable (the ranking classes S1 to S4 are unchanged compared to the base run). The analysis of the sensitivity criteria shows the robustness of the results. The results from 400 simulation runs (200 runs for simulations of factor weights for soil retention capability and 200 runs for soil transportation capability) show that:

• For soil retention capability, the factors soil pH and soil texture are more sensitive than the factors of soil weathering and organic matter (cf. Table 4 and Figure 4). Soil pH and soil texture cause significant capability classes modification from the base map (which used weight values from Table 3); the number of cells changed (cell size 100 m²) from the base map are 650,708, 195,402, 846,259, and 149 for S1, S2, S3, and S4 when their weights decreased or increased outside the range (−18%, +23%) for pH factor, and the differences amount to 673,378, 443,563, 814,192, and 302,723 cells when their weights decreased or increased outside the range (−16%, +23%) for the factor soil texture.
• For the soil capability in transporting Al³⁺, the soil texture and slope are the most sensitive factors; the weight of the factor soil texture causes significant changes to the soil transportation levels (the number of changed cells reached up to 733,149, 567,986, 166,706, and 116) compared to the base map, when its weight decreased or increased outside the range (−2%, +16%) from the base weight value. For the soil slope, the changed cells reached up to 733,037, 568,009, 166,582, and 116 when its weight decreased or increased outside the range (−10%, +21%) from the base weight value. The factors pH and organic matter in turn also made changes to the soil transportation levels when their weights fell outside the ranges (−25%, +4%) for pH and (−17%, +25%) for organic matter, but the cell differences were not so large (cf. Figure 5).

Figure 4. Soil retention capability: Cell differences compared to the base model. Summary results from 200 simulations (50 runs for each criterion). Factor 1: soil pH, factor 2: ferrallitisation, factor 3: organic matter, factor 4: soil texture. S1: no capability for Al³⁺ retention in the soil, S2: little capability, S3: moderate capability, and S4: strong capability.

Figure 4. Soil retention capability: Cell differences compared to the base model. Summary results from 200 simulations (50 runs for each criterion). Factor 1: soil pH, factor 2: ferrallitisation, factor 3: organic matter, factor 4: soil texture. S1: no capability for Al³⁺ retention in the soil, S2: little capability, S3: moderate capability, and S4: strong capability.

Figure 5. Soil transportation capability: Cell differences compared to the base model. Summary results from 200 simulations (50 runs for each criterion). Factor 1: soil pH, factor 2: soil texture, factor 3: soil slope, factor 4: organic matter. S1: no capability for Al³⁺ transport in the soil, S2: little capability, S3: moderate capability, and S4: strong capability.

Figure 5. Soil transportation capability: Cell differences compared to the base model. Summary results from 200 simulations (50 runs for each criterion). Factor 1: soil pH, factor 2: soil texture, factor 3: soil slope, factor 4: organic matter. S1: no capability for Al³⁺ transport in the soil, S2: little capability, S3: moderate capability, and S4: strong capability.

Table 4. Summary of ranges of weight values of each criterion from the 50 sensitivity analysis simulation runs compared to the base run.

	Changed Factor	Percent Change (%) from Base Value Which Keeps the Model Stable	Range of Weight Values Which Keeps the Model Stable	Range of Weight Values Which Keeps the Model Stable and Satisfies CR < 0.1
Retention Capability	pH	(−18, +23)	0.376–0.48	0.376–0.4
	Ferrallitisation	(−25, +23)	0.16–0.24	0.17–0.22
	Organic matter	(−25, +23)	0.073–0.12	0.1–0.101
	Soil texture	(−15, +23)	0.252–0.36	0.303–0.342
Transportation Capability	pH	(−25, +3)	0.096–0.137	0.1056–0.1248
	Soil texture	(−2, +16)	0.501–0.6084	0.5096–0.5616
	Slope	(−10, + 21)	0.229–0.336	0.28–0.266
	Organic matter	(−17, +25)	0.05–0.097	0.074–0.0808

Table 4. Summary of ranges of weight values of each criterion from the 50 sensitivity analysis simulation runs compared to the base run.

	Changed Factor	Percent Change (%) from Base Value Which Keeps the Model Stable	Range of Weight Values Which Keeps the Model Stable	Range of Weight Values Which Keeps the Model Stable and Satisfies CR < 0.1
Retention Capability	pH	(−18, +23)	0.376–0.48	0.376–0.4
	Ferrallitisation	(−25, +23)	0.16–0.24	0.17–0.22
	Organic matter	(−25, +23)	0.073–0.12	0.1–0.101
	Soil texture	(−15, +23)	0.252–0.36	0.303–0.342
Transportation Capability	pH	(−25, +3)	0.096–0.137	0.1056–0.1248
	Soil texture	(−2, +16)	0.501–0.6084	0.5096–0.5616
	Slope	(−10, + 21)	0.229–0.336	0.28–0.266
	Organic matter	(−17, +25)	0.05–0.097	0.074–0.0808

The S4 (strong capability) in the soil retention capability map is stable in all simulations despite a certain degree of variations in the weights of the factors (Figure 4). In the map of soil transportation capability, S4 remains stable when the factors of soil texture and soil slope vary (Figure 5). By varying the weights of the factors of soil pH and organic matter to a certain extent, S4 changed compared to the base map (and also to the other capability levels, S1–3). This reveals that the degree of domination of the S4 cells is independent of changes in the decision weights for the soil retention capability, and that the S4 level fluctuates for the soil transportation capability when changing the weights of the factors pH and OM (Figure 5).

S1, S2, and S3 in the soil retention capability map are rather stable to variations in the weights of the soil factors of weathering and OM. The differences of most cells from the base run happen when the weights of the factors of weathering and OM are increased to more than 24% from the base values. For the runs for the soil transportation capability map, the number of cells in S1, S2, and S3 changed for all factor weights variation (Figure 5). In general, the perturbation of the decision weights has a small impact on the ranking of all levels from S1 to S4 in the map of soil retention capability, but this perturbation has a large impact on the levels of soil transportation capability.

3.2. The Maps of Soil Al³⁺ Retention and Transportation Capabilities

Based on the calculated weights in Table 3, the ranking criterion layers in Figure 6 and Figure 7 were aggregated. The integration of these criterion layers gives the resultant maps, as is shown in Figure 8. These two maps show the spatial patterns and distribution of the soil capability classes of Al³⁺ retention and transportation.

The information given by the two resultant maps show that most of the soils in the area have medium to high capability of Al³⁺ retention (99% of the area, Figure 8a), and about 65% of the soils have medium to high capability of transporting Al³⁺ (Figure 8b). For the agricultural land, about 65% of the land is ranked from a high to a very high soil Al³⁺ retention capability, and about 58% of the land is ranked from a medium to a high capability of transporting Al³⁺ (

Table 5. Total area for each level of soil retention and transportation capability in Trang Bom.

Soil reSoil Retention Capability of Al3+
	Total area		Agricultural area
	Area (ha)	Area (%)	Area (ha)	Area (%)
S1	212.5	0.68	53.93	0.28
S2	11,566.13	37.13	6761.58	35.65
S3	19,370.27	62.18	12,150.27	64.06
S4	4.21	0.014	1.7	0.009
SUM	31,153.11	100	18,967.48	100
Soil Transportation Capability of Al3+
	Total area		Agricultural area
	Area (ha)	Area (%)	Area (ha)	Area (%)
S1	9331.471	29.98	7969.81	42.03
S2	17,695.52	56.85	9425.94	49.71
S3	4101.4	13.18	1565.03	8.25
SUM	31,128.39	100	18,960.78	100

and Figure 8c,d).

For soil retention capability, the S2 (medium capability) and S3 (high capability) occupy most of the area (~99% of total area). They are expressed in yellow and orange in Figure 8a. The soil characteristics of these classes include soil pH from 4.5 to 6, loam and clay loam soils, an OM content from 2 to 5%, and the soil has a medium weathering process (Acrisols). The S1 (no capability for Al³⁺ retention in the soil) is small and is found mainly in the south-west of the area (212.5 ha, accounting for 0.68% of the total area). The S1 level has soil pH > 6, is mostly sandy soil with OM content < 2%, and has a low weathering process (Luvisols and Gleysols) (Table 2 and Figure 8a). The S4 level appears scattered with very small areas, accounting for 0.014% of the total area. This area is mostly occupied in the center with soil pH < 4.5, a high clay content, a high content of OM (>5%), and the soil has a strong weathering process (Ferralsols). For the map of soil capability in transporting Al³⁺ (Figure 8b), the area of S1 (no capability) occupies 9331.47 ha, accounting for 29.98% of the area. This area is distributed from the south-east to the north-east of Trang Bom with the factors of soil pH > 6, sandy soil, soil slope < 2°, and OM content > 5% (Table 2). The medium level S2 occupies 56.85% of the total area, and is distributed from the center to the west and one part from the south-east to the north-east of the area. The soil pH is from 5 to 6, the soil is clay loam soil, the soil slope ranges from 2° to 10°, and the content of the OM is from 3 to 5%. The soil that has a high capacity in transporting Al³⁺ (S3) occupies the area of 4101 ha, accounting for 13.18% of the area. This type of soil is mostly distributed in the south and south-west of the area. The soil pH is from 4.5 to 5, it is clay soil with a slope > 10°, and an OM content of < 2%. Table 5 shows the areas of each capability level.

3.3. Map Validation

We used our sampling data (21 samples at 7 positions and at 3 soil depths of 0–10 cm, 10–20 cm, and 20–50 cm) to evaluate the information given by these maps. The results are shown in Figure 9. The concentrations of Al³⁺ were in the range of fair agreement (R² = 0.670) with capability values predicted by the maps. For the soil retention levels, the Al³⁺ concentration is less in the low capability level (level 2) than those in the high capability level (level 3, cf. Figure 9a). A converse effect can be found in the map of the soil transportation capability; the Al³⁺ concentration is lower in the high transportation capability of soil than those at low capability (Figure 9b). The positions F5 to F7 are in the high level of soil transportation capability, thus they have lower Al³⁺ concentrations even though they are located on a high level of soil retention capability. This is because Al³⁺ is transported to adjacent areas. The converse effect can be observed at the positions F1 and F3 (Figure 9). The calculation of the soil samples’ properties, as is shown in Table 1 using Equation (3), shows that they fit into the capability levels compared to the levels from the maps (R² = 0.9), which were built based on collected data. It presents that our resultant maps performed well in classifying Al³⁺ retention and the transportation capabilities of the soil in the study area.

4. Discussion

Information given from the two maps shows that this study area is characterized by 99% of the soils having a medium to high capability of Al³⁺ retention and about 65% have a capability of transporting Al³⁺. On the land for agricultural use, 65% and 58% of the total land belong to medium and high capabilities of retaining and transporting Al³⁺, respectively. With this complexity of the agricultural soil matrix, the managers or landowners should identify necessary control measures, especially the usage of plant protection chemicals. One should limit the activities of Al³⁺ replenishment or the activities which trigger the release of Al³⁺ from available Al compounds in the soil. Because the sources that put Al³⁺ in the soils of this area are mostly from agricultural activities, the use of fungicides or pesticides containing Al³⁺ should be carefully controlled, especially for agricultural lands in the northern and southern areas of Trang Bom (Figure 8). Al³⁺ retained excessively in the soil can be transported to neighboring areas, causing negative effects to the aquatic ecosystems in the north and the residential areas in the south (Figure 8c,d). Al³⁺ can also be transported to aquifers, affecting the groundwater quality in these areas.

Growing anthropogenic pressures continue to intensify and affect natural and semi-natural (agricultural areas) environments by putting in lots of different kinds of chemical products which might lead to pollution. Different soil conditions have different abilities to accumulate and retain a certain substance. Thus, the assessment of the levels of soil capabilities to retain and transport chemical substances, in order to be able to position and estimate the how hazardous the soil is, is very important.

Previous studies usually focused on building maps of agricultural land suitability, such as the determination of suitable sites for agricultural use [62,63,64,65]. Some authors used soil characteristics and slope conditions in combination with information extracted from remote sensing data to evaluate the agricultural support plans, such as the work of Zolekar and Bhagat [66], or the work of Ostovari, Honarbakhsh, Sangoony, Zolfaghari, Maleki, and Ingram [30] with the assessment of land suitability for rapeseed production. Other researches focused on building maps for pollution risk assessment [67,68]. For the studies on pollution risk assessment, the authors assessed the spatial patterns of pollutants and evaluated the pollution risk by measuring their concentrations in the soils and spatially zoning them [69,70,71]. The innovation of the present study compared with other studies is that it focused on the assessment of soil capability in retaining and transporting Al³⁺ based on some soil properties (pH, OM, soil texture, weathering conditions, and soil slope). This approach can also be applied to assess other toxic substances or pesticides.

The resultant maps of soil capabilities for Al³⁺ retention and transportation (Figure 8) demonstrate the effectiveness of the GIS-based AHP method. This method is rather simple but provides insights into spatial patterns through the characterized factors. Many studies with a variety of disciplines have used the GIS-based AHP method. For assessing the fate of substances (pollutants) in the soil environment, modelling is another effective tool which can predict the transportation and reaction of these substances. However, the application of these models is quite complex and requires lots of data and parameters input. Moreover, they need a certain level of expert knowledge, thus leading to difficulties in application, especially for the farmers. From this perspective, our created maps show their effectiveness. These maps can contribute to the efforts to effectively control soil pollution and monitor soil quality in an easy-to-use way. Based on information from these maps, we can design experiments to quantify the tolerance and thresholds of Al³⁺ on each soil zone to delineate the areas that allow for more or less chemical usage. These maps are useful tools for management at the landscape scale, for quickly and effectively making decisions without spending too much time on data collection, model running, and analyzing the results.

5. Conclusions

In this study, the soil capabilities for the retention and transportation of Al³⁺ were determined. The method of GIS-based AHP was applied. These created maps provide useful information and a scientific basis for farmers and managers to develop the framework on the usage of chemical products containing Al³⁺ to prevent soil degradation and other environmental consequences caused by excessive Al³⁺ concentration. To this end, these maps can be used for (i) managing and controlling the dosage and frequency of the application of pesticides containing Al³⁺ in each soil zone, (ii) determining solutions to renovate and reclaim Al³⁺ pollution in the soil, corresponding to each soil zone condition, and (iii) estimating the extent of Al³⁺ contamination in different areas and prioritizing the areas that need reclamation.

Over and inappropriate usage of chemical plant protection products are the major reasons for agricultural soil degradation. Thus, from the results of our study, we infer that it is important to invest in the prior characterization of soil quality, for instance the soil suitability in accumulating Al³⁺, in order to well manage the soil quality. Reliable and accurate soil quality assessment is therefore indispensable to any practical soil management for a sustainable agriculture.