Predicting evapotranspiration from drone-based thermography – a method comparison in a tropical oil palm plantation

For the assessment of evapotranspiration, near-surface airborne thermographyoffers new opportunities for studies with high numbers of spatial replicatesand in a fine spatial resolution. We tested drone-based thermography and thesubsequent application of the DATTUTDUT energy balance model using the widely accepted eddy covariance technique as a referencemethod. The study site was a mature oil palm plantation in lowland Sumatra,Indonesia. For the 61 flight missions, latent heat flux estimates of theDATTUTDUT (Deriving Atmosphere Turbulent Transport Useful To Dummies UsingTemperature) model with measured net radiation agreed well with eddy covariancemeasurements (r2 = 0.85; MAE = 47; RMSE = 60) acrossvariable weather conditions and times of day. Confidence intervals for slope andintercept of a model II Deming regression suggest no difference betweendrone-based and eddy covariance methods, thus indicating interchangeability.The DATTUTDUT model is sensitive to the configuration of the net radiation assessment. Overall, weconclude that drone-based thermography with energy balance modeling is areliable method complementing available methods for evapotranspirationstudies. It offers promising, additional opportunities for fine grain andspatially explicit studies.