A global microbiome survey of vineyard soils highlights the microbial dimension of viticultural terroirs

Abstract

The microbial biodiversity found in different vitivinicultural regions is an important determinant of wine terroir. It should be studied and preserved, although it may, in the future, be subjected to manipulation by precision agriculture and oenology. Here, we conducted a global survey of vineyards' soil microbial communities. We analysed soil samples from 200 vineyards on four continents to establish the basis for the development of a vineyard soil microbiome's map, representing microbial biogeographical patterns on a global scale. This study describes vineyard microbial communities worldwide and establishes links between vineyard locations and microbial biodiversity on different scales: between continents, countries, and between different regions within the same country. Climate data correlates with fungal alpha diversity but not with prokaryotes alpha diversity, while spatial distance, on a global and national scale, is the main variable explaining beta-diversity in fungal and prokaryotes communities. Proteobacteria, Actinobacteria and Acidobacteria phyla, and Archaea genus Nitrososphaera dominate prokaryotic communities in soil samples while the overall fungal community is dominated by the genera Solicoccozyma, Mortierella and Alternaria. Finally, we used microbiome data to develop a predictive model based on random forest analyses to discriminate between microbial patterns and to predict the geographical source of the samples with reasonable precision.

Fig. 1. Coverage of the study.

The countries highlighted in green are represented in the study’s dataset. The two isothermal lines define the range for optimal conditions for grapevine cultivation (isothermal source: https://www.thirtyfifty.co.uk/spotlight-climate-change.asp). Each panel shows the country, the number of samples (n) and information about the average maximum temperature (Tmax) and the level of precipitation (Prec) up to 2 weeks before sample collection. Finally, H′P and H′F define the Shannon Index for the prokaryotic and fungal community respectively.

Fig. 2. Evaluation of core microbiota.

a Co-occurrence of different genera constituting the core microbiome at continent-level for the prokaryotic (left) and fungal (right) community visualised with Venn diagrams; b Heatmaps for the top 25 genera shared between continents for the prokaryotic (left) and fungal communities (right), including information on their relative abundance and prevalence.

Fig. 3. PCoA plots of beta diversity based on CLR.

Fig. 3a, b represent the national-scale samples ordination for prokaryotes (a) and fungal communities (b); Fig. 3c, d represent the national-scale samples ordination for prokaryotes (c) and fungal communities (d) on a subset of samples from Spain.

Fig. 4. Random forest results at national scale including only microbial data.

The top charts show the 20 best predictors for prokaryotic (a) and fungal (b) communities; the bottom charts show the confusion matrixes for the prokaryotic (c) and fungal communities (d).