Fig. 3

Dissimilarities between microbial communities. Beta diversity analyses for fungal (a, b) and bacterial (c, d) communities originating from 420 wheat residue samples. Hierarchical clustering (a, c) and multidimensional scaling (b, d) are based on the compositional distances between samples (Bray-Curtis distance matrix). a, c Visualization of compositional distances between samples through hierarchical clustering with the average linkage method. The samples (15 sampling points per treatment) corresponding to the two cropping seasons (year) are represented by the two colored horizontal series (2016–2017, 2017–2018). Effects of seasonality are highlighted by different colors, corresponding to the different sampling dates (July: green; October: red; December: blue; February: gray). The intensity of the colors distinguishes between samples obtained from plants inoculated with Z. tritici (I, dark hues) and non-inoculated samples (NI, light hues). Aboveground and soil-contact treatments are represented by horizontal lines, with each sample considered separately. b, d Visualization of compositional distances between samples through multidimensional scaling (MDS). Each data point corresponds to one sample of wheat residues. The shape of the points (circles, 2016–2017; triangles, 2017–2018) corresponds to the cropping season (year effect); the colors, similar to those used in graphs a and c, correspond to the sampling dates (seasonality effect). For fungal communities, MDS analysis was performed on all samples together, whereas for bacterial communities, the analyses of the July samples and samples from all other sampling dates (October, December, and February) were separated, in accordance with the large differences between the communities of these samples shown in the clustering analysis (c). For the sake of clarity, the MDS are shown according to the soil-contact condition