Assessment of land suitability for agriculture production using hybrid approaches: trapezoidal fuzzy number analytic hierarchy process, remote sensing and GIS

1 Introduction

In recent years, climate change has significantly affected global agriculture productivity and food security, as reported by the IPCC (IPCC, 2014). This impact is particularly critical in developing countries, where it is essential to evaluate land suitability for different agriculture practices and to develop crop specific assessment indices (Sabljić et al., 2024a). One of the major challenges facing global agriculture production today is the decline in crop yields, largely driven by climate change and anthropogenic activities (Sabljić et al., 2026; Sabljić et al., 2024b). Assessing land suitability is a fundamental step in planning and implementing effective land use strategies for sustainable agriculture development (El Baroudy, 2016; AbdelRahman et al., 2025). Variations in soil properties, climatic conditions, and topography across regions highlight the need for site specific land suitability evaluations. The soil, climate, and terrain are essential for accurately determining the agriculture potential of land (FOTH (1991) (Ramamurthy et al., 2020).

In 2022, agriculture occupied approximately 60.05% of India’s total land area. In 2023, nearly 61% of the country’s population was engaged in agriculture activities (Ramamurthy et al., 2020). The study region alone supports about 210,300 farmers, including nearly 190,000 small and marginal farmers. Agriculture forms the economic backbone of the present study area, with around 70% of the population depending on farming and related activities for their livelihood. The present study area covers 443,741 ha, of which 166,843 ha are under cultivation.

There is no standard fixed criteria for assessing agriculture land suitability. Evaluating a range of physical factors, including temperature, precipitation, moisture availability, soil aeration, fertility, depth, texture, salinity, slope, flood risk, and accessibility (Wang, 1994). Similarly, parameters such as soil texture, moisture, consistency, pH, drainage, and organic content along with slope (Perveen et al., 2007). Incorporated factors such as land capability, soil depth, erosion, slope, water availability, climate, vegetation cover, and transportation access (Zeng et al., 2019). The land suitability analysis in Tabriz, Iran, using parameters including elevation, slope, aspect, soil fertility, pH, temperature, precipitation, groundwater, and other related physical and climatic characteristics (Feizizadeh and Blaschke, 2013).

Most of these studies highlight the importance of accurately identifying land suitability for agriculture use to support effective land management and environmental sustainability (Akpoti et al., 2019; Patel et al., 2026). However, this task remains complex, as it requires the integration of multiple physical, chemical, and climatic parameters (Zhang et al., 2015; Pramanik, 2016). While socio-economic factors can be modified through policy and management interventions, natural factors remain the primary focus in land suitability assessments.

The integration of remote sensing and GIS technologies has significantly improved the collection and analysis of spatial and temporal data over large geographic areas. These technologies enable rapid, efficient, and comprehensive evaluation for natural resource management (Shaw et al., 2025). GIS-based tools provide a strong framework for handling spatial datasets essential for sustainable agriculture. The GIS tools are particularly effective in managing and analyzing complex geospatial datasets, making them essential for sustainable land and resource management (Jothimani et al., 2021; Azadi et al., 2024; Sabljić et al., 2024a).

The combination of fuzzy set theory with the Analytic Hierarchy Process (AHP) provides a robust framework for assigning weights to different criteria based on their relative importance (Eskandari et al., 2022; Azadi et al., 2023b). This integrated approach, along with GIS, is widely applied in land suitability analysis due to its ability to handle uncertainty and complex decision-making processes (Sarkar et al., 2022).

GIS-based AHP and other Multi-Criteria Decision-Making (MCDM) techniques are widely used to identify and evaluate the key factors influencing land suitability (Rangzan et al., 2024; Baskaran et al., 2023). Furthermore, the integration of remote sensing, GIS, MCDM approaches, hierarchical analysis, fuzzy logic, Boolean logic, and weighted linear combination methods enhances the accuracy and effectiveness of land suitability mapping (Emadi et al., 2010; Mendas et al., 2014; Montgomery et al., 2016; Azadi et al., 2023a).

In the literature reviews on land suitability for agriculture production, many methods have been used, such as the AHP (Yaman and Mutlu, 2025; Hosen et al., 2025; Mercan and Acibuca, 2023; Sadiq et al., 2023; Günal et al., 2022; Aburas et al., 2017; Rukhsana and Molla, 2022; Upadhyay et al., 2022; Everest and Gür, 2022; Karakuş and Yıldız, 2022; Nguyen et al., 2020), triangular fuzzy numbers AHP (TFN-AHP) (Tashayo et al., 2020b; Vahidi et al., 2023; Jamil et al., 2018; Abimbola et al., 2021; Sarkar et al., 2022; Sengupta et al., 2022), fuzzy inference systems (Seyedmohammadi and Navidi, 2022), machine learning (Chaibi et al., 2025), Pythagorean fuzzy set AHP (Gökler, 2025).

The AHP plays a crucial role in spatial analysis by organizing criteria into hierarchical structures and applying pairwise comparisons (Jothimani et al., 2021). The AHP method is commonly used to structure and prioritize multiple criteria, enabling the evaluation of their relative significance in decision-making. It also incorporates a consistency check mechanism to ensure the reliability of pairwise comparisons made by decision-makers (Karakuş and Yıldız, 2022). One of the key advantages of AHP lies in its ability to integrate both qualitative and quantitative factors within a hierarchical framework, facilitating the analysis of complex problems and allowing the measurement of consistency in judgments (Upadhyay et al., 2022).

The AHP is widely applied across various regions for assessing complex decision-making problems. This method facilitates the prioritization of multiple criteria by evaluating the relative importance of parameters or criteria. However, AHP is limited by its inability to effectively capture uncertainty in expert judgment. To address this limitation, researchers have integrated fuzzy set theory with AHP. In fuzzy set theory, commonly used fuzzy numbers include TFN, TRFN, and generalized fuzzy numbers. Unlike TFN, which represents uncertainty around a single central range, TRFN captures uncertainty over a broader interval, allowing a better representation of imprecise, variable, and hesitant expert opinions. This characteristic makes TRFN more suitable for synthesizing diverse judgments in pairwise comparisons and incorporating variations in expert evaluations. The TRFN-AHP approach provides enhanced uncertainty handling, reduces ambiguity in the decision-making process, generates more reliable parameter weights, and improves land suitability classification. Consequently, it enables a more comprehensive assessment and a more stable suitability ranking for complex agro-environmental decision-making problems. Therefore, this study adopts the TRFN-AHP approach to achieve a more robust assessment of land suitability for agriculture production.

The novelty of the study is as follows:

In the literature review on land suitability, no existing studies have applied the TRFN-AHP method. Therefore, the TRFN-AHP is used to determine land suitability zones in the study area.

The contributions of the study are as follows:

• To calculate parameter weights, the TRFN-AHP method is used
• To determine land suitability for agriculture production, hybrid approaches: TRFN-AHP, remote sensing, and GIS are used
• Testing the prediction accuracy, using the ROC curve.

2 Study area

Dharmapuri District is located in the northwestern part of Tamil Nadu, India, sharing its northern boundary with the state of Karnataka. The district is situated at approximately 12.1340° n latitude and 78.1580° E longitude. It lies roughly 300 km southwest of Chennai, the capital city of Tamil nadu. The district stands at an average elevation of about 450 m a.s.l. As shown in Figure 1, Dharmapuri District exhibits diverse topographic features, including plains, hills, and plateau regions that form part of the Eastern Ghats. This positioning plays a vital role in shaping its climate and agriculture practices. The northeast Monsoon, which typically occurs between October and December, is the primary source of rainfall in the district. Agriculture activities are closely aligned with the seasonal rainfall, which also impacts groundwater recharge and potential contamination levels. Based on temperature data obtained from the Indian Meteorological Department (IMD) for the period 1991–2020, the average annual temperature in Dharmapuri is approximately 25 °C. The hottest months occur between March and May, when temperatures can reach up to 35 °C, whereas the coolest period extends from December to February, with temperatures decreasing to around 17 °C. With the onset of the northeast Monsoon, temperatures begin to fall, and although the weather remains warm, the intensity of the heat reduces. This increase in moisture benefits both crops and the natural environment. The post-monsoon period, from December to February, is characterized by gradually decreasing temperatures. Despite being part of the winter season, this period remains important for agriculture as it supports the cultivation of various crops. Winters in Dharmapuri, particularly during January and February, are mild compared to northern India, with warm days and cooler nights. Frost is rare, making the winter season relatively moderate and favorable for year-round farming.

2.1 Materials and methods

In the present study area, hybrid approaches: TRFN-AHP, remote sensing, and GIS are used to assess land suitability. The geographical extent of the study area is depicted in Figure 1, which was prepared using QGIS. Table 1 Parameters with their data sources. Table 2 presents the parameters along with their corresponding weights and score for the subparameters. An overview of the methodology adopted is illustrated in Figure 2. The steps in the methodological framework are listed below:

• Step 1: Selection of parameters.
• Step 2: To prepare the parameters maps, using ArcGIS software.
• Step 3: To obtain parameter weights, using the TRFN-AHP method.
• Step 4: Determination of score for subparameters.
• Step 5: The result of land suitability index for agriculture production was validated using the ROC curve.

TABLE 1. Parameters with their data sources.

TABLE 2. Parameters with TRFN-AHP method weights (W) and subparameters with score (R) for agriculture production.

2.2 Data set

In the study area, parameters such as land use and land cover, slope, depth to water level, soil texture, annual rainfall, pH, nitrogen, organic carbon, temperature, soil moisture, distance to river, aspect, elevation, distance to road, normalized difference built-up index, and salinity were selected based on previous studies on land suitability (Upadhyay et al., 2022; Jamil et al., 2018; Everest and Gür, 2022; Farnood Ahmadi and Farsad Layegh, 2015; Sadiq et al., 2023; Yaman and Mutlu, 2025; Abimbola et al., 2021; Seyedmohammadi and Navidi, 2022; Karakuş and Yıldız, 2022; Mercan and Acibuca, 2023; Aburas et al., 2017). These parameters directly influence agriculture productivity and land suitability and are widely recognised as essential indicators for identifying suitable areas for agriculture production. They represent the major climatic, soil, topographic, hydrological, and land-use factors affecting agriculture suitability. Moreover, all selected parameters have been extensively applied in previous studies. In addition, the selection of these parameters was supported by expert knowledge to ensure their relevance to the local environmental and agriculture conditions of the study area.

The temperature was obtained from the IMD and processed in ArcGIS using the Spatial Analyst - Inverse Distance Weighting (IDW) tool. The annual rainfall was sourced from IMD and processed using the IDW tool to produce a spatially continuous rainfall map. Depth to water table was obtained from the Central Ground Water Board, Chennai, Tamil Nadu, India (CGWB) and processed in ArcGIS using the Spatial Analyst -IDW tools. The nitrogen (Cg/kg) and pH data sourced from CGWB were interpolated using the IDW technique to produce spatially continuous nitrogen and pH map. Organic carbon was obtained from the SoilGrids database, and the dataset was processed using mask extraction and spatial analysis tools to generate a raster layer. The elevation was obtained from the USGS Earth Explorer platform. Soil moisture was obtained from the ISRIC World Soil Information repository (https://data.isric.org). Topographic parameters, including aspect and slope were derived from Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) data downloaded from USGS Earth Explorer and processed using surface analysis tools in ArcGIS. The LULC was prepared from Sentinel satellite imagery through supervised classification techniques. The NDBI was derived from Google Earth Engine using Landsat 9 imagery via spectral analysis. Salinity was obtained from the ISRIC–World Soil Information database (https://data.isric.org/). The distance from rivers was derived from the SRTM DEM.

The satellite images and parameter datasets were acquired for the period from April 20 to 27, 2025. The dataset was processed in ArcGIS using the surface tools to generate the elevation raster. In the study area, to ensure uniform analysis, all the parameter layers were resampled to a resolution of 30 m. Consequently, the soil texture layer, which originated as a vector dataset, was digitized, converted to a raster format, and resampled.

In the study area, annual rainfall was used as the rainfall parameter in the land suitability evaluation (Dornik et al., 2022; Nguyen et al., 2020; Mandal et al., 2020). The present study area was selected based on the availability of rainfall data at the required spatial resolution, which facilitated direct integration into the TRFN-AHP and GIS analysis. Using single-year rainfall parameters captures the actual water availability experienced by farmers during that season, thereby making the assessment more relevant to practical agriculture decision-making. Long-term averages are often used to reduce variations from year to year. However, the rainfall of the selected year was taken as representative of in the study area.

The accuracy of the land-use classification was evaluated using a confusion matrix. Based on the confusion matrix, two widely used accuracy measures, namely, Overall Accuracy (OA) and the Kappa Coefficient (K), were calculated to assess the reliability of the classification results. These metrics quantify the level of agreement between the classified map and the reference data. The equations used to compute OA and K are presented in Equations 1, 2, respectively.where the number of correctly classified samples indicates the count of observations that were accurately classified, whereas the total number of samples denotes the complete set of observations considered in the analysis, including both accurate and inaccurate classifications.where “overall accuracy” represents the proportion of correctly classified samples relative to the total number of samples and reflects the overall performance of the classification. A value of 1 indicates complete agreement between the classified and reference data, corresponding to perfect classification. Estimated chance agreement refers to the level of agreement that may occur randomly and is incorporated to account for agreements that are not attributable to the actual performance of the classifier.

OA = Number of correctly classified samples Number of total samples (1)

K = Overall accuracy − estimated chance agreement 1 − estimated chance agreement (2)

Based on confusion matrix analysis, class-specific accuracy measures, including User’s Accuracy (UA), Producer’s Accuracy (PA), and the F1-score, were computed for each land-use class using Equations 3, 4. The F1-score, which represents the harmonic mean of UA and PA, provides a balanced assessment of classification performance at the individual class level. The equations used to determine these accuracy metrics are given below.where the number of correctly classified samples in each class indicates the observations correctly identified within a given class, while the number of samples classified into that class represents all observations assigned to that class, including both correct and incorrect classifications.where the number of correctly classified samples in each class denotes the observations that were accurately assigned to their actual class in the classification output, while the number of reference samples in each class represents the total observations belonging to that class based on the reference (ground-truth) data (Equation 5).where PA measures the likelihood that samples belonging to a specific class are correctly classified, and UA measures the likelihood that samples assigned to a class truly belong to that class. The F1-score integrates these two metrics into a single indicator by calculating their harmonic mean, offering a balanced assessment of class-level classification performance. It is widely used in previous studies due to its equal consideration of both PA (as a precision) and UA (as a recall) (Sabljić et al., 2024a).

UA = Number of correctly classified samples in each classs Number of samples classified to that class (3)

PA = Number of correctly classified samples in each class Number of samples from reference data in each class (4)

F 1 = 2 × UA × PA UA + PA (5)

The land suitability assessment results were integrated with agricultural areas obtained from the supervised classification of land-use data. The goal of the process is to determine land suitability for agriculture production.

2.3 Trapezoidal fuzzy numbers

Fuzzy set theory, introduced by Zadeh (1965), provides an effective framework for addressing uncertainty in decision-making processes. In this framework, a fuzzy set A ̃ defined on a universe of discourse X is described by a membership function μ A ̃ ( x ), which assigns to each element x ∈ X a membership value between 0 and 1, representing the degree of belongingness of x to A ̃. A TRFN, denoted as A ̃ = ( a 1 , b 2 , c 3 , d 4 ), is a specific type of fuzzy set that maps each element of X to a value within the interval [0, 1]. The mathematical formulation of the TRFN is given in Equation 6.

μ A ̃ x = x − a 1 b 2 − a 1 if a 1 ≤ x ≤ b 2 1 if b 2 ≤ x ≤ c 3 d 4 − x d 4 − c 3 if c 3 ≤ x ≤ d 4 0 if x ≥ d 4 & x ≤ a 1 (6)

2.4 Trapezoidal fuzzy numbers analytic hierarchy process method

Fuzzy set theory was first incorporated into the Analytic Hierarchy Process (FAHP) by Van Laarhoven and Pedrycz (1983) to address the limitations of traditional AHP in handling uncertainty and vagueness in expert judgments (Nukala and Gupta, 2005; Van Laarhoven and Pedrycz, 1983). By integrating fuzzy logic, the method became more effective for multi-criteria decision-making problems involving imprecise or subjective evaluations (Li, 2005). FAHP has since gained widespread acceptance and has been successfully applied in various fields to evaluate alternatives under uncertain conditions (Weck et al., 1997). Typically, TFN are used in FAHP to represent expert preferences, as they effectively approximate uncertain values. However, TRFN provides greater flexibility and more accurately captures the ambiguity inherent in human judgments. Consequently, TRFNs are increasingly employed in FAHP to enhance decision-making accuracy (Zheng et al., 2012). The weights of the parameters are determined using the following steps (Babu and Maury, 2025).

Step 1: Determine fuzzy pairwise comparison matrix:

This matrix is essential for determining the relative importance of each parameters using Equations 7, 8. It is developed based on expert judgment using trapezoidal fuzzy numbers, which represent the relative importance of parameters under uncertainty. These fuzzy numbers are derived by modifying the value of x in the intermediate scale. The values of x range from 1 to 9, corresponding to the Saaty scale (Saaty, 1980; Saaty, 1990). To better capture the uncertainty and vagueness inherent in expert opinions, these Saaty values are adjusted to generate TRFN. Assuming that there are a total of n factors ( i = 1,2 , … , n ) with the i-th criterion denoted by C i, experts construct the fuzzy pairwise comparison matrix A ̃ as shown in Table 3 using TRF-AHP scale in Table 4. P ̃ y x = ( P x y 1 , P x y 2 , P x y 3 , P x y 4 ) represents TRFN.

A ̃ = 1,1,1,1 P ̃ 12 P ̃ 13 ⋯ P ̃ 1 n 1 P ̃ 12 1,1,1,1 P ̃ 23 ⋯ P ̃ 2 n 1 P ̃ 13 1 P ̃ 23 1,1,1,1 ⋯ P ̃ 3 n ⋮ ⋮ ⋮ ⋱ ⋮ 1 P ̃ 1 n 1 P ̃ 2 n 1 P ̃ 3 n ⋯ 1,1,1,1 (7)

P ̃ y x = 1 P ̃ x y = 1 P x y 4 , 1 P x y 3 , 1 P x y 2 , 1 P x y 1 (8)

TABLE 3. The fuzzy pairwise comparison matrix is constructed by an expert using the TRFN scale to determine the weights of parameters for land suitability zones.

TABLE 4. Trapezoidal fuzzy numbers scale.

Step 2: Determination of fuzzy geometric mean: The fuzzy geometric mean is determined using the following Equations 9–13. Fuzzy weights are calculated by following Equations 14–18: where, q x represents fuzzy geometric mean values of the i t h parameters of each criteria and ω ̃ x is the fuzzy weights of j t h parameter.

X ̃ y = P ̃ x 1 × P ̃ x 2 × ⋯ × P ̃ x n 1 n = R x 1 , R x 2 , R x 3 , R x 4 (9)

R x 1 = ∏ y = 1 n P x y 1 1 n (10)

R x 2 = ∏ y = 1 n P x y 2 1 n (11)

R x 3 = ∏ y = 1 n P x y 3 1 n (12)

R x 4 = ∏ y = 1 n P x y 4 1 n (13)

ω ̃ x = q ̃ x × q ̃ x 1 + q ̃ x 2 + ⋯ + p ̂ x n − 1 = w x 1 , w x 2 , w x 3 , w x 4 (14)

w x 1 = q x 1 ∑ y = 1 n r x 4 − 1 (15)

w x 2 = q x 2 ∑ y = 1 n r x 3 − 1 (16)

w x 3 = q x 3 ∑ y = 1 n r x 2 − 1 (17)

w x 4 = q x 4 ∑ y = 1 n r x 1 − 1 (18)

Step 3: Defuzzification of fuzzy weights.

To obtain crisp weights, the Equation 19 is used.Step 4: Normalization of weights Finally, the crisp weights are normalized using the following Equation 20:Thus, the final weight vector is represented as:The TRFN-AHP method was used to determine parameter weights such as LULC (0.1332), slope (0.1141), depth to water level (0.1030), soil texture (0.0969), rainfall (0.0932), pH (0.0759), nitrogen (0.0700), organic carbon (0.0610), temperature (0.0501), soil moisture (0.0464), distance to river (0.0399), aspect (0.0338), elevation (0.0265), proximity to road (0.0240), nDBI (0.0212), and Salinity (0.0109). These weights were applied to generate the land suitability zones for agriculture production in the study region. In Table 3, to calculate consistency ratio (CR), use the following equation:where, λ Max is the maximum eigenvalue, n is several parameters.

w y ′ = w x 1 + 2 w x 2 + w x 3 + w x 4 6 (19)

w y = w y ′ ∑ y = 1 n w y ′ (20)

W = w 1 , w 2 , … , w n

C o n s i s t e n c y i n d e x C I = λ Max − n n − 1 C R = C I R I

The fuzzy pairwise comparison matrix for agriculture production parameters, λ max = 17.2 − 16, CI = (17.2–16)/(16–1) = 0.08, and CR = 0.087/1.45 = 0.06. The CR is 0.06. So the fuzzy pairwise comparison matrices are acceptable.

In the study area, the fuzzy pairwise comparison matrix was constructed based on a literature review on land suitability studies and expert knowledge (Aburas et al., 2017; Jamil et al., 2018; Tashayo et al., 2020b; Abimbola et al., 2021; Rukhsana and Molla, 2022; Sengupta et al., 2022; Seyedmohammadi and Navidi, 2022; Yaman and Mutlu, 2025). Relevant published studies were carefully examined to determine the relative importance of each parameter influencing land suitability. The Saaty scale was adapted into trapezoidal fuzzy numbers to better represent the uncertainty and variability reported in previous studies.

3 Assessment of land suitability for agriculture production

Each parameter was classified into five suitability classes and assigned scores ranging from 1 to 5, where 1 = not suitable, 2 = marginally suitable, 3 = moderately suitable, 4 = highly suitable, and 5 = very highly suitable. The basis for the classification and scoring is presented in Table 2, which provides the corresponding suitability ranges and criteria adopted from relevant literature and expert recommendations. The classification intervals were established based on previous studies, crop requirements, expert knowledge, and regional environmental conditions. For example, lower slope values were assigned higher suitability scores because they provide favourable conditions for agriculture production and reduce the risk of soil erosion, whereas steeper slopes received lower scores. Similarly, optimal ranges of rainfall, soil pH, soil moisture, nitrogen, organic carbon, temperature, and groundwater availability were assigned higher suitability scores due to their positive influence on agriculture productivity. Continuous variables were reclassified into discrete suitability classes using the Reclassify tool in ArcGIS, while categorical variables, such as land use/land cover and soil texture, were assigned suitability scores according to their relative suitability for agriculture production. After reclassification, all thematic layers were converted into raster format and integrated using the Fuzzy Weighted Linear Combination (FWLC) method, with parameter weights derived from the TRFN-AHP approach. The detailed classification intervals, suitability scores, and scientific basis for assigning these scores are presented in Table 2, thereby enhancing the transparency of the methodology and clarifying the contribution of each parameter to the final agriculture production suitability map. The Land Suitability Index (LSI) was then computed using the FWLC method in ArcGIS, as expressed in the following equation (Tashayo et al., 2020a; Karakuş and Yıldız, 2022):where n is the total number of parameters w i is a weights of parameters and r i is a scores of subparameters.

LSI = ∑ i = 1 n w i r i

3.1 Selection of parameters

3.1.1 LULC

Land use plays a crucial role in determining land suitability for agriculture production, as it directly influences soil conditions, water availability, and overall crop productivity. Different land use types affect key factors such as infiltration, moisture retention, and nutrient availability. Forested areas improve soil structure and enhance water infiltration, which supports long-term soil fertility, while excessive built-up areas with impervious surfaces restrict water movement and reduce the availability of arable land. Agriculture lands (farms) are generally more suitable for agriculture production. However, their productivity depends on proper land management practices, irrigation, and soil conservation measures. In addition, land use and land cover patterns significantly influence agriculture production, where areas with sparse vegetation or intensive urbanisation are more prone to moisture stress and reduced crop yield. In the study area, the land use/land cover classification includes major categories such as built-up land, agriculture land (farms), and water bodies, as illustrated in Figure 3.

3.1.2 Slope

Slope plays a significant role in determining land suitability for agriculture production, as it directly influences water movement, soil retention, and erosion processes. It affects the balance between infiltration and surface runoff, which in turn impacts soil moisture availability essential for crop growth. Steep slopes tend to generate higher surface runoff and lower infiltration, leading to increased soil erosion and reduced water availability for crops, making such areas less suitable for agriculture. In contrast, gentle slopes promote greater water infiltration and better moisture retention, supporting favorable conditions for agriculture production. Therefore, slope is generally negatively related to infiltration and positively related to runoff (Das and Pal, 2020). The slope is classified into five suitability classes: very highly (225–389), highly (389.01–512), moderately (512.01–658), marginally (658.01–848), and not suitable (848.01–1,271) Figure 4.

3.1.3 Depth to water level

Depth to water level represents the vertical distance from the ground surface to the groundwater table. This zone acts as a pathway through which water and dissolved substances move before reaching the aquifer. In the context of land suitability for agriculture production, a shallower water table is generally more favorable, as it ensures easier access to groundwater for irrigation, whereas deeper water levels may limit water availability and increase extraction costs. Therefore, greater depth to the water table is typically associated with lower land suitability (Babu and Maury, 2025). This parameter is an important factor in the ranking and weighting of criteria used in the analysis. In the present study area, the depth to water level ranges from 6.76 to 198.89 m (Figure 5). Based on this range, it is classified into suitability categories such as very high suitability (6.76–25.6), high suitability (25.61–39.91), moderate suitability (39.91–69.30), low suitability (69.31–121.29), and very low suitability (121.29–198.89).

3.1.4 Soil texture

In the land suitability, soil characteristics are a key factor, especially with respect to irrigation potential (Xiang et al., 2022; Jamil et al., 2018). Soils are commonly classified into five categories based on their infiltration capacity and water-holding characteristics. Clay soils generally have low permeability and tend to retain water for longer periods, which may restrict aeration and limit their suitability for continuous cultivation. In contrast, sandy soils exhibit high infiltration rates, allowing rapid water movement but often leading to low moisture retention. Although a variety of crops can grow across different soil types, excessive or insufficient soil moisture can adversely affect root growth and crop productivity (Mandal et al., 2020). Among all soil types, loamy soil provides a balanced combination of drainage and moisture retention, making it most suitable for agriculture production. The soil texture is classified as shown in Figure 6.

3.1.5 Annual Rainfall

Annual rainfall, derived from both natural processes and human influences, infiltrates into the soil and contributes to groundwater recharge. The amount of water that enters and moves through the subsurface is referred to as net recharge, which is controlled by factors such as slope gradient, soil permeability, precipitation, and land use/land cover (Rangzan et al., 2024). This process is particularly important because water acts as a major carrier for nutrients as well as potential contaminants (Tashayo et al., 2020b). In the present study area, the annual rainfall ranges from 709 to 826.36 mm/year (Figure 7). For land suitability assessment, these rainfall values were categorized into five classes: not suitable (709–741.37 mm), marginally suitable (741.38–758.37 mm), highly suitable (758.38–777.2 mm), and very highly suitable (777.21–799.72 mm). This classification is essential for identifying suitable zones for agriculture production, where areas under very high suitability are expected to support better crop growth. In contrast, those under very low suitability may face limitations in water availability. Annual rainfall data were utilised to estimate recharge using the method proposed by the Research Institute of UP Irrigation (Chaturvedi, 1973).where R denotes the effective rainfall (mm), P represents the total rainfall or precipitation (mm), 1.35 is an empirical coefficient, 14 is a constant accounting for initial rainfall losses, and the exponent 0.5 represents the square root function used to estimate effective rainfall.

R = 1.35 P − 14 0.5

3.1.6 pH

pH is largely controlled by its mineral composition and serves as an important indicator of nutrient availability, potential toxicity, and overall crop suitability (Mandal et al., 2020; Tashayo et al., 2020b). The pH range supports effective root development, nutrient uptake, and healthy plant growth. In general, most crops perform well under neutral to slightly alkaline conditions. A pH value below 7.0 indicates acidic conditions, while values above 7.5–8.0 reflect increasing alkalinity. Although many crops can tolerate slightly acidic soils, excessive alkalinity or salinity can significantly limit crop growth and productivity. Therefore, both soil and irrigation water pH must be carefully considered when evaluating land suitability. In the study area, pH ranges from 7.19 to 8.20 (Figure 8). The pH map was classified into five suitability categories: not suitable (7.19–7.41), representing optimal conditions; marginally suitable (7.42–7.57), indicating favourable conditions; moderately suitable (7.58–7.70), where some limitations exist; highly suitable (7.71–7.87), where alkaline conditions begin to affect nutrient availability; and very highly suitable (7.87–8.20), where high alkalinity restricts effective cultivation without proper soil management.

3.1.7 Nitrogen

Nitrogen represents the concentration of available nitrogen in the soil and is one of the most important macronutrients required for plant growth. It plays a vital role in essential physiological processes such as vegetative growth and overall crop productivity. The availability of nitrogen is controlled by factors including organic matter content, microbial activity, and mineralization processes. Adequate nitrogen levels promote healthy plant development, improve nutrient use efficiency, and enhance crop yield, whereas nitrogen deficiency can result in stunted growth, chlorosis (yellowing of leaves), and reduced productivity. On the other hand, excessive nitrogen levels may lead to nutrient imbalance, weak root systems, lodging, and negative environmental impacts. Therefore, evaluating soil nitrogen content is essential for determining land suitability for agriculture production. Nitrogen values range from 0 to 515 (Figure 9). Based on their suitability for crop growth, nitrogen levels were classified into five categories: very highly suitable (0–515), representing optimal nutrient conditions; highly suitable (352–515), indicating favourable levels; moderately suitable (276–352), where some limitations may occur; marginally suitable (218–276), where nitrogen is insufficient for optimal growth; and not suitable (0–218), where severe deficiency limits crop establishment without significant soil amendments.

3.1.8 Organic carbon

Organic carbon is a critical indicator in assessing land suitability for agriculture production, as it directly influences soil fertility and productivity. It contributes significantly to both the chemical and physical properties of soil, including structure, water-holding capacity, temperature regulation, pore distribution, and microbial activity (Ramamurthy et al., 2020; Abimbola et al., 2021). Soils with adequate organic carbon enhance nutrient availability, improve fertilizer efficiency, and support sustainable crop growth. In contrast, low organic carbon content reduces soil quality, limits nutrient retention, and negatively affects crop yield. The level of organic carbon varies with land management practices and natural soil conditions. Higher organic carbon content is generally more favourable for agriculture, as it improves moisture retention and promotes healthy root development, which are essential for crop production (Kauer et al., 2021). In the present study area, organic carbon values range from 0 to 864 (Figure 10). For evaluating land suitability for agriculture production, organic carbon was classified into five categories: highly suitable (410–864), moderate suitable (326–410), moderately suitable (300–362), marginally suitable (263.01–326), and not suitable (0–263). This classification helps identify areas with optimal soil fertility, thereby supporting effective agriculture planning and sustainable crop production.

3.1.9 Temperature

Climatic conditions play a vital role in determining land suitability for agriculture production, particularly through their influence on temperature. Crops can tolerate a moderately wide range of temperatures; however, extreme conditions can adversely affect plant growth, flowering, and yield. Optimal temperature levels are essential for efficient photosynthesis, vegetative development, and grain formation. In the study area, temperature was used as a key parameter to evaluate land suitability for agriculture production. The temperature ranges from 68°F to 88°F, indicating varying levels of suitability across the region. The temperature was classified into five categories: very highly suitable (84.17–88), highly suitable (81.42°F–84.16°F), moderately suitable (81.42°F–84.16°F), marginally suitable (73.50°F–78.20°F), not suitable (68°F–73.49°F) and not suitable (68°F–73.49°F). The spatial distribution of temperature classes is shown in (Figure 11).

3.1.10 Soil moisture

Land characteristics such as soil, rock formations, water bodies, vegetation, and built-up areas significantly influence spectral reflectance properties. Spectral reflectance is an important parameter obtained through remote sensing techniques, enabling effective identification and analysis of different land cover types. Among these, water bodies typically exhibit low reflectance due to their absorption characteristics. Soil moisture, derived from spectral analysis, is a key factor in determining land suitability for agriculture production, as it directly influences soil health, nutrient availability, and crop productivity (Pramanik, 2016; Hosen et al., 2025). Maintaining optimal soil moisture is essential for supporting plant growth, root development, and efficient nutrient uptake. Both insufficient and excessive moisture levels can adversely affect crop performance. In the study area, soil moisture values range from 0.19 to 0.28 (Figure 12). Soil moisture was classified into five categories: not suitable (0.19–0.21), marginally (0.22–0.23), moderately suitable (0.24–0.25), highly suitable (0.26–0.26), and very highly suitable (0.27–0.28). This classification helps identify areas with optimal moisture conditions for agriculture production, thereby supporting effective land management and improving overall crop productivity.

3.1.11 Distance to river

The distance from rivers is an important factor in determining land suitability for agriculture production, as it directly influences groundwater availability and soil moisture conditions. Areas located closer to rivers generally have higher groundwater levels due to continuous recharge and lateral water movement, which can support irrigation and enhance crop growth. In contrast, regions situated farther away from rivers may experience reduced groundwater availability and increased dependence on rainfall or artificial irrigation sources. However, the impact of river proximity on groundwater conditions can vary depending on local geology, soil characteristics, and hydrological conditions. In the study area, the distance from rivers ranges from 0 to 12,600 m (Figure 13).

3.1.12 Aspect

Aspect refers to the orientation or direction that a slope faces, which plays a significant role in determining land suitability for agriculture production. It controls the amount of solar radiation received by the land surface, thereby influencing temperature, evaporation rates, and soil moisture conditions. Slopes facing different directions experience variations in microclimate, which can affect crop growth, water requirements, and overall productivity. For instance, slopes receiving higher sunlight exposure tend to have increased evaporation and lower soil moisture, whereas less exposed slopes may retain more moisture and provide relatively favorable conditions for agriculture production. In the study area, aspect values range from 1 to 359.80 (Figure 14), representing all possible slope orientations. Aspect was classified into five categories: not suitable (1–66.92), marginally suitable (66.93–139.08), moderately suitable (139.09–211.24), highly suitable (211.25–284.81), and very highly (284.82–359.80). Aspect refers to the orientation or direction that a slope faces, which plays a significant role in determining land suitability for agriculture production. It controls the amount of solar radiation received by the land surface, thereby influencing temperature, evaporation rates, and soil moisture conditions. Slopes facing different directions experience variations in microclimate, which can affect crop growth, water requirements, and overall productivity. For instance, slopes receiving higher sunlight exposure tend to have increased evaporation and lower soil moisture, whereas less exposed slopes may retain more moisture and provide relatively favorable conditions for agriculture production.

3.1.13 Elevation

Elevation is an important factor in determining land suitability for agriculture production, as it significantly influences microclimatic conditions such as temperature, rainfall distribution, wind patterns, and evapotranspiration. Variations in elevation can lead to differences in soil moisture availability and crop growth conditions across this region. Generally, higher elevations are associated with increased runoff, lower water retention, and greater agriculture production, which can limit agriculture productivity. In contrast, lower elevations often have better moisture accumulation and more favourable conditions, although they may also be prone to waterlogging in some cases. Elevation ranges from 225 m to 1,271 m (Figure 15). The elevation is classified as follows: not suitable (225–389), marginally suitable (389–512), moderately suitable (512–658), highly suitable (658–848), and very highly (848–1,271). These are classes indicating considerable topographic variation in the study area. This variation plays a key role in shaping local agriculture potential and water availability. By incorporating elevation into the suitability analysis, it becomes possible to identify areas with favourable environmental conditions and to support effective land use planning and sustainable agriculture management.

3.1.14 Proximity to road

Proximity to roads is a significant factor in assessing land suitability for agriculture production, as it directly influences accessibility to transportation networks, farm inputs, labor, and markets. Areas located closer to roads benefit from reduced transportation costs, easier movement of goods, timely access to fertilizers and equipment, and improved market connectivity, all of which enhance agriculture productivity and profitability. In contrast, lands situated farther from roads may face challenges such as higher transportation expenses, delayed input supply, and limited market access, which can negatively impact farming activities and economic returns. In the study area, the proximity to roads ranges from 0 to 0.19 km (Figure 16). The proximity to road is classified as follows: very highly (0–0.02), highly (0.03–0.04), moderately suitable (0.05–0.08), marginally suitable (0.09–0.12), and not suitable (0.13–0.19).

3.1.15 Normalized difference built up index

The normalized difference built-up Index is primarily used to identify urban and built-up areas using remote sensing data, rather than directly evaluating agriculture land suitability. However, it plays an important supporting role in land suitability analysis by helping to distinguish non-agriculture zones. Built-up areas are generally unsuitable for farming due to the presence of impervious surfaces such as buildings and roads, which limit soil availability, water infiltration, and crop growth. By mapping and excluding these areas, NDBI allows for a more accurate identification of land that is potentially suitable for agriculture production. In the study area, NDBI values range from 0.29 to 0.34 (Figure 17). The NDBI was classified into five categories: very highly suitable (0.29–0.04), highly suitable (0.05–0.09), moderately suitable (0.10–0.13), marginally (0.14–0.17), and not suitable (0.18–0.34).

3.1.16 Salinity

Salinity is a critical factor in evaluating land suitability for agriculture production, as it directly affects soil fertility, water uptake, and overall crop performance. Salinity is commonly assessed using electrical conductivity. This threshold may vary depending on crop tolerance, as some crops can withstand higher salinity levels while others are more sensitive. High salinity reduces the ability of plants to absorb water due to osmotic stress and can also lead to ion toxicity, ultimately limiting plant growth and yield. In the study area, salinity is classified as shown in the (Figure 18).

3.2 Result and discussion

In the study area, several parameters were selected, including LULC, Slope, depth to water level, soil texture, annual rainfall, pH, nitrogen, organic carbon, temperature, soil moisture, distance to river, aspect, elevation, proximity to road, NDBI, and salinity. The land suitability map was generated using the fuzzy weighted overlay tool in ArcGIS. The land suitability for agriculture production map (Fig: Figure 19). Based on the results, the land suitability for agriculture production is classified as follows: not suitable (8.86%), marginally suitable (20.70%), moderately suitable (24.88%), highly suitable (27.12%) and very highly suitable (18.44%) as shown in Table 5.

TABLE 5. TRFN-AHP on land suitability zone for agriculture production.

The TRFNAHP results indicate that LULC (0.1332), slope (0.1141), depth to water level (0.1030), soil texture (0.0969), and rainfall (0.0932) are the most influential parameters affecting agricultural land suitability in the study area. In contrast, salinity (0.0109), NDBI (0.0212), proximity to road (0.0240), and elevation (0.0265) received relatively lower weights. These findings suggest that land-use conditions, topography, groundwater availability, soil characteristics, and climatic factors have a greater influence on agricultural suitability than accessibility and salinity-related factors in the study area. A comparison with previous studies using AHP and TFN-AHP approaches (Sadiq et al., 2023; Everest and Gür, 2022; Abimbola et al., 2021; Tashayo et al., 2020b) shows that the criterion weights and rankings vary among studies. Such differences are expected because of variations in study objectives, crop requirements, environmental conditions, selected parameters, and expert judgments. Therefore, direct comparison of numerical weights is challenging. Nevertheless, the present results demonstrate the effectiveness of the TRFNAHP approach in prioritizing land suitability criteria while accounting for uncertainty in expert judgments, resulting in criterion weights that reflect the local agricultural and environmental conditions of the present study area.

Areas classified as not suitable are predominantly concentrated in the northwestern, western, southwestern, and southeastern parts of the district, particularly around Periyanoor, Marandahalli, Hogenakkal, Neruppur, and Sitteri. These regions are generally associated with hilly terrain, steep slopes, poor soil properties, low rainfall, high temperatures, and limited water availability. Such conditions reduce soil fertility, restrict moisture retention, and hinder crop growth. Furthermore, deeper groundwater levels, greater distance from water sources, and the presence of built-up land limit irrigation potential and reduce cultivable areas. Consequently, these regions experience low agriculture productivity, a higher risk of crop failure, and limited suitability for sustainable farming.

Marginally suitable and moderately suitable zones are widely distributed throughout the central and northern parts of the district, forming transitional areas between highly suitable and unsuitable regions. These areas are characterized by moderate constraints, including gentle slopes, suboptimal soil fertility, limited soil moisture, moderate rainfall, and moderate groundwater availability. They also exhibit relatively favorable conditions, such as improved soil texture, balanced pH, and moderate levels of nitrogen and organic carbon, which support stable crop growth and moderate agriculture productivity. Furthermore, reasonable proximity to water sources enhances irrigation potential and reduces the risk of crop failure. However, agriculture production may still be constrained by insufficient nutrient availability, periodic water stress, and partial limitations in irrigation access. Therefore, appropriate management practices, including soil amendments, efficient irrigation systems, and improved nutrient management, are necessary to enhance agriculture productivity and sustainability in these areas.

Highly suitable and very highly suitable zones are mainly concentrated in the central, eastern, and northeastern parts of the district, particularly around Dharmapuri, Nallampalli, Kadathur, Morappur, Harur, Adilam, Bodipatti, and Komabainallur. These regions represent the most productive agriculture areas and are characterized by favorable topographic conditions, including nearly flat terrain and low slopes, together with fertile soils, ideal pH, high levels of organic carbon and nitrogen, adequate soil moisture, and reliable water availability from both rainfall and shallow groundwater resources. Favorable climatic conditions, such as adequate rainfall and moderate temperatures, further support crop development, while better water accessibility, lower salinity levels, and minimal built-up interference enhance agriculture productivity. These factors collectively promote optimal crop growth, higher yields, reduced cultivation risk, and sustainable agriculture practices, making these areas highly suitable for intensive and diversified agriculture production.

From an agriculture planning perspective, the highly suitable and very highly suitable zones should be prioritized for sustainable crop production and efficient resource allocation. In contrast, the marginally suitable and unsuitable regions may require soil and water conservation measures, improved irrigation facilities, and appropriate land management practices to enhance their agriculture potential.

The spatial distribution of land suitability zones obtained in this study indicates that highly suitable and very highly suitable areas are predominantly concentrated in the central and eastern parts of the district, whereas unsuitable and marginally suitable zones are mainly distributed in the northwestern, western, and southeastern hilly regions. Similar spatial patterns have been reported in previous land suitability studies conducted in semi-arid regions, where highly suitable agriculture lands were concentrated in low-slope areas with favorable soil and water conditions, while hilly and water-deficient regions exhibited lower suitability (Ganie et al., 2023; Mandal et al., 2020; Tashayo et al., 2020a). This similarity suggests that topography, soil fertility, rainfall, and groundwater availability are the primary factors controlling agriculture suitability in the study area.

Lower agriculture production zones are primarily influenced by limiting factors such as high elevation, low rainfall, deeper groundwater levels, low nitrogen content, and unfavourable climatic conditions. These factors interact to reduce soil moisture availability through increased runoff and evapotranspiration, thereby limiting nutrient uptake and root development, which ultimately results in reduced crop growth and lower agriculture productivity.

To compare present study results with existing studies results, existing studies have used five zones of land suitability and methods such as the AHP method (Bozdağ et al., 2016; Tashayo et al., 2020a; Ramamurthy et al., 2020; Dedeoğlu and Dengiz, 2019) and triangular fuzzy number (Kılıc et al., 2022). The present results differ from those of existing studies. This deviation is primarily attributed to the use of the TRFN-AHP method, which more effectively handles uncertainty and provides a clearer representation of expert judgments, leading to a more accurate and reliable assessment of land suitability (Malmir et al., 2016). Furthermore, the observed differences may also result from variations in parameter selection and weighting approaches. In addition, factors such as the spatial resolution of datasets and local environmental conditions, including vegetation cover and rainfall variability. These findings highlight that both the choice of method and the selection of parameters play a critical role in determining land suitability outcomes (Selassie et al., 2014; Vahidi et al., 2023; Koca et al., 2025; Tuesta-Trauco et al., 2025; Chandrakala et al., 2025; Sabljić et al., 2024a).

The incorporation of fuzzy set theory into the AHP framework has been extensively utilized to manage uncertainty and vagueness in expert judgments (Karthikeyan et al., 2016). Therefore, many studies have adopted fuzzy AHP approaches to deal with uncertain data related to socioeconomic conditions and climatic variability (Vaz et al., 2025). The integration of fuzzy set theory with the AHP method has been widely applied to improve the accuracy of decision-making in pairwise comparisons, especially in land suitability analysis (Ayalew, 2014; Zhang et al., 2015; Malmir et al., 2016; Rashidi and Sharifian, 2022; Tiruneh et al., 2023; Koca et al., 2025).

4 Validation

The LSI for agriculture production was validated using the ROC curve. The ROC method was selected because it is widely used in spatial suitability, vulnerability, and environmental modelling studies to evaluate predictive performance (Bera et al., 2021); Babu and Maury (2026)). It provides a robust and threshold-independent measure of the model’s discrimination ability, indicating how effectively the LSI distinguishes between suitable and unsuitable areas. The ROC analysis is particularly advantageous due to its cutoff independence and its ability to assess model performance across the entire range of threshold values, thereby offering a more objective and reliable evaluation (Frattini et al., 2010).

The validation was carried out using ground truth occurrence data collected through field observations. The use of field data is essential to ensure the accuracy and reliability of the land suitability map. The LSI serves as a direct indicator of model performance and provides valuable information for validation. For the ROC analysis, a dataset representing land suitability zones for agriculture production was prepared. A total of 232 field-verified locations were used to compute the Area Under the Curve (AUC), where presence points indicate areas confirmed as suitable for agriculture through field observations. The ROC–AUC analysis was conducted using these field-based data. The LSI was treated as a continuous variable, and its predictive performance was evaluated using an independent binary reference dataset (very highly suitable = 1, not suitable = 0) derived from threshold values. The threshold was defined based on natural breaks in the LSI distribution and supported by agronomic suitability criteria. Using this threshold, validation metrics such as overall accuracy, sensitivity, specificity, and ROC–AUC were calculated to quantitatively assess the model performance.

The ROC curve is a graphical representation plotted between sensitivity (true positive rate) on the vertical axis and 1–specificity (false positive rate) on the horizontal axis. Sensitivity represents the proportion of correctly predicted suitable areas, while specificity indicates the proportion of non-suitable areas that are correctly identified for agriculture production (Khosravi et al., 2018; Torkashvand et al., 2021; Hoque et al., 2020; Karakuş and Yıldız, 2022). Due to its clear, comprehensive, and visual nature, the ROC curve is widely used for quantitative accuracy assessment (Tehrany et al., 2013). The ROC method is based on the relationship between true positive and false positive rates. The performance of the agriculture suitability model is evaluated using the AUC of the ROC. Sensitivity and specificity are calculated using the following equations (Karakuş and Yıldız, 2022).

S e n s i t i v i t y = T P T P + F N

S e n s i t i v i t y = T N F P + T N

where, False Positive (FP), False negative (FN), True Positive (TP) and True negative (TN). The AUC was calculated to evaluate the predictive performance of the model, with values closer to 1 indicating higher accuracy (Tehrany et al., 2013; Rahmati et al., 2016). The AUC value serves as an important indicator of prediction reliability and is commonly classified as follows: 0.5–0.6 (poor), 0.6–0.7 (moderate), 0.7–0.8 (good), 0.8–0.9 (very good), and 0.9–1.0 (excellent). In the present study area, the AUC value is 0.76 (Figure 20), which indicates good predictive accuracy in the study area.

5 Conclusion

In the study area, to assess land suitability for agriculture production, using hybrid approaches: trapezoidal fuzzy number analytic hierarchy process, remote sensing and GIS were used. These parameters were selected based on expert judgment and previous studies. TRFN-AHP was applied to assign criterion weights. The AUC value is 0.76, which indicates good prediction accuracy.

The results revealed that soil type, water availability, land use, climatic conditions, and topographic characteristics are the primary factors controlling agriculture suitability. Areas with favorable soil properties, adequate moisture availability, and reliable water resources exhibited higher suitability for agriculture production, whereas regions characterized by poor soil conditions, water scarcity, and land degradation showed lower suitability. The spatial distribution of suitability zones provides valuable information for identifying priority areas for agriculture development and sustainable land management.

The main contribution of the integration of TRFN-AHP method and GIS for agriculture land suitability assessment which enables a more effective representation of uncertainty in expert judgments compared with conventional AHP-based approaches. The proposed framework provides a systematic and reliable decision-support tool for evaluating agriculture potential and optimizing land-use planning in semi-arid environments. The findings have broader significance for policymakers, land-use planners, and agriculture stakeholders by supporting evidence-based decision-making, sustainable resource allocation, and improved agriculture productivity.

To promote sustainable agriculture, farmers should adopt environmentally friendly practices such as proper soil management, drip irrigation, rainwater harvesting, crop rotation, agroforestry, and efficient waste management. These practices can improve soil health, enhance water-use efficiency, reduce land degradation, and support long-term agriculture sustainability. Future research should further investigate the impacts of climate change, extreme weather events, and temporal variability on land suitability to improve the resilience and adaptability of agriculture production.

The limitations of the study are outlined below and should be addressed in future work:

• In assessing land suitability, the determination of parameter weights plays a crucial role, and the use of different fuzzy set types in AHP may lead to variations in the assigned weights, which in turn influence the resulting land suitability index.
• Several extensions of fuzzy set theory can be integrated with the AHP method, including Intuitionistic Fuzzy Sets, Type-2 Fuzzy Sets, Pythagorean Fuzzy Sets, Spherical Fuzzy Sets, Fermatean Fuzzy Sets, and Bipolar Intuitionistic Fuzzy Sets, which are increasingly being applied in GIS and environmental studies.
• In addition, the inclusion of a broader range of environmental, climatic, and socio-economic parameters, supported by expert judgment, can improve the reliability of the LSZ. This framework may also be extended to other regions with different agro-climatic settings.
• Parameter weighting was conducted based on the judgment of a single domain expert, consistent with several previous GIS-based land suitability studies. While the expert possesses considerable experience, the inclusion of multiple experts could further minimize individual subjectivity and enhance the robustness of the weighting process.
• The robustness of the TRFN-AHP-derived weights and the corresponding land suitability zones was not assessed using a formal sensitivity analysis. Therefore, future studies should incorporate group-based expert decision-making and conduct sensitivity analysis to quantify the impact of variations in parameter weights on the spatial distribution of land suitability zones.

Generative AI statement

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

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

Publisher’s note

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

Footnotes

Edited by: Henry Bulley, Borough of Manhattan Community College, United States

Reviewed by: Luka Sabljic, University of Banjaluka, Bosnia and Herzegovina

Mohamed Chaibi, Moulay Ismail University, Morocco