Figure
6. The immediate effect of a disturbance is not related to the
microbiome’s long-term compositional recovery.
LEGENDS
Figure 1. Measuring microbial recovery after disturbance. Community
richness can be measured before disturbance (yellow), immediately after
disturbance (green), and over longer time periods (blue, a).
Compositional dynamics can be assessed using compositional dissimilarity
metrics (i.e., Bray-Curtis, b). Dispersion can be measured as the
pairwise dissimilarity between experimental replicates, and turnover can
be measured as the pairwise dissimilarity between pre-disturbance and
disturbed communities (red). Following disturbance, communities
dispersion can increase or decrease over time if communities become more
or less dissimilar, respectively (c). They may also return to their
pre-disturbance composition (negative turnover trend) or become
increasingly dissimilar from the pre-disturbance community (positive
turnover trend) Insets are the relationship between time (x-axis) and
turnover (dashed line, y-axis) or dispersion (solid line, y-axis).
Figure 2. Samples used in this synthesis. We selected time series that
had control samples and multiple samples after disturbance (a). A
vertical black line denotes a disturbance event in all cases; samples
taken on the day of disturbance (before or after) are shown along this
line. A vertical grey line indicates the fourth day after disturbance.
Studies which had not sampled the recovering microbiome within
<4 days after disturbance were excluded from assessments of
the immediate impacts of disturbance on richness and dispersion. All
samples were standardized to 1,500 observations per sample, and had an
average coverage >0.9 (b).
Figure 3. The effect of disturbance on microbiome richness, immediately
(<4 days) after disturbance (a), and over 50 days of recovery
(b). Richness was calculated as the number of observed taxa in each
sample, and is presented with a log2-transformed y-axis.
Circles represent samples, and are colored by study. In a, solid
black circles indicate the mean across time series per environment with
a 95% CI indicated by error bars. In b , regression lines for
each time series are colored by study, and the solid black line shows
the mean response across time series per environment. The 95% CI is
displayed as a grey shaded area, and environments for which overall
trends deviate from zero are indicated with an asterisk (*) on the
bottom right corner.
Figure 4. Consistent effect of disturbance on microbiome dispersion are
not detected, immediately (<4 days) after disturbance (a), or
over 50 days of recovery (b). Dispersion was calculated as the pairwise
Bray-Curtis distance between replicates for each time point within each
time series, and each circle is a pairwise comparison, colored by study.
In a, solid black circles indicate the mean across time series
per environment with a 95% CI. In b , regression lines for each
time series are colored by study, and a solid black line indicates the
mean response across time series per environment. The 95% CI of the
overall response in each environment is displayed as a grey shaded area.
y-axis scales differ per environment, and the overall trends from each
environment did not deviate from zero.
Figure 5. The effect of disturbance on community recovery is
environment-dependent. For each time series, recovery was calculated as
the pairwise distance between post-disturbance samples and
pre-disturbance controls. Each point is a pairwise comparison, colored
by study. recover to increasingly different conformations Regression
lines for each time series are colored by study, and a solid black line
indicates the mean response across time series per environment. The 95%
CI is displayed as a grey shaded area, and environments for which
overall trends deviate from zero are indicated with an asterisk (*) on
the bottom right corner.
Figure 6. The immediate effect of a disturbance is not related to the
microbiome’s long-term compositional recovery. Each point is a time
series, colored by its environment. Immediate richness responses were
calculated as the effect of disturbance on community richness (Figure
3a), and log-transformed. Turnover rates were calculated as the slope
estimate for turnover over time. Error bars show the 95% CI for both
metrics. Large circles indicate the mean response per environment.