Continuous estimation of evapotranspiration and gross primary productivity from an Unmanned Aerial System

Model prediction of evapotranspiration (ET) and gross primary productivity (GPP) using optical and thermal satellite imagery is biased towards clear-sky conditions. Unmanned Aerial Systems (UAS) can collect optical and thermal signals at unprecedented very high spatial resolution (< 1 meter) under sunny and cloudy weather conditions. However, methods to obtain model outputs between image acquisitions are still needed. This study uses UAS based optical and thermal observations to continuously estimate daily ET and GPP in a Danish willow forest for an entire growing season of 2016. A hexacopter equipped with multispectral and thermal infrared cameras and a real-time kinematic Global Navigation Satellite System was used. The Normalized Differential Vegetation Index (NDVI) and the Temperature Vegetation Dryness Index (TVDI) were used as proxies for leaf area index and soil moisture conditions, respectively. To obtain continuously daily records between UAS acquisitions, UAS surface temperature was assimilated by the ensemble Kalman filter into a prognostic land surface model (Noilhan and Planton, 1989), which relies on the force-restore method, to simulate the continuous land surface temperature. NDVI was interpolated into daily time steps by the cubic spline method. Using these continuous datasets, a joint ET and GPP model, which combines the Priestley-Taylor Jet Propulsion Laboratory ET model (Fisher et al., 2008; Garcia et al., 2013) and the Light Use Efficiency GPP model (Potter et al., 1993), was applied. The simulated ET and GPP were compared with the footprint of eddy covariance observations. The simulated daily ET has a RMSE of 14.41 W•m-2 and a correlation coefficient of 0.83. The simulated daily GPP has a root mean square error (RMSE) of 1.56 g•C•m-2•d-1 and a correlation coefficient of 0.87. This study demonstrates the potential of UAS based multispectral and thermal mapping to continuously estimate ET and GPP for both sunny and cloudy weather conditions.