Introduction

Evaluating land suitability of agricultural lands plays an important role to protect agricultural land resources, manage, and create a sustainable agriculture system (Mendas and Delali 2012; Kazemi et al. 2016). A land suitability evaluation for a specific crop is a complex process. In this way, environmental and agro-ecological factors, such as topo-position, soil properties, and climate data, are the main factors that affect land suitability evaluation (Falasca et al. 2012; Baroudy 2016). Land suitability at the field scale changes in each part of a local area because of variation in its topo-positions and soil properties. Hence, this is necessary to evaluate, classify, and manage land units for improving land productivity based on local potentials and limitations (Food and Agriculture Organization (FAO) 1990). Increasing population and industrial revolution in Iran have gradually caused degradation of land resources. Therefore, for ensuring sustainability in land productivity and food supplies, attention has been increased on the local potentials and limitations of the land for crop production, specifically in calcareous, saline, and sodic soils in southwest Iran.

Land suitability assessment enables farmers and managers to make the best decision for the land-use planning and crop management system (Recatal and Zinck 2008; Fontes et al. 2009). The high number of environmental and agro-ecological factors related to assessing land suitability could create a power struggle over using long-term sustainability in agriculture (Bandyopadhyay et al. 2009; Akinci et al. 2013). Hence, some methods have been used to create a land suitability evaluation map. The parametric method (Sys et al. 1991) and a framework presented by FAO (1976) were widely used dealing with soil properties, terrain, and climate data. In this way, the analytical hierarchy process (AHP) incorporates multicriteria factors that give different dimensions of scores for land suitability assessments (Banai 2005). One of the best methods for controlling the heterogeneous factors and for representing the relationships between agro-ecological and environmental factors in a hierarchic structure is the AHP method suggested by Saaty (1980). In addition, the integration of the AHP method with geographic information system (GIS) presents a new approach for land suitability evaluation (Akinci et al. 2013; Ostvari et al. 2019). Generally, GIS is a valuable tool for computing and storing the extremely large amount of data needed to map the land suitability for special crop farming. Moreover, spatial analysis methods in GIS aid to determine the relative importance of multicriteria factors, which are considered important roles in land suitability assessment (Banai 2005).

This AHP algorithm–GIS approach has various advantages, such as the ability to develop a weighting scheme, and flexibility to change different criteria and AHP hierarchical structures. Recently, several studies have applied the GIS technique and AHP methods to analyze the land suitability. For example, Akinci et al. (2013) evaluated the land suitability for the agriculture productions in Yusufeli district in Turkey by applying AHP in GIS. They indicated that the soil depth is inadequate for agricultural farming and the erosion rate is too high. Zolekar and Bhagat (2015) developed a GIS-based multicriterion decision-making approach framework for agriculture in the hilly zone by using slope degree, soil data, and IRS P6 LISS-IV dataset. Kazemi et al. (2016) assessed the land suitability for rain-fed Faba-bean farming in Golestan province, north of Iran by using a combined GIS–AHP method. They indicated that climate, topography, and soil data were the most important factors for rain-fed faba-bean production. Ostvari et al. (2019) used a combined AHP–GIS to assess the land suitability for rapeseed production with considering climate, topography, soil, and remote-sensing data. They highlighted that soil properties had the highest specific weight, indicating the most effective parameters.

Fuzzy set is a quantitative method for determining with vagueness in a complicated system (Zadeh 1965). The fuzzy set was applied for the land suitability assessment studies (Banai 1993; Sicat et al. 2005; Keshavarzi 2010). Banai (1993) expressed that a fuzzy-set methodology has some advantages: (i) permits flexibility in defining the boundary of the object in the set, (ii) a degree of closeness to the ideal point is considered in the inclusion, and (iii) account of partial membership of an element in a set. Sicat et al. (2005) applied a fuzzy modeling along with the farmers’ knowledge to assess land suitability maps for agriculture productions in Nizamabad region in India. Keshavarzi (2010) used fuzzy-set theory for land suitability evaluation in Ziaran region, Iran. They showed that Fuzzy-set methodology could continually indicate the land continuity in various land classes. Wu et al. (2019) applied a fuzzy logic model for assessing soil quality and evaluating of the spatial distribution in land-use types in the Yellow River Delta. Their results showed that the use of land resources in the study site should be reasonably planned. The AHP or fuzzy-set methods themselves provided a particularly weak result in controlling weights of land features and calculating the land suitability value. Hence, Keshavarzi (2010) and Zhang et al. (2015) employed AHP for determining weights of land feature for developing fuzzy membership functions. Therefore, the integrated method of AHP, fuzzy set, and GIS techniques could be a strong method to increase the accuracy of land suitability evaluation for a particular crop production.

Wheat occupies ~6.0 million ha of the total irrigated agricultural lands worldwide (FAO 2005). Cereal grains (wheat, corn, and rice) are the most important food source for human consumption (about 68%) worldwide. Wheat is a cereal crop and is grown widely throughout the world in a range of agro-ecological and environmental conditions (USDA 2016). The major staple field crop in Iran is wheat, including bread and durum wheat (Triticum aestivum and T. turgidum); both types are planted either in winter or in spring. It is predicted that about 2/3 of all wheat production in Iran is used for human food. However, wheat farming needs to adapt to local conditions by considering soils (calcareous, saline, and sodic soils), climate, and topography. The calcareous, saline, and sodic soils are known for their high calcium carbonate equivalent and pH (pH > 7.5), high electrical conductivity (EC > 4 dS m−1), and high sodium (ESP > 15), respectively (Brady and Weil 2002). The aims of the current study were to generate a Geostatistical–AHP–Fuzzy set in GIS techniques to determine the land suitability evaluation for wheat farming in the calcareous, saline, and sodic soils of Marvdasht Plain located in Fars Province, southern Iran, which is famous for being the highest wheat-producing region in Iran, and is dominated by the wheat–corn–fallow-based cropping rotational systems.

Materials and Methods

Study Area

This study was conducted on the Marvdasht plain in the northeast of Fars province, Iran (Fig. 1). The study area coordinates, with an area of ~51,831 ha, range from 29° 44′ to 30° 26′ N latitudes and from 52° 17′ to 53° 30′ E longitudes (Fig. 1). Wheat, maize, and vegetables are the main cultivated crops in this region, although wheat farming covers about 70%. The most common soil orders are Inceptisols, Entisols, and Aridisols based on the Soil Taxonomy (USDA 2010).

Fig. 1
figure 1

Location of study area in the Marvdasht plain, northeast Fars province in Iran (a), distribution of sample points (b), and average monthly precipitation and temperature (from 2001 to 2018) (c)

Full size image

Climate Factors

The average monthly precipitation and temperature distribution in the period 2001–2018 are shown in Fig. 1. The climate of the study area is semiarid, and the annual average precipitation and temperature are 291.7 mm and 17.5 °C, respectively.

Topography Factors

The DEM (digital elevation model) of the study area (with a cell size of 10 × 10 m) was obtained from the agricultural research service of Fars province. The DEM data were used to derive the elevation and slope factors. Elevation in the study region varies from 1574 to 3121 m above sea level (Fig. 2). Most of the environmental data, such as soil water content, precipitation, radiation and temperature, vary in line with the elevation factor. Elevation from sea level plays a crucial role in the crop yield, growth, and distribution. About 80,781 ha (40.8%), 18,658 ha (9.4%), 9353 ha (4.7%), 6231 ha (3.1%), and 82,900 ha (41.9%) of the study region were located in the slope classes of 0–2, 2–6, 6–12, 12–16, and >16%, respectively (Fig. 2). Generally, low-slope land is more suitable for wheat farming (Fu et al. 2011).

Fig. 2
figure 2

Topographic maps, including elevation and slope maps of the study area

Full size image

Soil Sampling and Analysis

A stratified random sampling method was used to select the positions of 65 cropland soil samples, taken from a depth of 0–30 cm (Fig. 1). The soil samples were then air-dried and passed through a 2-mm sieve. Soil particle size, including clay (<0.002 mm), silt (0.05–0.002 mm), and sand (0.05–2 mm) content, was determined using the hydrometer method (Gee and Bauder 1986). The soil pH and electrical conductivity were determined in the extracted saturated soil (Saturation solution 1:5). The optimum pH value for crop growth is between 6.5 and 7.0 (Thompson and Troeh 1973). The ESP (exchangeable sodium percentage) parameter was determined in the soil solution by the ratio of sodium content to the sum of sodium and other cations, such as calcium, magnesium, and potassium (Brady and Weil 2002). The CCE (calcium carbonate equivalent) parameter was measured by the back-titration methods (Nelson and Sommers 1986).

Geostatistical Analysis

Soil parameter thematic maps were obtained using the ordinary kriging (OK) interpolation method in ArcGIS v10.3. The empirical semi-variogram was generated by Eq. (1) proposed by Matheron (1963) (Webster and Oliver 2001)

$${\mathrm{\Upsilon }}\left( {\mathit{h}} \right) = \frac{1}{{2{\mathit{N}}\left( {\mathit{h}} \right)}}\mathop {\sum}\limits_{{\mathrm{i}} = 1}^{{\mathit{N}}\left( {\mathit{h}} \right)} {\left[ {{\mathit{Z}}\left( {{\mathit{x}}_{\mathrm{i}}} \right) - {\mathit{Z}}\left( {{\mathit{x}}_{\mathrm{i}} + {\mathit{h}}} \right)} \right]^2},$$
(1)

where Z(xi), ϒ(h), and N(h) are measured soil properties at the location of xi, the variogram for a lag distance h between Z(xi) and Z(xi + h), and the number of data pairs, respectively. The semi-variogram is a geostatistical tool for visualizing, modeling, and interpreting the spatial dependence in a regionalized variable (Webster and Oliver 2001; Zekai 2009). The basic definition of the semi-variogram is the half squared-difference variation of the regionalized variable by distance, and is a practical measure of average spatial changes (Zekai 2009). Some important features were used to describe the variogram, including (i) Range (radius of influence), (ii) Sill (observed variance when the semi-variogram levels off at large distances), and (iii) Nugget (observed variance at zero distance). In addition, the semi-variogram generated just based on distance is called isotropic and, based on distance and orientation is called anisotropic. After that, the main theoretical (mathematical function) semi-variogram, such as linear, spherical, exponential, and Gaussian, is fitted to the experimental data (Zekai 2009). The semi-variogram is the source of information applied in kriging method to obtain an optimum weighting function.

The OK equation is given by the following equation (Webster and Oliver 2001):

$$\widehat {\mathrm{Z}}\left( {{\mathit{x}}_{\mathrm{i}}} \right) = \mathop {\sum}\limits_{{\mathrm{i}} = 1}^{\mathit{N}} {\lambda _{\mathrm{i}}{\mathrm{Z}}\left( {{\mathit{x}}_{\mathrm{i}}} \right)},$$
(2)

where λi and \(\widehat {\mathrm{Z}}\left( {{\mathit{x}}_{\mathrm{i}}} \right)\) are the weight of a specific point and predicted soil parameters at the selected points, respectively. OK is an exact interpolation procedure, which used weighted average of the measurement values to estimate soil attributes in unsampled points with a linear combination of surrounding measurements (Mirzaee et al. 2016). OK is a weighted linear combination of the measured data with the ME, equal to 0, and aims at minimizing the variance of the errors. The importance of the OK is that we do not have information about the ME and variance of the errors. Hence, the distinguishing feature of the OK method is that try to build the best model for minimizing the variance of the errors.

Land Suitability Assessment

The analysis of land suitability for wheat farming includes multicriteria decision-making by hierarchical structure (AHP algorithm), namely land suitability classification, selection of effective factors, assessment of the weights, and building a Fuzzy set–GIS-based model to create a land suitability map.

The diagram of the land suitability assessment procedure used in the present study is illustrated in Fig. 3. Scientific resource information was investigated and used to determine the agro-ecological and environmental data requirements for wheat farming (Table 1). In order to classify the data, knowledge of 25 local experts, having lot of experience and appropriate expertise in wheat farming, was taken into account. According to land characteristics, the framework for land evaluation suggested by FAO (United Nations FAO) was applied to classify land suitability map, i.e., S1 (highly suitable), S2 (moderately suitable), S3 (marginally suitable), and N (not suitable) based on the environmental and agro-ecological requirements for wheat farming (Table 1). In order to determine the most suitable land for wheat farming, some environmental and agro-ecological properties, such as topography, climate, and soil data, were used as presented in Fig. 3.

Fig. 3
figure 3

Hierarchy structure of evaluation factors for wheat farming

Full size image
Table 1 Criteria for delineating land suitability of wheat crop
Full size table

Fuzzy Logic presented by Zadeh (1965) is applied for modeling human knowledge and experience. As shown in Fig. 4, a fuzzy inference system has four significant sections, including a rule base, a fuzzifer, an inference engine, and a defuzzifer. A fuzzy set is the rule-based system and behaves as a system dynamic. The inference engine is used to make an inference from fuzzy rules (Keshavarzi 2010; Ostovari et al. 2016). Defuzzification is applied for mapping fuzzy data to output crisp factors. The fuzzy membership tool reclassifies or transforms the input data to a 0–1 scale based on the possibility of being a member of a specified set. In the present study, various membership functions, including sigmoid, inverted sigmoid, linear, and parabola type, were employed.

Fig. 4
figure 4

Fuzzy inference system

Full size image

For determining the weight of factors for land suitability evaluation, as aforementioned, the opinions of 25 local experts were considered. The weights of factors determined to generate multicriteria decision-making analysis data were in the range of 0–1 (Malczewski 1999; Girvan et al. 2003).

For the pairwise comparison of factors, the environmental and agro-ecological properties were ranked on a scale from 1 to 9 following the fundamental scale presented by Saaty (1980) (Table 2). In this phase, a square pairwise comparison matrix was developed in which all its elements are compared with each other (Saaty 1994). The pairwise comparison matrix A of n factors can be defined as follows:

$$A = \left[ {a_{ij}} \right]_{n \times n} = \left[ {\begin{array}{*{20}{c}} {a_{11}} & {a_{12}} & \cdots & \cdots & {a_{1n}} \\ {a_{21}} & {a_{22}} & \cdots & \cdots & {a_{2n}} \\ \vdots & \vdots & \cdots & \cdots & \vdots \\ \vdots & \vdots & \cdots & \cdots & \vdots \\ {a_{n1}} & {a_{n2}} & \cdots & \cdots & {a_{nn}} \end{array}} \right],$$
(3)

where A is the matrix with aij elements, i and j are the agro-ecological and environmental factors. This matrix can be mathematically expressed as in Eq. (4).

$$a_{ij} = 1/a_{ji}.$$
(4)
Table 2 The fundamental scale for pairwise comparison used for the AHP
Full size table

After designing all pairwise comparison matrices of factors, the vector of weights, w = [w1, w2, w3…wn] was calculated according to the Saaty’s eigenvector method.

In this way, the Consistency Index (CI) was used to determine consistency in A binary comparison matrix. The CI value is calculated by Eq. (5)

$${\mathrm{CI}} = \frac{{\lambda _{\mathrm{max}} - n}}{{n - 1}},$$
(5)

where λmax the maximum eigenvalue and n is the number of the binary comparison matrix.

Consistency ratio (CR) is calculated by Eq. (6).

$${\mathrm{CR}} = \frac{{\mathrm{CI}}}{{\mathrm{RI}}}.$$
(6)

If the A binary comparison matrix is not fully consistent, the value λmax will deviate from n and the other eigenvalues will deviate from zero. These deviations are determined by the CI given below. The RI is random index values that were used to calculate the CR. The CR ratio is a measure of consistency or inconsistency of the weights determined. The less than 0.10 shows a valid pairwise comparison matrix (Saaty 2008).

Land suitability assessment according to the AHP–GIS can be expressed by Eq. (7). The classes of the land suitability assessment and the weighted linear additive combination approach were calculated.

$$S = \mathop {\sum}\limits_{i = 1}^n {WiXi}$$
(7)

S is the land suitability classes (i.e., highly suitable, moderately suitable, marginally suitable, and not suitable, and the evaluation index with an effective factor on the suitability class and indicates the decision rule) (Zhang et al. 2013). The Wi, Xi, and n are the weight of factors (i.e., topography, climate, and soil data), the subfactor score of i factors, and the total number of factors. Finally, multicriteria decision-making analysis and the AHP method were used to generate a land suitability assessment map for wheat farming using the ArcGIS 10.0 software. The land suitability evaluation for the different factor weights was integrated in the GIS Model Builder. The ArcGIS system links the suitability of different layers to calculate the land suitability map at the study area for wheat farming.

Performance Evaluation

The cross-validation statistical procedure was applied to show the performance of models. The accuracy of the geostatistical method, i.e., OK in this study, was examined with the coefficient of determination (R2) (Eq. 8), mean error (ME) (Eq. 9), and root mean square error (RMSE) (Eq. 10)

$${\mathit{R}}^2 = 1 - \frac{{\mathop {\sum }\nolimits_{{\mathrm{i}} = 1}^{\mathit{N}} \left( {{\mathrm{Y}}_{\mathrm{i}} - \widehat {{\mathrm{Y}}_{\mathrm{i}}}} \right)^2}}{{\mathop {\sum }\nolimits_{{\mathrm{i}} = 1}^{\mathit{N}} \left( {{\mathrm{Y}}_{\mathrm{i}}^2} \right) - \frac{{\left( {\mathop {\sum }\nolimits_{{\mathrm{i}} = 1}^{\mathit{N}} {\mathrm{Y}}_{\mathrm{i}}} \right)^2}}{{\mathit{N}}}}},$$
(8)
$${\mathrm{ME}} = \frac{1}{{\mathit{N}}}\mathop {\sum}\limits_{{\mathrm{i}} = 1}^{\mathit{N}} {\left( {\widehat {{\mathrm{Y}}_{\mathrm{i}}} - {\mathrm{Y}}_{\mathrm{i}}} \right)},$$
(9)
$${\mathrm{RMSE}} = \left[ {\frac{{\mathop {\sum }\nolimits_{{\mathrm{i}} = 1}^{\mathrm{N}} \left( {\widehat {{\mathrm{Y}}_{\mathrm{i}}} - {\mathrm{Y}}_{\mathrm{i}}} \right)^2}}{{\mathit{N}}}} \right]^{0.5},$$
(10)

where N is the number of data, Yi is measured, and \(\widehat {{\mathrm{Y}}_{\mathrm{i}}}\) is estimated data.

Results

Soil Characteristics

A summary statistic of soil properties, including clay, sand, silt, CCE (calcium carbonate equivalent), pH, EC (electrical conductivity), and ESP, is shown in Table 3. A negative skewness was observed in the soil properties for EC, ESP, and the silt parameters. Whereas, a slight positive skewness was found for sand, clay, pH, and the CCE parameters. However, the statistical analyses showed that all of the soil parameters were distributed normally (p < 5%) when the Kolmogorov–Smirnov test was applied (Cressie 1993). In addition, a CV (Coefficient of variation) test with a categorization suggested by Wilding (1985) revealed that soil properties having a CV > 35% are considered as high-variability factors, between CV > 15 and CV < 35% as moderate-variability factors, and CV < 15% as low-variability factors. According to the Wilding (1985) categorization, sand, pH, and silt indicated a low variability; CCE and clay indicated moderate variability and ESP and EC factors indicated a high variability. However, a soil factor with a broad range of data could be a relevant factor for modeling land suitability for wheat farming.

Table 3 Descriptive statistics of study soil
Full size table

Due to the existence of calcareous formation in the study area, geomorphological conditions have the highest influence on various soil characteristics, such as soil texture, OM, pH, and accumulated calcium carbonate in the surface horizon. The diagnostic subsurface horizons, Cambic and Calcic with high amounts of calcium carbonate, are the distinguishing features of the soils studied. In the study site, the most influential soil-farming processes are the movement and leaching of calcium carbonate in the soil profile, and the accumulating of organic matter in the surface horizon. Soil types in the lowland areas under agricultural activities are Aridsols and Inceptisols; however, the soils formed in the mountain and highland areas are Entisols with very fine soil layers.

The Geostatistical Analysis of Soil Attributes

Table 4 indicates the results and parameters of variography modeling for soil attributes. Based on the results, all soil attributes displayed an anisotropic semi-variogram at the studied region, except silt content. The Spherical and Gaussian variogram models matched well to different soil attributes (Table 4), being in line with the results by Pilevar et al. (2020), who used Spherical and Gaussian variograms for the geostatistical analysis of soil properties for land suitability assessment in Northeast Iran. In this study, the spatial dependency for soil attributes was identified through the C0/silt ratio (nugget-to-silt ratio). According to the C0/silt ratio, a moderate spatial dependency (i.e., 0.292–0.604) was observed for all soil attributes (Table 4) (Zhu and Lin 2010; Mirzaee et al. 2016; Mirzaee et al. 2020). Moreover, the ratio of major to minor range is applied as the parameter k (anisotropic ratio). For all soil features, the k parameter was greater than one (Table 4). In addition, the results of OK methodology for mapping soil attributes are given in Table 5. Several studies were reported that the OK method is the best method for mapping studied soil features (Eldeiry and Garcia 2010; Li 2010; Piccini et al. 2014). The estimated maps for all soil attributes are demonstrated in Fig. 5.

Table 4 Semi-variograms of soil property data
Full size table
Table 5 The performance of ordinary kriging methods for mapping of soil properties
Full size table
Fig. 5
figure 5

Soil properties’ map such as pH, EC, CCE, ESP, clay, and sand parameters, and land suitability map for wheat farming in the Marvdasht plain

Full size image

Land Suitability Assessment

First, climate parameters, soil, and topography information transformed to fuzzy-set data with a numerical value between 0 and 1 based on the different fuzzy membership function (Table 6). As aforementioned, the local experts’ opinions were used as the basis to determine the weights of the effective factors. The impact of each environmental and agro-ecological data inclusive is shown in Table 6. For applying the AHP method, determining the percentage of factor role for each of the parameters is the first step. The results showed that the main limiting factors for wheat farming in the Marvdasht plain are soil texture with weight of 0.207, pH with weight of 0.121, slope with weight of 0.120, EC with weight of 0.113, ESP with weight of 0.111, and CCE with weight of 0.105 for this study region, respectively (Table 6). As Table 6 shows, there is a high dependence of wheat farming on soil properties, such as soil texture and pH parameters.

Table 6 The results of AHP method in land suitability for wheat cropping
Full size table

Finally, the best land suitability maps by integrating AHP, fuzzy set, and GIS techniques were produced for wheat farming (Fig. 5). As shown in Fig. 5, the S1 class (highly suitable land) with an area of 48,306.6 ha (25.65%) is located at the southeast and central parts of the study site. This class is recognized using environmental and agro-ecological factors, such as annual average temperature (15–20 °C), annual minimum temperature (10–15 °C), annual maximum temperature (20–25 °C), slope <2%, elevation <1700 m, soil textures: loam, clay loam, sandy clay, and clay, pH 6.5–7.3, EC < 4 dS/m, CCE < 25%, and ESP < 15% (Table 1). The S2 class (moderately suitable land) is located mainly in the south and around the S1 class over all parts of the study region (Fig. 5). This class with a total area of 71,939.7 ha (38.2%) (Table 7) is identified by annual average temperature ranges of 12–15 °C in winter and 20–25 °C in summer with annual minimum temperature of 9 °C and annual maximum temperature of 26 °C, pH ranges 6.0–7.6, soil textures sandy loam and sandy clay loam, ESP 15–20%, CCE 25–40%, and EC 4–8 dS/m (Table 1).

Table 7 The distribution of land suitability analysis results for wheat farming in the Marvdasht plain
Full size table

The marginally suitable class (S3) covering 52,017.2 ha (27.63%) (Table 7), shown in Fig. 5, is located mostly in the northern areas of the study area. Class S3 characteristics include annual average temperature within the range of 10–12 °C in winter and 25–32 °C in summer, annual minimum temperature 6–8 °C, annual maximum temperature 27–32 °C, pH 5.0–6.0 and 7.6–8.5, soil textures loamy sand, silty loam, and silty clay loam, ESP 20–45%, CCE 40–60%, and EC 8–16 dS/m (Table 1). The land suitability analysis for wheat farming indicated that some small areas of the study area were classified as not-suitable land (N class), covering 8.52% of the area (16,042.4 ha) (Table 7). The N-class characteristics include annual average temperature <10 and >32 °C, annual minimum temperature <6 °C, annual maximum temperature >32 °C, pH <5.0 and >8.5, soil textures silty clay, sandy, and silty, ESP > 45%, CCE > 60%, and EC > 16 dS/m (Table 1). All the land suitability classes for wheat farming on the Marvdasht plain are indicated in Fig. 5.

Discussion

It is important to analyze suitable land for maximizing the yield of specific crops. Many studies investigated the land suitability for specific crops in the available literature (FAO 1976; Recatal and Zinck 2008; Fontes et al. 2009; Bandyopadhyay et al. 2009; Grassano et al. 2011; Akinci et al. 2013; Zolekar and Bhagat 2015; Kazemi et al. 2016). Crop production on irrigated land at field scale is mostly determined by soil properties, climatic data, topography factors, and site-specific management based on local expert opinion (Grassano et al. 2011). Therefore, it is an important issue to evaluate the potential of the Marvdasht plain for wheat farming in the presence of calcareous, saline, and sodic soil conditions. In the present study, soil properties, such as soil texture, pH, slope, EC, ESP, and CCE presented in section “The Geostatistical Analysis of Soil Attributes”, are significant for maximizing yield in wheat farming. These soil properties have been considered for land suitability modeling in Kurdistan province, Western Iran, by Taghizadeh-Mehrjardi et al. (2020) and in Khorasan Province, Eastern Iran, by Pilevar et al. (2020), where soils are calcareous and controlled by a high content of calcium carbonate.

Recently, a number of studies have used the AHP method to analyze land suitability at a hierarchical structure (Lai et al. 2002; Zhang et al. 2015; Kazemi et al. 2016, Seyedmohammadi et al. 2019; Ostovari et al. 2019; Pilevar et al. 2020; Mandal et al. 2020) and fuzzy set (Sicat et al. 2005; Nikseresht et al. 2019; Keshavarzi 2010; Hoseini 2019). In addition, the weighting of factors, as Table 6 shows, can be compared with other studies, such as by Kazemi et al. (2016), that reported on rain-fed faba bean. In the present studies for modeling land suitability for wheat farming under calcareous, saline, and sodic soil conditions, the soil factors are identified as the most important. However, according to the local experts’ opinions, the wheat yield in the Marvdasht plain is dependent on soil factors such as texture and pH. Ostovari et al. (2019), Mandal et al. (2020), and Pilevar et al. (2020) reported the importance of soil texture as the most influential soil property affecting land suitability for maize and wheat farming in calcareous soils of Iran. The higher relative impact of the auxiliary data extracted from the DEM for modeling the land suitability indicates the influence of topography, as this has significant effects on many aspects of the important processes for soil formation (Taghizadeh-Mehrjardi et al. 2020). The slope degree (Table 6) ranging between 0 and 100% was the second most influential factor in the study area, which is in line with the results by Ostovari et al. (2019) and Taghizadeh-Mehrjardi et al. (2020). This topography factor has a big impact on mechanization during agricultural practices. In addition, the slope degree has negative effects indirectly on soil properties’ crop yield. Elevation and slope are higher in the upper catchment and mountainous area, indicating lower suitability for wheat farming compared with other parts of the study area. Elevation mostly has significant influences on temperature and soil microbiological activities linked to soil productivity and plant growth. It is worth mentioning that due to decreasing temperatures along with the increase in elevation, the plant growth stages, particularly sowing and rosette for wheat, are negatively influenced. Furthermore, according to Seyedmohammadi et al. (2019) and Taghizadeh-Mehrjardi et al. (2020), the slope factor had the highest weight in assessing the land suitability for irrigated farming in calcareous soils of Northwestern Iran. Many studies have also highlighted the significant effects of topographical factors on soil properties, resulting in affecting suitability of agricultural lands. For instance, Nabiollahi et al. (2018) and Ostovari et al. (2016) showed that the soil loss in slopes > 10% was significantly higher than slopes <10%, resulting in lower soil quality in steeper slopes.

The innovation of the present study compared with other research studies, that focused on land suitability assessment, is the assessment of the unique influence of saline, sodic, and calcareous materials in the Marvdasht plain soils posing a serious problem for wheat farming. The range of calcium carbonate equivalent in the study area varied from 11.5 to 66.3% (Table 3). The presence of high calcium carbonate in a soil creates some high nutritional problems mainly for micronutrients, especially for zinc and iron elements. In these soils, the hydrolysis of carbonate ions in soil solution creates OH– and controls soil pH (Foth 1990). Seyedmohammadi et al. (2019) and Hoseini (2019) reported the significant impact of CaCO3− content on controlling the soil pH, resulting in affecting land suitability for maize and wheat production. However, according to Ostovari et al. (2016) and Ostovari et al. (2020), Ca2+ ion in CaCO3 plays a vital role as a binding agent among the soil particles for creating stable and resistant aggregates, resulting in not only decreasing soil erodibility and increasing soil permeability, but also sustaining soil productivity by preserving soil fertility. In addition, the salinity ranging from 0.04 to 12.7 dS/m was the third important soil property in the study area. The importance of soil salinity in land suitability modeling has been frequently reported in calcareous soil of Iran (Hoseini 2019; Ostovari et al. 2019; Pilivar et al. 2020). In particular, in calcareous and saline–sodic soils, the soil salinity condition affects the uptake of water, nutrient elements, the microbiological activity, crop growth at different stages, and yield of most crops, especially wheat (Brady and Weil 2002). The salt stress affects crop yield in three ways: (1) osmotically, (2) imbalance uptake of nutrient elements, and (3) uptake of Na+ and Cl− by the plant (Munns and Tester 2008; Boyer et al. 2008). Furthermore, sodic soil with high sodium levels negatively influences wheat growth (Brady and Weil 2002). This type of soil swells excessively and results in convex surfaces after irrigation or rainfall. For this reason, the air and water movement through these soils is very difficult (Brady and Weil 2002). Hence, it was concluded that the land suitability evaluation model for wheat farming is needed for calibrating in the variable field conditions, such as those found in the Marvdasht plain soils affected by saline, sodic, and calcareous materials.

Conclusion

In the current study, a Geostatistics–AHP–Fuzzy-set algorithm in GIS-based modeling approach in order to produce a land suitability map for wheat farming was used in Iran’s calcareous, saline, and sodic soils of the semiarid regions. Land suitability evaluation is an important process to maintain the sustainability of Marvdasht plain for wheat farming. The land suitability analysis by integrating Geostatistic, fuzzy set, GIS, and AHP techniques for sustainable wheat farming showed that 25.65% (48,306.6 ha), 38.2% (71,939.7 ha), 27.63% (52,017.2 ha), and 8.52% (16,042.4) were highly suitable (S1 class), moderately suitable (S2 class), marginally suitable (S3 class), and not-suitable (N class) class, respectively. Our findings could well be a guide for future land-use management to investigate the impact of soil and environmental data on crop yields in southwest Iran and other similar regions. Further work would focus on the quantifying of the potential and limitations for other crops, mainly cultivated and affected by saline, sodic, and calcareous soils in this region in Iran. We concluded that combining fuzzy–AHP–Geostatistics in GIS is an applicable and effective way for making a better decision regarding land-use planning for agricultural purposes.