Grape Expectations: Disentangling Environmental Drivers of Microbiome Establishment in Winegrowing Ecosystems

Microbial communities play an integral role in agricultural systems, where their composition and function are shaped by environmental factors and interactions with the plant host. In viticulture and winemaking, the resulting distinct microbial biogeographic patterns influence wine characteristics – a well-established concept termed microbial terroir . Yet, due to the complexity of interactions, the specific environmental drivers shaping these microbiomes remain poorly understood.

We conducted a multi-year, large-scale survey (N = 680 samples) of Swiss vineyards (N = 95 sites), following a subset (N = 12, within 2.46 km) longitudinally over 3 years. We integrate microbiome (bacterial and fungal marker-gene sequencing), untargeted metabolomics (GC-MS and LC-MS/MS), environmental monitoring, and sensory data to disentangle abiotic factors influencing community assembly and interactions in vineyards and in wine fermentations.

Results show that topography and climate collectively structure microbial communities, yet exhibit distinct influences on soil- and plant-associated microbiomes. Berry-associated fungal communities exhibit the strongest site-specific imprint, enabling machine-learning-based prediction of even subtle microclimatic differences. Climatic factors and berry chemistry exhibit an inverse relationship, selectively favoring either Hanseniaspora sp. or Saccharomyces cerevisiae – each correlating with distinct metabolite and aroma profiles. Additionally, plant stress-response metabolites were associated with shifts in microbial composition and fermentation outcomes.

By integrating multi-omics and machine learning approaches, this study enhances our understanding of microbial biogeography and its role in winegrowing. Our findings highlight the need for further research on microbial transmission, vineyard management, and implications for wine quality, offering a foundation for a more precise, science-driven approach to terroir expression.

Graphical Abstract

We sampled a total of 95 vineyards, with longitudinal sampling of 12 over three years, controlling for cultivar while capturing variations in topography, soil properties, and microclimate. Data collected in vineyards further included berry chemistry, vineyard management, and phenology. We sequenced microbial communities in both vineyard and fermentation samples, and the wines were additionally analysed with three different untargeted metabolomics approaches. Supervised and unsupervised modeling revealed spatio-temporal variability, the influence of environmental factors on microbial communities, and their impact on wine characteristics.