Technical note: Investigating the potential for smartphone-based monitoring of evapotranspiration and land surface energy-balance partitioning

Abstract

Evapotranspiration plays a key role in the terrestrial water cycle, climate extremes, and vegetation functioning. However, the understanding of spatio-temporal variability of evapotranspiration is limited by a lack of measurement techniques that are low cost and that can be applied anywhere at any time. Here we investigate the estimation of evapotranspiration and land surface energy-balance partitioning by only using observations made by smartphone sensors. Individual variables known to effect evapotranspiration as measured by smartphone sensors generally showed a high correlation with routine observations during a multiday field test. In combination with a simple multivariate regression model fitted on observed evapotranspiration, the smartphone observations had a mean RMSE of 0.10 and 0.05 mm h-1 during validation against lysimeter and eddy covariance observations, respectively. This is comparable to an error of 0.08 mm h-1 that is associated with estimating the eddy covariance ET from the lysimeter or vice versa. The results suggests that smartphone-based ET monitoring could provide a realistic and low-cost alternative for real-time ET estimation in the field.