Sensitivity of reference evapotranspiration to weather variables across seven regions of Turkey

Daily weather data from 1984 to 2018 for 14 weather stations representing seven regions of Turkey were used to estimate reference evapotranspiration (ETref). We performed sensitivity analyses of input weather variables (air temperature, wind speed, and solar radiation) to determine the relative effects of measurement errors in weather data on the accuracy of calculated grass (ETos) and alfalfa (Medicago sativa L., ETrs) reference ET. In addition, changes in estimated irrigation water requirement for maize (Zea mays L.), a major crop in all regions of Turkey, due to errors in major climatic variables were also determined across seven regions of Turkey. Comparison of the sensitivity coefficients indicated that ETref was most sensitive to wind speed (sensitivities from -2.46 to 5.56 mm d(-1)), followed by air temperature (sensitivities from -2.13 to 3.1 mm d(-1)) across Turkey. However, responses of ETos and ETrs to changes in weather variables varied across the regions. For examples, the Marmara region had the largest impact of errors in air temperature but the smallest impact of errors in wind speed on the calculated ETref. Likewise, the Southeastern and Eastern Anatolia regions had the greatest impact of errors in wind speed but the smallest impact of errors in air temperature on the calculated ETref. The errors in weather variables led to substantial over- or underestimation of irrigation water requirement for maize. In some cases, irrigation water requirement was overestimated by more than double. These results suggest that different sensor accuracies are needed for weather stations in different regions of Turkey to accurately estimate ETref and irrigation water requirement.