Challenges in water stress quantification using small unmanned aerial system (sUAS): Lessons from a growing season of almond

Abstract

With water shortages and drought affecting many regions of the world, it becomes urgent to increase water use efficiency (WUE) by optimizing irrigation schedule. Proper irrigation scheduling, which includes integrating of soil moisture monitoring, surface evapotranspiration loss calculation, and plant based measurements is required for high WUE. Stem water potential (SWP) has become one of the more common methods to measure water status. It is, however, labor intensive and time consuming, and adoption has been slow. This study aims to build the link between SWP and canopy normalized difference vegetation index (NDVI) based on aerial multi-spectral images and ground-truth measurement of an almond orchard. Data suggests that the correlation between SWP and canopy NDVI can be improved by tuning canopy NDVI threshold, as indicated by the coefficient of determination (R2). Also, NDVI shows good correlation with SWP in different growing stages — fruit development and post-harvest. Finally, it is demonstrated canopy NDVI distribution from different missions are significantly different, even if the interval between two flights is less than one hour. This poses the challenge that further calibration is needed to conduct quantitative measurement in long flight missions. Meanwhile, quantitative consideration of characteristic of bi-directional reflectance distribution function makes it necessary to obtain stable performance of canopy NDVI.