Assessing climate change effects on Turkish tea farming through a dual approach using MMQR and machine learning

Abstract

Abstract Climate change increasingly threatens the productivity of region-specific strategic agricultural products such as tea cultivation in Türkiye, posing a serious risk to both food security and rural economies. However, existing literature is notably limited in terms of studies that draw attention to this risk and examine the effects of climate change on tea productivity at a regional scale through rigorous quantitative methods. To this end, this study investigates the influence of climate change on tea productivity in Türkiye’s tea–growing provinces (Artvin, Giresun, Ordu, Rize, and Trabzon) between 2004 and 2022. Distinct from previous studies, we integrate advanced machine learning techniques with the method of moments quantile regression (MMQR) approach to provide comprehensive, reliable, and methodologically robust results for the first time in this context. The results of the MMQR demonstrate that although humidity reduces tea productivity, temperature and precipitation significantly increase it. Furthermore, the results of machine learning research indicate that the tea farming area is the variable with the highest importance, whereas humidity emerges as the least influential factor. These findings indicate that policymakers need to implement integrated agricultural policies in the five tea–growing provinces of the Eastern Black Sea region, including effective moisture management, soil fertility, erosion control, and irrigation infrastructure tailored to the climate and land conditions.