Methodological considerations on local temperature zoning – where to install sensors, how many sensors and what interpolation method could be used? A case study in Morey-Saint-Denis and Chambolle Musigny (Bourgogne, France)

A critical sensor density threshold of at least 5 sensors per km² (40 sensors) was identified to ensure reliable local temperature zoning in vineyards, with Regression-Kriging outperforming other methods like Random Forest for high-density systems. The study, conducted over 18 months in Burgundy's Morey-Saint-Denis and Chambolle Musigny regions using 112 grape-cluster-level sensors, revealed that interpolation accuracy for maximum temperatures (Tmax) is significantly lower than for minimum temperatures (Tmin), suggesting micro-scale factors override topographical signals at high planting densities. While 40–60 sensors suffice for mapping frost-related indices, higher densities or additional covariates are needed for heat stress and thermal accumulation indices.

Bois, B., Nicolas, S., Roig Puscama, F., Giraudo, J., Rega, M., Gavrilescu, C., Jacquin-Ponsort, W., & Ouvrié, M. (2026). Methodological considerations on local temperature zoning – where to install sensors, how many sensors and what interpolation method could be used? A case study in Morey-Saint-Denis and Chambolle Musigny (Bourgogne, France). OENO One, 60(3). https://doi.org/10.20870/oeno-one.2026.60.3.9882