Climate variability impacts on rainfed cereal yields in west and northwest Iran.

In order to assess the response of wheat and barley to climate variability, the correlation between variations of yields with local and global climate variables was investigated in west and northwest Iran over 1982-2013. The global climate variables were the El Niño-Southern Oscillation (ENSO), Arctic Oscillation (AO), and North Atlantic Oscillation (NAO) signals. Further, minimum (T min ), maximum (T max ), and mean (T mean ) temperature, diurnal temperature range (DTR), precipitation, and reference evapotranspiration (ET0 ) was used as local weather factors. Pearson's correlation coefficient was applied to analyze the relationships between climatic variables and yields. Unlike T min , T mean , ET0 , and T max , the yields were significantly associated with the entire growing season (EGS) DTR in most sites. Therefore, considering weather extreme variables such as DTR sheds light on the crop-temperature interactions. It is also found that the April-May-June (AMJ), October-November-December (OND), and EGS rainfall variations markedly influence the yields. Unlike the AO and NAO indices, the Niño-4 and SOI (the ENSO-related signals) were significantly correlated with the OND and EGS precipitation and DTR. Thus, the ENSO anomalies highly impact rainfed yields through influencing the OND and EGS rainfall and DTR in the studied sites. As the correlation coefficient of the OND and July-August-September (JAS) Niño-4 with yields was significant (p < 0.05) for almost all locations, the JAS and OND Niño-4 may be a good proxy for cereal yield forecasting. Further, an insignificant increment and a significant reduction in yields are expected in La Niña and El Niño years, respectively, relative to neutral years.