Quality Prediction of a Tunisian Virgin Olive Oil Based on the Effect of Climate Change Using ANFIS Method

ABSTRACT An assessment of the interannual influence of climatic factors, namely temperature, precipitation, and humidity, on the quality of Tunisian virgin olive oil from the Zalmati cultivar was conducted using samples collected over three consecutive crop seasons (2018/2019–2020/2021), resulting in a total of 28 samples representing different ripening stages. An adaptive neuro‐fuzzy inference system (ANFIS) model was developed to predict key olive oil quality parameters, including oleic acid, chlorophyll, and total phenols, based on climatic data. The ANFIS model demonstrated good predictive performance, with root mean square error (RMSE) values of 0.76, 0.87, and 12.40 during training, and 1.12, 0.79, and 24.68 during testing for oleic acid, chlorophyll, and phenols, respectively. Mean absolute percentage error (MAPE) values remained low, confirming the robustness of the model. A strong agreement between experimental and predicted results was observed. These findings demonstrate the potential of ANFIS as a reliable predictive tool for assessing climate‐driven variations in olive oil composition and may support climate‐adaptive strategies in olive production. Practical Application : This study provides a practical predictive tool for evaluating the impact of climatic variability on virgin olive oil quality under arid Mediterranean conditions. The developed ANFIS model can support olive growers and producers in anticipating changes in key quality parameters such as oleic acid, chlorophyll, and phenolic content based on climatic data. This predictive framework may assist in optimizing harvest timing, irrigation management, and crop adaptation strategies under climate change conditions. Furthermore, the model may support policymakers and agricultural stakeholders in designing sustainable management practices for olive production in vulnerable regions.