Climate variability plays a pivotal role in rainfed agriculture, especially within arid regions. Analyzing these fluctuations across diverse climatic conditions establishes a foundation for subsequent investigations. In Iran, the FAO56 aridity index categorizes the nation into very dry, dry, semidry, and humid climate classifications. This study aimed to explore equations derived from multiple linear regression (MLR) and the disparities between predicted and observed yields of rainfed wheat and barley across Iran's varying climates. Meteorological data, encompassing rainfall (R), mean temperature (Tmean), solar radiation (S), and wind speed (U2), were compiled from 44 synoptic stations spanning 1981–2020. These data constituted inputs for the MLR models employed to simulate rainfed wheat and barley yields. The Global Performance Indicator (GPI), a 5-point statistical criteria index, was utilized to assess MLR model performance. The findings unveiled superior MLR model performance in dry climates (R2=0.84 for wheat and R2=0.9 for barley) compared to humid climates (R2=0.69 for wheat and R2=0.66 for barley), evidenced by lower statistical error criteria values. Moreover, across all climates, the MLR models exhibited more accurate predictions for rainfed wheat yield (GPI=1559.3) in contrast to rainfed barley (GPI=1536). In conclusion, this study sheds light on the notable role of climate in rainfed agriculture, showcasing the efficacy of MLR models in predicting yields across varying climatic contexts.
