Comparison of advanced multicriteria decision and FAO models for land suitability assessment 

Land suitability assessment is an important process in modern agricultural management, involving the evaluation of various factors such as soil properties, climate, relief, hydrology, crop varieties and socio-economic considerations. Various methods have been used to assess land suitability, such as the parametric method developed by Sys et al. (1991) and the FAO (1976) approach to land evaluation. Determining the relative weighting of the factors that influence land suitability is a particularly challenging step in the evaluation process. An alternative to these procedures is the use of multi-criteria decision making (MCDM) techniques, which enable land managers and policy makers to make informed decisions about land use and development. Spatial MCDM techniques include complex spatial data and methods such as the Technique of Preference Ordering by Similarity to the Ideal Solution (TOPSIS), which are widely used in the agricultural sector. TOPSIS determines the optimal alternative according to the principle of minimizing the proximity to the ideal solution and maximizing the distance to the negative ideal solution. The aim of this study was to assess the suitability of land for wheat cultivation in western Iran, a country facing the challenge of becoming self-sufficient in wheat. Seventy soil profiles were selected and described on the basis of a geomorphologic map and the content of various soil properties and wheat yield were determined. MCDM (TOPSIS) and FAO models were applied and evaluated according to wheat yield. The Shannon entropy method (SHE) was used to extract the criteria weights. Land suitability assessment was mapped using a Random Forest machine learning model and auxiliary variables. According to the results of the Shannon entropy method, slope, cation exchange capacity (CEC) and calcium carbonate equivalent (CCE) are the most important criteria for wheat cultivation. Furthermore, the results are also confirmed by the spatial autocorrelation between the key criteria and wheat yield. These results also show that the soil suitability values calculated with the TOPSIS model have a higher correlation with wheat yield than the values calculated with the FAO model (0.73 and 0.67, respectively). The spatial distribution of the suitability values for wheat cultivation showed that 30 to 33% of the areas were very suitable, 13-16% moderately suitable and 51% and 57% unsuitable. For the areas with high and medium suitability, the TOPSIS and FAO results were largely in agreement, in contrast to the areas with low suitability. This study provided a comprehensive approach to land suitability for wheat cultivation using advanced MCDM techniques and machine learning, which can be beneficial for sustainable land management and food security in Iran and similar regions.