Background The geographic and temporal distributions of bacterial and fungal populations are poorly understood within the same wine grape cultivar. In this work, we describe the microbial composition from ‘Pinot noir’ must with respect to vintage, growing region, climate, and must chemistry across the states of California and Oregon, USA. Materials and Methods We sampled ‘Pinot noir’ clone 667 clusters from 15 vineyards existing in a latitudinal gradient spanning nearly 1,200 km in California and Oregon for two vintages (2016 and 2017). Regions included five American Viticultural Areas (AVA). In order from southern California to Oregon, these AVAs were Santa Barbara, Monterey, Sonoma, Mendocino, and Willamette Valley. Uninoculated grape musts were subjected to 16S rRNA gene and ITS-1 amplicon sequencing to assess composition of microbial communities. We also measured grape maturity metrics. Finally, to describe regions by precipitation and growing degree days, we queried the Parameter-elevation Regressions on Independent Slopes Model (PRISM) spatial climate dataset. Results Most of the dominant bacterial taxa in must samples were in the family Enterobacteriaceae, notably the lactic acid bacteria or the acetic acid bacteria groups, but some, like the betaproteobacterial genus Massilia, belonged to groups not commonly found in grape musts. Fungal communities were dominated by Hanseniaspora uvarum (Saccharomycetaceae). We detected relationships between covariates (e.g., vintage, precipitation during the growing season, pH, titratable acidity, and total soluble solids) and bacterial genera Gluconobacter and Tatumella in the family Enterobacteraceae, Sphingomonas (Sphingomonodaceae), Lactobacillus (Lactobacillaceae), and Massilia (Oxalobacteraceae), as well as fungal genera in Hanseniaspora, Kazachstania, Lachancea, Torulaspora in the family Saccharomycetaceae, as well as Alternaria (Pleosporaceae), Erysiphe (Erysiphaceae), and Udeniomyces (Cystofilobasidiaceae). Fungal community distances were significantly correlated with geographic distances, but this was not observed for bacterial communities. Climate varied across regions and vintages, with growing season precipitation ranging from 11 mm to 285 mm and growing degree days ranging from 1,245 to 1,846. Discussion We determined that (1) bacterial beta diversity is structured by growing season precipitation, (2) fungal beta diversity reflects growing season precipitation and growing degree days, and (3) microbial differential abundances of specific genera vary with vintage, growing season precipitation, and fruit maturity metrics. Further, the correlation between fungal community dissimilarities and geographic distance suggests dispersal limitation and the vineyard as a source for abundant fungal taxa. Contrasting this observation, the lack of correlation between bacterial community dissimilarity and geographic distance suggests that environmental filtering is shaping these communities.

Understanding how periodical disturbances affect the community assembly processes is vital for predicting temporal dynamics in microbial communities. However, the effect of dilutions as disturbances are poorly understood. We used a marine bacterial community to investigate the effect of disturbance (+/−) and carrying capacity (high/low) over 50 days in a dispersal-limited 2 × 2 factorial study in triplicates, with a crossover in the disturbance regime between microcosms halfway in the experiment. We modelled the rate of change in community composition between replicates and used this rate to quantify selection and ecological drift. The disturbed communities increased in Bray-Curtis similarity with 0.011 ± 0.0045 (Period 1) and 0.0092 ± 0.0080 day −1 (Period 2), indicating that selection dominated community assembly. The undisturbed communities decreased in similarity at a rate of −0.015 ± 0.0038 day −1 in Period 1 and were stable in Period 2 at 0.00050 ± 0.0040 day −1 , suggesting drift structured community assembly. Interestingly, carrying capacity had minor effects on community dynamics. This study is the first to show that stochastic effects are suppressed by periodical disturbances resulting in exponential growth periods due to density-independent biomass loss and resource input. The increased contribution of selection as a response to disturbances implies that ecosystem prediction is achievable.

Understanding how periodical disturbances affect the community assembly processes is vital for predicting temporal dynamics in microbial communities. The effect of dilutions as disturbances are poorly understood. We used a marine bacterial community to investigate the effect of disturbance (+/-) and carrying capacity (high/low) over 50 days in a dispersal-limited 2x2 factorial crossover study in triplicates. The community's disturbance regime was crossed halfway. We modelled the rate of change in community composition between replicates and used this rate to quantify selection and ecological drift. The disturbed communities increased in Bray-Curtis similarity with 0.011+/-0.0045 (Period 1) and 0.0092+/-0.0080 day-1 (Period 2), indicating that selection dominated community assembly. The undisturbed communities decreased in similarity at a rate of -0.015+/-0.0038 day-1 in Period 1 and were stable in Period 2 at 0.00050+/-0.0040 day-1, suggesting drift structured community assembly. Interestingly, carrying capacity had minor effects on community dynamics. This study is the first to show that stochastic effects are suppressed by periodical disturbances resulting in exponential growth periods due to density-independent biomass loss and resource input. The increased contribution of selection as a response to disturbances implies that ecosystem prediction is achievable.