Water Stress Estimation from Leaf Turgor Pressure in `Arbequina' Olive Orchards Based on Linear Discriminant Analysis

Abstract

In this work, we focus on estimating water stress based on leaf turgor pressure in `Arbequina' olive orchards. Currently, this analysis is time-consuming and primarily conducted by experts, rendering it impractical for integration into commercial deficit irrigation applications. Our study leverages a dataset that includes leaf turgor pressure measurements and meteorological variables collected between 2014 and 2019 from an orchard in Spain. We propose an innovative automated system based on machine learning techniques to classify trees into three distinct water stress levels by analyzing their daily trajectory. In this approach, both the probe and meteorological data undergo normalization before classification, employing principal component analysis and linear discriminant analysis.

Keywords: IoT, machine learning, expert system, Turgor pressure, Stem water potential, Irrigation scheduling