Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach

Upscaling instantaneous evapotranspiration retrieved at any specific time-of-day (ETi) to daily evapotranspiration (ETd) is a key challenge in mapping regional ET using polar orbiting sensors. Various studies have unanimously cited the shortwave incoming radiation (RS) to be the most robust reference variable explaining the ratio between ETd and ETi. This study aims to contribute in ETi upscaling for global studies using the ratio between daily and instantaneous incoming shortwave radiation (RSd ∕ RSi) as a factor for converting ETi to ETd. This paper proposes an artificial neural network (ANN) machine-learning algorithm first to predict RSd from RSi followed by using the RSd ∕ RSi ratio to convert ETi to ETd across different terrestrial ecosystems. Using RSi and RSd observations from multiple sub-networks of the FLUXNET database spread across different climates and biomes (to represent inputs that would typically be obtainable from remote sensors during the overpass time) in conjunction with some astronomical variables (e.g. solar zenith angle, day length, exoatmospheric shortwave radiation), we developed the ANN model for reproducing RSd and further used it to upscale ETi to ETd. The efficiency of the ANN is evaluated for different morning and afternoon times of day, under varying sky conditions, and also at different geographic locations. RS-based upscaled ETd produced a significant linear relation (R2 = 0.65 to 0.69), low bias (−0.31 to −0.56 MJ m−2 d−1; approx. 4 %), and good agreement (RMSE 1.55 to 1.86 MJ m−2 d−1; approx. 10 %) with the observed ETd, although a systematic overestimation of ETd was also noted under persistent cloudy sky conditions. Inclusion of soil moisture and rainfall information in ANN training reduced the systematic overestimation tendency in predominantly overcast days. An intercomparison with existing upscaling method at daily, 8-day, monthly, and yearly temporal resolution revealed a robust performance of the ANN-driven RS-based ETi up