Abstract
<title>Abstract</title> <p>Global climate change seriously threatens agricultural production, which is directly dependent on biological processes and open-air conditions. The purpose of this study is to examine the causal relationship between climate variables (temperature, humidity, and precipitation) and apricot yield using data from the 1991–2024 period in Malatya province, which is the most important apricot production center in Türkiye and the world. Unlike traditional studies in the literature, alongside the Toda-Yamamoto causality test that fixes the relationship between series with a single coefficient, the dynamic Bootstrap Rolling Window causality analysis, capable of capturing structural breaks and periodic climate shocks, was utilized. According to the analysis findings, a general causal relationship covering the entire analysis period could not be detected as a result of the Toda-Yamamoto test. However, the dynamic Bootstrap Rolling Window approach determined strong causalities during crisis periods when climatic stresses intensified. It was found that in the years 2006, 2007, 2008, and 2010, when precipitation levels fell below the critical threshold, the yield was positively and directly affected. During the 2008–2011 period, when winter and early spring temperatures increased, a significant negative causality on yield was identified due to early flowering and the risk of frost. Humidity ratio, on the other hand, was observed to play a conditioning role and have a positive effect only during the recovery process in 2015 following the severe frost disaster in 2014. Consequently, static causality models overlook the periodic shocks of climate change on agricultural yield. By demonstrating that agricultural risk management and climate change adaptation strategies should be handled in a dynamic structure rather than a fixed one, this study presents a new methodological perspective to the time series literature.</p>