Advancing the monitoring of soil-plant-atmosphere interactions: insight into Cosmic-Ray Neutron Sensing and multi-sensor monitoring in vineyards and orchards.

The sustainable management of water in agriculture demands accurate and continuous monitoring of soil and plant water status across multiple spatial and temporal scales. This thesis develops and validates a monitoring framework integrating Cosmic‑Ray Neutron Sensing (CRNS) with plant physiological and atmospheric measurements to quantify soil water content (SWC) and its influence on crop functioning within the soil–plant–atmosphere continuum (SPAC). The work is structured in three experimental chapters addressing (1) CRNS‑based SWC detection under contrasting irrigation systems, (2) eco‑physiological interpretation of soil and atmospheric drivers in vineyards, and (3) integration of continuous multi‑sensor data to model SPAC dynamics. At two commercial sites in northern Italy (a subsurface drip‑irrigated vineyard and a micro‑sprinkler‑irrigated walnut orchard), CRNS was calibrated using soil sampling and corrected for atmospheric pressure, humidity, and cosmic‑ray intensity. CRNS proved sensitive to irrigation‑induced SWC changes, capturing short‑term variations in the main root zone and outperforming point‑scale sensors by providing smoother, field‑averaged patterns. In the walnut orchard, irrigation heterogeneity reduced CRNS correlation strength, highlighting its sensitivity to spatial and temporal water distribution. The vineyard study linked CRNS‑derived extractable soil water (ESW) with midday stem water potential (Ψstem), berry composition, vapor pressure deficit (VPD), and NDVI. In a wet year (2023), vines showed limited water stress and weak coupling with atmospheric demand; in the drier 2024 season, more negative Ψstem and stronger VPD links accelerated sugar accumulation. Principal Component Analysis (PCA) revealed ESW and VPD as dominant drivers of vine performance. In Renmark, South Australia, continuous CRNS, Ψstem, sap flow, and meteorological data were combined and modelled with Linear Mixed‑Effects Models (LMMs). VPD emerged as the primary predictor of vine water status, while SWC enhanced night‑time rehydration. These results demonstrate that an integrated CRNS‑based framework can support adaptive irrigation and holistic water‑governance strategies in Mediterranean and semi‑arid regions.