Search

As 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 ). 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 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 ), compared with the WCM (RMSE = 0.06 m3 m−3, 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. [ABSTRACT FROM AUTHOR]

This study aims to determine whether Hass avocado irrigation can be triggered based on the surface soil water content (SSWC). To address this question, the soil water dynamics from three Hass avocado orchard plots located in Valle del Cauca (Colombia) was simulated using a model provided by Hydrus-1D software, which was calibrated through the genetic algorithm NSGA-II and validated using the soil matric potential measured at several depths at nine monitoring stations installed in the three plots. The influence of each superficial (0–5, 5–10, and 10–15 cm) and deeper (15–30 and 30–60 cm) available water (AW) computed from the simulated moisture on the SWB at 0–60 cm was estimated from the artificial neural network (ANN) trained weights. The most influential depth range was used to predict the soil water balance at 0–60 cm using ANN. For validation, the RMSE slightly increased regarding calibration, varying from 1.73 to 8.20, while the R2 value varied from 0.61 to 0.89 (P < 0.001 for all cases). The AW at 5–10 cm depth had a significant influence on SWB with an average relevance index of 2.87 (Wilcoxon signed-rank test P ≤ 0.05) for Laurentina farm. The AW at 0–5 cm depth had not significant influence on SWB with an average relevance index of 1.34 (independent group) and 0.97 (P < 0.05) for Laurentina and Poncena, respectively. The ANN model predicted the SWB with a RMSE no bigger than 13.76 mm. In conclusion, the SSWC at a depth of 5–10 cm can be used as an indicator for scheduling Hass avocado irrigation. • Irrigation and precipitation modify differently the soil matric potential. • Water changes significantly in the top 15 cm of soils cropped with Hass avocado. • Soil water at 5–10 cm depth range influences the avocado irrigation scheduling. • Soil water in the top 5 cm does not influence the avocado irrigation scheduling. • Genetic algorithm NSGA-II improves the performance of Hydrus 1D calibration. [ABSTRACT FROM AUTHOR]