Using remote sensing for precise nitrogen fertilization to estimate the spatial variability in yield and fruit quality in avocados
Research team: Dr. Tarin Paz-Kagan (BGU, PI), Dr. Ran Erel (ARO, Co-PI), Dr. Shahar Baram (ARO, CI), and Dr. Noam Alkan (ARO, CI).
Today, farmers do not have the tools to identify excess nitrogen fertilization. Therefore, the efficiency of nitrogen application in orchards is often low since there is no reference to the existing variation, and the application is made uniformly. Our previous studies identified the relationship between nitrogen concentrations in the canopy and multispectral information that can detect nitrogen excess/deficiency in the plant and map interface areas for precise fertilization in orchards and almonds. In this study, we propose to rely on previous research efforts that provided a unique infrastructure and a combination of controlled experiments, field experiments, and commercial plots to construct large databases and advanced statistical models to develop a decision-support tool for nitrogen fertilization in avocados. This study will make it possible to point out the relationship between nitrogen fertilization - and development indicators – fertility, and fruit quality in a commercial avocado orchard to optimize inputs and promote sustainable agriculture. We will calibrate statistical models guided by machine learning to predict nitrogen values in avocado tree canopies. Later, we will develop artificial intelligence-based models for evaluating the seasonal changes in the nitrogen content and link them to measures of development, yield, and fruit quality in avocados. We will examine the effect of nitrogen fertilization on N2O fluxes from avocado tree leaves. Finally, we will draw up recommendations for sampling in commercial areas and an original decision-making system for optimal nitrogen fertilization for the accuracy and timing of fertilization in orchards. This is by studying the nutrient needs at the level of the tree in the field and examining their relationship to the crop and its quality. Thus, by combining frequent measurements on a large scale and using several experimental systems, we will establish extensive databases describing several fertilization levels and diverse growing conditions. This information will support decision-making systems to optimize fertilization in avocado orchards with tools based on remote sensing and artificial intelligence.
