Reasonable estimation of reference evapotranspiration (ET_o) requires some climatic inputs which might be missing in areas with sparse data recording. This study aimed to assess performance of FAO56 Penman-Monteith (PM-ET_o) fed by ERA5, MERRA2 and GLDAS2 outputs in estimating daily and monthly ET_o under data limitation. The accuracy of PM-ET_o calculated by interpolated factors and the temperature-based PM-ET_o (PMT) was also studied. Additionally, performance of PM-ET_o fed by the bias-corrected reanalysis products against the PMT with updated constant, i.e. recalibrated PMT, was investigated. Climatic data required to run PM-ET_o were collected from 146 stations over Iran for 25 years. Results revealed that ERA5 provides more realistic daily and monthly ET_o estimates relative to MERRA2 and GLDAS2 in 84% of cases. Furthermore, ERA5 surpassed the others in producing daily and monthly wind speed, vapor pressure deficit and mean temperature for the majority of locations. The average relative Mean Bias Error (rMBE) of − 7.3% and 8.1% at monthly scale and of − 11.1% and 9.8% at daily scale were found for MERRA2- and GLDAS2-estimated ET_o, respectively, indicating ET_o overestimation and underestimation by MERRA2 and GLDAS2, respectively. The ERA5 provided more satisfactory results, with normalized Root Mean Square Error of 15.2% and 22.7% for daily and monthly steps, respectively, relative to PMT for approximately 70% of sites. Moreover, ET_o estimated by ERA5 had a smaller nRMSE than that simulated using the interpolated variables in around 60% of the sites. Therefore, under temperature data availability or existence of nearby sites, application of ERA5 is better suited to estimate ET_o in our study area. The PM-ET_o fed by bias-corrected ERA5 outputs also outperformed recalibrated PMT, illustrating that bias-correction seems to be a more accurate modification when complete datasets are available at least for a limited time. Overall, ERA5 products are robust surrogates for simulating ET_o under data limitation on different temporal resolutions which is needed for decision making and planning processes.

参考蒸散量 (ET _o ) 的合理估计需要一些气候输入，这些输入在数据记录稀少的地区可能会丢失。本研究旨在评估由 ERA5、MERRA2 和 GLDAS2 输出提供的 FAO56 Penman-Monteith (PM-ET _o )在数据限制下估算每日和每月 ET _{o 的性能}。还研究了通过内插因子计算的 PM-ET _o和基于温度的 PM-ET _o (PMT)的准确性。此外，PM-ET的性能_Ó由偏置校正的再分析产品针对与更新常数PMT馈送，即重新校准PMT，进行了调查。气候数据需要运行PM-ET _Ø25 年来从伊朗上空的 146 个站点收集。结果显示，在 84% 的案例中，ERA5 提供了比 MERRA2 和 GLDAS2更现实的每日和每月 ET _o估计值。此外，ERA5 在大多数地区产生的每日和每月风速、蒸汽压差和平均温度方面都超过了其他地区。MERRA2 和 GLDAS2 估计的 ET _o的平均相对平均偏差误差 (rMBE) 分别为每月规模的 − 7.3% 和 8.1% 以及每日规模的 − 11.1% 和 9.8% ，表明 ET _oMERRA2 和 GLDAS2 分别高估和低估。ERA5 提供了更令人满意的结果，相对于大约 70% 的站点的 PMT，每日和每月步骤的归一化均方根误差分别为 15.2% 和 22.7%。此外，ERA5 估计的ET _o的 nRMSE 小于使用内插变量在大约 60% 的站点中模拟的 nRMSE。因此，在温度数据可用或附近站点存在的情况下，ERA5 的应用更适合于估计我们研究区的ET _o。PM-ET _o由偏差校正的 ERA5 输出馈送的结果也优于重新校准的 PMT，说明当完整的数据集至少在有限的时间内可用时，偏差校正似乎是一种更准确的修改。总体而言，ERA5 产品是在数据限制下模拟 ET _{o 的}强大替代品，不同时间分辨率是决策和规划过程所需的。