Dynamics of evapotranspiration over a non-irrigated alfalfa field in the Southern Great Plains of the United States

Accurately quantifying the dynamics of evapotranspiration (ET) is crucial for efficient water management and improved water use efficiency. However, details on the magnitudes and annual dynamics of ET with respect to environmental/biophysical factors and harvesting of hay in non-irrigated alfalfa (Medicago sativa L.) are lacking. Using the eddy covariance (EC) technique, daily magnitudes and seasonal/annual dynamics and budgets of ET were quantified from April 2016 to May 2018 over a non-irrigated alfalfa field in central Oklahoma, USA. The field was harvested periodically for hay, and cumulative dry forage yield was approximately 7.5 and 10 t ha(-1) in 2016 (dry year) and 2017 (wet year), respectively. Daily ET reached up to 6.9mm d(-1) and 8-day average ET reached up to 5.64mm d(-1). Cumulative seasonal (April-October) ET was 652mm (similar to 1.3 times of precipitation) in 2016 and 734mm (similar to 0.8 times of precipitation) in 2017. Annual ET in 2017 was similar to 900mm (similar to 0.8 times of annual precipitation). Optimum air temperature (T-a) and vapor pressure deficit (VPD) for ET were approximately 30 degrees C and 3 kPa, respectively. Higher forage production was associated with a greater increase (similar to 22%) in carbon uptake (gross primary production, GPP) than ET (similar to 13%) in 2017 compared to 2016. Consequently, ecosystem water use efficiency (EWUE) at the seasonal scale (seasonal sums of GPP to ET) was 2.38 and 2.57 g C mm(-1) ET in 2016 and 2017, respectively. Despite strong correspondence (R-2 = 0.73) between EC-measured ET and Moderate Resolution Imaging Spectroradiometer (MODIS)-derived ET (ETMOD16), the standard ETMOD16 product underestimated ET by 36% compared to EC-measured ET. The MODIS-derived enhanced vegetation index (EVI) and photosynthetically active radiation (PAR) explained 83% of variations in alfalfa ET, indicating the potential of integrating remote sensing observations and climate data to extrapolate site-level alfalfa ET at larger areas.