IS-SAR: an irrigation scheduling web application for Hass avocado orchards based on Sentinel-1 images

Abstract

AbstractAs the Hass avocado crop expands exponentially in Colombia, concern about its increasing water use is on the rise. This research aimed to develop IS-SAR, a free-access web application to schedule irrigation for Valle del Cauca’s Hass growers. We calibrated the water cloud (WCM) and artificial neural network (ANN) models using field data measurements from Hass avocado orchard plots in Valle del Cauca (Colombia) and Sentinel 1 (S1) satellite imagery measurements and evaluated their performance computing the root-mean-square error (RMSE) and Pearson correlation coefficient ($$r$$ r ). IS-SAR estimates the surface soil water content from the most recent S1 image, becomes it in water depth, and recommends to users apply irrigation according to allowable depletion limits, computed from a spatially distributed $$\theta$$ θ at field capacity and permanent wilting point obtained from a soil database of the study area. Our results indicate that the surface soil water content was retrieved with a better performance by the ANN (RMSE = 0.05 m3 m−3, $$r=0.74$$ r = 0.74 ), compared with the WCM (RMSE = 0.06 m3 m−3, $$r=0.47$$ r = 0.47 ). IS-SAR simulations in validation orchard plots result in irrigation events of up to 107 L tree−1 for 3.4 h. The IS-SAR web application provides near-real-time irrigation information to assist Hass avocado growers in designing better irrigation routines and improving the regional understanding of water crop consumption.