Land is a vital resource directly impacting agricultural development and food security. Growing populations increase pressure on land resources, necessitating efficient land use and management. Fertile farmlands are negatively affected by industrialization, urbanization, and improper land use, leading to soil degradation and irreversible consequences. Turkey, with approximately 24 million hectares of agricultural land, faces land degradation risks. Identifying suitable areas for cultivation is crucial for sustainable land use planning. Land evaluation methods range from qualitative (expert knowledge) to quantitative (simulation models). MCDA approaches are suitable for addressing the complexity of land suitability assessment involving multiple criteria. Fuzzy logic enhances MCDA by incorporating uncertainties related to human cognitive processes. This study applies Fuzzy Analytic Hierarchy Process (FAHP) integrated with GIS techniques to identify potential agricultural areas in Central Anatolia, a semi-arid region, focusing on sustainable land use and management by considering soil characteristics.

Existing land evaluation methodologies include qualitative methods based on expert knowledge and quantitative models based on simulation models. Quantitative models offer high detail but demand significant data, time, and cost. Qualitative approaches utilize mathematical formulas to express land and soil features impacting agricultural land suitability. Multi-Criteria Decision Analysis (MCDA) offers a more suitable approach for land evaluation due to the inherent complexity of the problem. Remote sensing, Geographic Information Systems (GIS), and MCDA are frequently combined to overcome the challenges of land suitability assessment. Analytic Hierarchy Process (AHP) is a commonly used MCDA method for assigning weights to evaluation criteria, capable of handling inconsistencies in decision-making. However, assigning precise numerical values to criteria can be challenging. Fuzzy logic addresses this by incorporating linguistic terms, allowing for uncertainty contemplation in the decision-making process. The integration of fuzzy sets with AHP (FAHP) has been employed across various fields, including visual merchandising prioritization, mine security risk assessment, surface water quality assessment, occupational stress evaluation, chromite processing plant location selection, optimal maintenance strategy selection, and geological risk ranking. Existing land quality indices generally apply to specific purposes and environments, lacking universal applicability. The study highlights the need for a regionally specific model that considers the unique ecological and socio-cultural factors of Central Anatolia.

The study area encompasses approximately 195,013 km² in Central Anatolia, characterized by an average elevation of 1200 m and an average slope of 9.6%. The region exhibits a semi-arid climate with hot, dry summers and cold, snowy winters. Geological materials include metamorphic, granitic, and ophiolitic units, with sedimentary rock units dating back to the Precambrian and early Paleozoic eras. The region's limited rainfall and vegetation lead to soils that are poor in organic matter. A total of 4517 soil samples (0-30 cm depth) were collected for physico-chemical analysis (particle size distribution, pH, EC, CaCO3 content, and organic matter content). Different interpolation methods (Inverse Distance Weighing (IDW), Radial Basis Function (RBF), and Kriging) were applied to predict the spatial distribution of soil parameters. Root mean square error (RMSE) was used to evaluate the accuracy of interpolation models. Nine criteria were selected to identify suitable areas for potential agricultural lands: slope, depth, erosion, parent material, soil texture, organic matter (OM), bulk density (BD), pH, and lime content (CaCO3). These criteria were chosen based on their influence on plant growth and were categorized with scores from 1 (optimal) to 4 (low suitability). Fuzzy Analytic Hierarchy Process (FAHP) was used to determine the relative importance (weight) of each criterion. Triangular fuzzy numbers were assigned to linguistic scales representing pairwise comparisons of criteria. Buckley's geometric mean method was used to calculate fuzzy weights, followed by defuzzification using the Center of Area (COA) method to obtain crisp weights. The Weighted Linear Combination (WLC) method was employed to calculate the suitability index for each area, classified into five levels (S1: very high, S2: high, S3: marginally suitable, N1: currently unsuitable, N2: permanently unsuitable) using the Natural Breaks Jenks method. The Revised Universal Soil Loss Equation (RUSLE) model was used to estimate soil erosion.

Descriptive statistics of soil properties revealed variability influenced by environmental factors and land use. pH ranged from 6.40 to 9.47, BD from 0.72 to 1.84 g/cm³, CaCO3 from 0.01% to 95.24%, and OM from 0.42% to 13.0%. The Kriging Simple Spherical model was identified as the most suitable interpolation model for organic matter, RBF Spline with Tension for bulk density, Kriging Simple Gaussian for pH, and RBF Completely Regularized Spline for lime and texture. FAHP analysis, involving three experts, determined weights for the nine criteria. The pairwise comparison matrix was converted into triangular fuzzy numbers. The fuzzy weights were calculated using Buckley's geometric mean method and then defuzzified using the COA method. Slope (0.245), depth (0.217), and erosion (0.155) had high weights; texture (0.118), OM (0.091), and BD (0.061) had moderate weights; and pH (0.048), parent material (0.041), and lime content (0.024) had low weights. Spatial distribution maps of the criteria were created. Approximately 7% of Central Anatolia exhibited severe or very severe erosion risk, while 83.6% had slight or low risk. Deep soils (90+ cm) covered 33.1% of the area, while shallow soils (0-20 cm) covered 38.8%. About 45.3% of the parent material consisted of acid magmatic rocks, while 37.3% were basic-ultrabasic magmatic rocks. 64.3% of the land had a slope less than 12%, suitable for mechanized agriculture. Most soils (84%) had very low OM content. Approximately 98% of the soils had medium to fine texture. More than 95% had medium to high bulk density. Over half of the lands were slightly to moderately alkaline. The FAHP-based suitability assessment classified approximately 30.7% of the total area as very suitable or suitable for agriculture (S1 and S2), 26.6% as marginally suitable (S3), and 42.7% as unsuitable (N1 and N2). Konya province had the largest area of suitable land, while Sivas had the largest area of unsuitable land. Comparison with the General Directory of Rural Service (GDRS) land capability classification revealed some differences, likely attributed to data quality, methodology, and changes in land use.

The study successfully integrated FAHP and GIS to assess land suitability for agriculture in Central Anatolia. The FAHP approach effectively handled the uncertainties inherent in expert judgments. The results highlight a significant portion of the region (42.7%) unsuitable for agriculture, emphasizing the sensitivity of the area to agricultural activities. Discrepancies with CORINE 2012 data suggest that agricultural activities are occurring in less-than-ideal areas. The weight values obtained for the criteria reflect the ecological context of the region, with slope, depth, and erosion having the highest importance due to their impact on soil erosion and agricultural practices. The moderate weights assigned to texture, organic matter, and bulk density emphasize their importance for soil quality and crop yield. The relatively low weights of pH, parent material, and lime content acknowledge the potential for improvements through land management practices. This study provides a valuable approach for regional-scale land suitability assessment.

This study provides a comprehensive assessment of land suitability for agriculture in Central Anatolia using FAHP integrated with GIS. The findings highlight the limitations of current agricultural practices in unsuitable areas and identify potentially suitable areas for sustainable agriculture. Future research could integrate economic, social, and political factors to refine the suitability assessment. The results can guide sustainable land resource management within the National Strategy and Action Plan.

The study focused primarily on biophysical factors and did not incorporate economic, social, or political considerations. The accuracy of the results depends on the quality of input data and the expertise of the decision-makers involved in the FAHP analysis. Future research should consider integrating more detailed data, such as climate data, socio-economic information, land use practices, and yield outputs, to enhance the accuracy and applicability of the model.