Suitability evaluation of potential arable land in the Mediterranean region

Abstract

The existing cultivated land in the Mediterranean region faces great pressure from various sources. A suitability evaluation of potential arable land is urgent for helping adaptation measures to mitigate the impacts of climate change and human pressure on agricultural production in the Mediterranean region. We integrated 15 biophysical and socio-economic factors from GIS and remote sensing data to perform a suitability evaluation of potential arable land in the Mediterranean region using analytical hierarchy process and radial basis function artificial neural network methods. Moreover, we analyzed the gap between potential arable land and existing cultivated land and compared the evaluation results between the analytical hierarchy process and artificial neural network methods. The results show that the suitability index of potential arable land based on artificial neural network with 6 neurons has the best correlation with average yield and average harvested area. The land area with a suitability grade over medium level accounts for 62.95% of the potential arable land area, of which 45.71% is uncultivated land. Cyprus, France, Greece, Italy, Lebanon, Portugal, Spain and Turkey have great opportunities for agricultural development. Radial basis function artificial neural network outperforms analytical hierarchy process, has better verification results, and requires less input. This study provides an initial insight into the agricultural land suitability of 16 countries around the Mediterranean Sea and introduces a research idea for agricultural land suitability evaluation.

Keywords: Agricultural land suitability evaluation; Analytical hierarchy process (AHP); CRiteria importance through intercriteria correlation (CRITIC); Mediterranean; Radial basis function (RBF).