Evaluation of UV–Vis–NIR model transferability for predicting grapevine water status across Spanish viticultural regions

Abstract

Climate change is exerting unprecedented pressure on global viticulture, altering precipitation patterns, intensifying drought events, and making efficient irrigation management a critical challenge for sustainable production. In this context, the development of rapid and cost-effective tools to assess vine water status is essential to support precision irrigation strategies and optimize water use. Ultraviolet-visible (UV–Vis) and near-infrared (NIR) spectroscopy, combined with chemometric modeling, have emerged as promising techniques for estimating stem water potential (Ψ_stem), a key physiological indicator of plant hydration. In previous work, a general predictive model was developed using UV–Vis–NIR spectral data from vineyards in Galicia (Spain) collected over three growing seasons, achieving a strong predictive performance (R²p = 0.85; RMSEP = 0.17 MPa). This model demonstrated robustness across diverse varieties and climatic conditions. The present study extends that research by evaluating the transferability and generalization capacity of this model to new datasets from two contrasting viticultural regions in Spain: Enguera (Valencia), characterized by a dry Mediterranean climate, and Quinta Couselo (Pontevedra), located in a humid Atlantic zone. By comparing the predictions of the general model with newly developed region-specific models, the study aims to identify how climatic variability affects model performance and calibration stability.

Ultimately, this work seeks to provide insights into the scalability and reliability of spectroscopy-based tools for vineyard management. Demonstrating the capacity of a generalizable UV–Vis–NIR model to predict vine water status under different environmental conditions would represent a significant advance toward adaptive irrigation management, contributing to resilient and sustainable viticulture in a changing climate context.

Funding

Grupo Operativo Supraautonómico GO-PRERIVID. Predicción de necesidades hídricas del viñedo para un uso sostenible del agua de riego. Funding: Ministerio de Agricultura, Pesca y Alimentación (MAPA) y (FEADER). Partners: USC, ICVV-CSIC, ITACYL and CIDE and companies: Monet, Quinta Couselo, Quinta Sardonia, Bodegas Enguera and FEUGA.