Precision soil sampling strategy for the delineation of management zones in olive cultivation using unsupervised machine learning methods.

Climate change and environmental degradation pose a significant threat to the global community. Soil management is one of the critical factors for achieving climate neutrality, as plants and soils together currently absorb approximately 30% of the CO emitted by human activities each year. This study focused on delineating soil management zones in olive groves to maintain soil health in complex environmental conditions and minimize adverse effects on the biological systems supported. The results of this study are crucial because they showed the potential of unsupervised machine learning techniques in this setting and important soil characteristics for defining management zones. They might significantly affect applying precision farming techniques and methods in olive groves, providing a possible remedy for the problems caused by climate change. A total of 222 soil samples at a depth of 0-30 cm were collected from three areas in the Region of Western Greece, at a density of 21 × 21 m in each area, and analyzed for physicochemical properties. Principal Component Analysis (PCA) was utilized to identify the critical soil properties for delineating the management zones. The soil samples were clustered using unsupervised machine learning methods, K-means, Hierarchical clustering, and DBSCAN. PCA is a method that can help in the selection of critical parameters for the delineation of management zones. Sand (S), Clay (C), Cation Exchange Capacity (CEC), Potassium (K), Calcium (Ca), Soil Organic Carbon (SOC), and total carbonates (CaCO) can delineate management zones in olive cultivation. The management zones and how each field is separated vary depending on the clustering method.