Estimation of evapotranspiration from ground-based meteorological data and global land data assimilation system (GLDAS)

Evapotranspiration (ET) is one of the most significant factors in understanding global hydrological budgets, and its accurate estimation is crucial for understanding water balance and developing efficient water resource management plans. For calculation of reference ET (ETref), the meteorological data from weather stations have been widely used for estimation at the point scale; however, meteorological data from the global land data assimilation system (GLDAS) at the regional scale are rarely used for the estimation of ET. In this study, 30 different equations provided in the Reference Evapotranspiration Calculator Software (REF-ET) were utilized for estimating ETref with GLDAS and point scale data collected at 14 observation sites in the Korean Peninsula during 2013. Using ETref calculated from observation and GLDAS, 30 equations were evaluated by estimating the overall rank number, as determined by the correlation coefficient, normalized standard deviation, bias, and root mean square error (RMSE). Results showed that the Penman (Proc R Soc Lond Ser A Math Phys Sci 193:120-145, 1948) FAO-56 Penman-Monteith, 1982 Kpen equation (combination equations), the 1957 Makkink, Priestley-Taylor equation (radiation based equation), and the 1985 Hargreaves equation had a good overall rank. Using the six selected equations, seasonal analysis was conducted and evaluated using the bias and RMSE. Comparison of the ETref gathered from observation and GLDAS revealed that both of them showed similar seasonal variation, although ETref calculated from GLDAS were underestimated. Sensitivity analysis conducted by changing three main climatic variables (i.e., temperature, wind speed, and sunshine hours) by +/- 1, +/- 5, +/- 10, +/- 15, and +/- 20 % with one variable fixed also revealed that ETref was more affected by air temperature than sunshine hours and wind speed throughout the 14 selected stations.