Research

Extreme rainfall affects assembly of the root-associated fungal

community
Christopher J. Barnes1,4 , Christopher J. van der Gast2, Niall P. McNamara3, Rebecca Rowe3 and
Gary D. Bending1
1
School of Life Sciences, University of Warwick, Gibbet Hill Campus, Coventry, CV4 7AL, UK; 2School of Healthcare Science, Manchester Metropolitan University, Manchester, M1 5GD,
UK; 3NERC Centre for Ecology & Hydrology, Lancaster Environment Centre, Library Avenue, Bailrigg, Lancaster, LA1 4AP, UK; 4 Present address: National History Museum of Denmar,
University of Copenhagen, 83 Sølvgade, Madison 1800, Denmark

Summary
Author for correspondence:  Global warming is resulting in increased frequency of weather extremes. Root-associated
Christopher J. Barnes fungi play important roles in terrestrial biogeochemical cycling processes, but the way in
Tel: +45 22 25 56 67
which they are affected by extreme weather is unclear. Here, we performed long-term field
Email: c.barnes@snm.ku.dk
monitoring of the root-associated fungus community of a short rotation coppice willow plan-
Received: 1 October 2017 tation, and compared community dynamics before and after a once in 100 yr rainfall event
Accepted: 3 December 2017
that occurred in the UK in 2012.
 Monitoring of the root-associated fungi was performed over a 3-yr period by metabarcod-
New Phytologist (2018) ing the fungal internal transcribed spacer (ITS) region. Repeated soil testing and continuous
doi: 10.1111/nph.14990 climatic monitoring supplemented community data, and the relative effects of environmental
and temporal variation were determined on the root-associated fungal community.
 Soil saturation and surface water were recorded throughout the early growing season of
Key words: extreme weather, mycorrhizal
fungi, root-associated fungi, soil fungi, 2012, following extreme rainfall. This was associated with a crash in the richness and relative
temporal variation in microbial communities. abundance of ectomycorrhizal fungi, with each declining by over 50%. Richness and relative
abundance of saprophytes and pathogens increased.
 We conclude that extreme rainfall events may be important yet overlooked determinants of
root-associated fungal community assembly. Given the integral role of ectomycorrhizal fungi
in biogeochemical cycles, these events may have considerable impacts upon the functioning
of terrestrial ecosystems.

Plants live in close association with distinct microbial commu-

Introduction
nities in the rhizosphere, the radial gradient spanning from the
There is mounting evidence that global warming is directly plant root into the soil (Hartmann et al., 2008) into which
increasing the frequency of weather extremes, including heavy microorganisms are selectively recruited (Berendsen et al., 2012).
rainfall events, droughts and high temperatures across Europe These root-associated communities are diverse, and although
and North America (Mallakpour & Villarini, 2015; Shepherd, most attention has focused on their bacterial and fungal
2015), which may be further increased during El Nino events communities (Smit et al., 1999; Smalla et al., 2001), they include
(Cai et al., 2014). Terrestrial ecosystems play a key role in deter- a plethora of other components including protists and nematodes
mining global climate–ecosystem feedbacks, due to their role in (Bonkowski, 2004). These communities likely play key roles in
the exchange of greenhouse gases (GHG) including CO2, terrestrial ecosystems, connecting above- and belowground diver-
methane and nitrous oxide, and storage of carbon (C) in soil and sity (Bardgett & van der Putten, 2014).
vegetation. Extreme weather may have major effects on climate– Mycorrhizal fungi, particularly ectomycorrhizal (ECM) and
terrestrial ecosystem interactions, with links to reductions of C arbuscular mycorrhizal (AM) types are some of the best-
stocks through a range of feedbacks operating on plant productiv- characterized fungal inhabitants of the rhizosphere. These fungi
ity and soil processes (Reichstein et al., 2013). However, studies form mutualistic symbioses with their host plants (Smith &
of the impacts of extreme weather events on soil systems are lim- Read, 2010), conferring a range of benefits, particularly by
ited due to their rarity and our inability to capture such events in increasing host uptake of nutrients (Garcia et al., 2014) and pro-
space and time. As a result, very little is known of the ways in viding disease resistance in exchange for C (Chakravarty &
which extreme weather, particularly extreme rainfall events, affect Unestam, 1987; Liu et al., 2007). These root-associated fungi can
soil communities and associated processes, and the consequences be major components of terrestrial ecosystems through their role
for climate–ecosystem feedbacks (Heimann & Reichstein, 2008). in driving biogeochemical cycles (H€ogberg et al., 2010). For

Ó 2018 The Authors New Phytologist (2018) 1

New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com
 New
2 Research Phytologist

example within temperate forest soils, they may contribute a Fungi can have varied responses to waterlogging within terrestrial
third of ecosystem C and nitrogen (N) mineralization (Finzi ecosystems, with ‘high moisture’ specialist AMF and saprophytic
et al., 2015). ECM dominate the root zone in temperate and species that are seemingly unaffected or even thrive in water-
boreal forest systems, supporting plant nutrition and productivity logged and anoxic conditions (Camy et al., 2003; Fougnies et al.,
while acting as major conduits by which C flows from the plant 2007; Garcıa et al., 2008; Yang et al., 2016). ECM fungal species
to the soil on a global scale (Leake et al., 2004). Nonmycorrhizal are known to have variable hydrophobicity (Lilleskov et al.,
fungi also inhabit the rhizosphere, but their diversity and func- 2011), which may be an adaptation to moisture amounts within
tional significance are typically less well understood (Kubartova habitats and across seasons (Unestam & Sun, 1995). Some ECM
et al., 2008; Dean et al., 2012). Some such fungi occur as endo- fungi inhabit environments which are permanently waterlogged
phytes within the root and can also confer growth benefits to the (Baar et al., 2002). However, increasing saturation may reduce
host plant, whilst saprophytes promote mineralization processes, ECM fungal abundance across ecosystems (Tedersoo et al.,
altering nutrient availability and indirectly influencing plant 2009), and even induce death of ECM fungal mycelium within
growth (Smith & Read, 2010; Bardgett & van der Putten, an ecosystem as soil becomes waterlogged (Coutts & Nicoll,
2014). Pathogens, meanwhile, can directly reduce plant growth 1990). However, the way in which change in weather, and in par-
and development. Even within each of these differing ‘life strate- ticular weather extremes, affects the composition and function of
gies’, fungi can also show considerable functional variability the rhizosphere fungal community remains unknown, represent-
(Kiers et al., 2011; Dean et al., 2012), and changes in the compo- ing a considerable obstacle in understanding future climate-
sition of root-associated fungal communities may have profound ecosystem feedbacks.
ecosystem-level effects on soil processes such as C, N and phos- Here, we quantify the relative effects of spatial (geographical
phorus (P) cycles within natural and agricultural systems. distance between sampling locations), environmental (soil and
Assembly of root-associated fungi has been shown to be regu- climatic properties) and temporal variation (seasonal and interan-
lated by a wide range of spatial and temporal variables. Soil pH nual) in determining the root-associated fungal community of
has a near ubiquitous effect on soil microbial community assem- willow over a 3-yr period. Soil cores were taken across line tran-
bly (Coughlan et al., 2000; Gosling et al., 2013; Tedersoo et al., sects in October 2010, July 2011, October 2011, July 2012 and
2014), whilst available P and N have been linked to changes in October 2012. Fine roots were extracted from soil cores and their
the composition of root-associated fungal communities, with associated fungal communities were profiled by high-throughput
increases of both linked to reduced mycorrhizal abundance sequencing. Significantly, sampling in 2012 coincided with
(Gosling et al., 2013). Climatic factors such as rainfall and tem- record levels of rainfall across the UK, and locally at our field site,
perature also have been indicated to drive fungal richness on a rainfall was at its highest annual levels since records began in
global level (Swaty et al., 1998), whilst increasing geographical 1947. This provided an unique opportunity to elucidate the mag-
distance between soil fungal communities has been linked with nitude to which extreme rainfall and associated soil saturation
increasing dissimilarity, independent to that of changing environ- influences assembly of communities in situ, including the
mental factors (Lilleskov et al., 2004; Barnes et al., 2016b). Root- responses of fungi with distinct nutritional modes (ectomycor-
associated fungal communities also change over time, with strong rhiza, pathogens, endophytes and saprophytes), relative to spatial
evidence for seasonality of community structure (Bohrer et al., and seasonal variation in soil and climatic factors.
2004; Dumbrell et al., 2011), and inter-annual shifts which can
take place across multiple years (Last et al., 1984; Visser, 1995;
Materials and Methods
Daniell et al., 2001; Husband et al., 2002), adding further com-
plexity to understanding the factors that regulate community
Study site and experimental design
assembly.
Given that soil biodiversity and functioning remains largely ‘a The study was performed in a short rotation coppice willow plan-
black box’, it is unsurprising that to date little is known of the tation near Lincoln, UK (53.3163°N, 0.5777°W), which was
way in which extreme weather affects the assembly of root- subject to minimal land management practices. The site had a
associated fungi, and the consequences for climate–ecosystem 30-yr mean air temperature of 9.9°C and a fine loam over clay
feedbacks (Bardgett et al., 2008). The data that are currently soil (15% clay, 49% sand and 36% silt). In 2000, the willow was
available on extreme weather–soil interactions are limited and planted at a density of 15 000 stools ha 1, covering an area of c.
derived from experimental manipulations (Barnard et al., 2013; 9.44 ha. Salix cuttings were planted in paired rows 0.75 m apart,
Amend et al., 2016); however, drought and rewetting of soil cores with 1.5 m spaces between the rows. Six closely related Salix
has revealed a high resilience of the fungal community to desicca- viminalis L. genotypes were planted in order to limit disease
tion. Soil saturation following intense rainfall can have profound spread, with Tora (60%) being the most abundant (the others
effects on plant productivity, and exposure for long periods can being Bjorn (10%), Bowles Hybrid (10%), Jorr (10%) and
even cause plant death (Rivest et al., 2013). Furthermore, soil sat- Jorunn (10%)). Willow genotypic variation was assessed along
uration leads to hypoxic or anoxic conditions, promoting micro- the line of the trees where line transects were performed, and no
bial reduction reactions, and in turn influencing the abundance pattern was observed in planting (data not shown). Given the
and composition of soil microbiota (Coutts & Nicoll, 1990; density of trees found within the field site, and that roots can
Unger et al., 2009; Wilson et al., 2011; Wagner et al., 2015). spread over 9 m from individual S. viminalis trees (Phillips et al.,

New Phytologist (2018) Ó 2018 The Authors

www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust
New
Phytologist Research 3

2014), roots were considered an homogenous mixture of multi- DNA extraction kit (MP Biomedicals, Cambridge, UK) as per
ple genotypes within the upper soil profile. Coppicing began in the manufacturer’s instructions. After extraction, the four sub-
2001, with further events in 2004, 2007 and 2010, providing an samples at each sampling location were equilibrated to 25 ng ll 1
average yield of 6.72 t ha 1. In February 2010, 660 kg ha 1 PK using a nanodrop ND1000 (Fischer Scientific, Loughborough,
fertilizer (Fibrophos, Melton, UK), 20 t ha 1 of lime and UK) and 10 ll of each was pooled to make the DNA template
20 t ha 1 of locally sourced green waste compost were added to for each of the eight sampling locations for each sampling time,
the field site; no exogenous nutrient inputs were added during with 40 samples in total used for sequencing (eight sampling
the duration of the experiment. After planting, soil remained locations, five time points).
untilled throughout the lifetime of the crop. Therefore, soil
microbial communities developed over this period without major
Weather data collection
disruption of the soil matrix, in a system with limited above-
ground diversity. Meteorological data were provided by a weather station located
Sampling was performed along single line transects (as per Bar- at the field site. These consisted of a cup anemometer, wind vane,
nes et al., 2016b) at each sampling time. Starting from the south- air and wet bulb temperatures (Didcot Instruments AWS, Did-
ernmost edge of the field heading due north, single line transects cot, UK) and a rain gauge (Rimco, Malton, UK). Data were
started 25 m into the field in order to avoid edge effects, sampling logged as 30-min averages with the exception of rainfall, which
eight locations every 20 m along the line. At each location, four was logged as the 30 min total. Temperature and rainfall data
subsamples were taken 1 m apart from north, south, east and west were measured continuously from April 2010 to December 2012.
directions from the central location, using a 4.5-cm diameter soil Measurements of volumetric soil moisture content (SWC;
auger (Van Walt Equipment, Haslemere, UK) to a depth of m3 m 3) were only recorded over the main growth period (June–
15 cm, which was wiped clean with ethanol between samples to November). SWC was recorded by two CS616 time domain
limit contamination. Transects were taken in October 2010, July reflectometer (TDR) probes (Cambell Scientific Inc., Logan,
2011, October 2011, July 2012 and October 2012, shifting a UT, USA), which measure the upper 0.3 m of the soil profile.
further 3 m north each time in order to avoid sampling of previ- SWC sensors scanned every 10 s and were logged as 30-min aver-
ously disturbed areas. ages. All meteorological data were recorded using CR10 data log-
gers (Campbell Scientific, UT, USA).
In order to test the soil saturation point, six plastic 66-mm
Soil nutrient analysis
diameter and 30-cm depth tubes were hammered into the soil to
Soil cores were homogenized evenly by gloved hand, before 100 g remove intact cores. Cores were bagged and returned to the lab,
was removed and loosely covered to air-dry, and sieved to before being cut into 0–15 cm and 15–30 cm sections. Volumet-
< 2 mm particle size. Soil pH and nutrient analysis were per- ric moisture content (VMC) was determined as per Laird et al.
formed as described previously in Barnes et al. (2016a), with pH, (2010), with cores saturated within a solution of 0.001 M CaCl2
NO3, Mg, available K and available P determined for each indi- and weighed, before being incubated in an oven until completely
vidual subsample (32 per transect) and averaged at each of the dry and reweighed. VMC was subsequently determined as a per-
eight sampling locations. This was repeated for each line transect centage of bulk density of soil.
performed at the different times. These soil properties were anal-
ysed as they have repeatedly been shown to influence root-
High-throughput amplicon sequencing and processing of
associated fungal community assembly (Geisseler & Scow,
sequencing data
2014).
DNA extracts underwent sequencing and subsequent bioinfor-
matics as described previously in Barnes et al. (2016b) using
Root-associated fungal DNA extraction
ITS1F and ITS4 primers (Gardes & Bruns, 1993; Yang et al.,
Following removal of c. 100 g (from a total of between 200 and 2012), before sequencing on a Roche 454 GS Junior pyrose-
300 g) of soil for nutrient analysis, the remaining soil from each quencer (454 Life Sciences/Roche Applied Biosystems, Nurley,
subsample was soaked in deionized water at room temperature NJ, USA) at Micropathology Ltd (Coventry, UK). Raw
for 1 h before all roots were hand extracted using forceps (be- sequences were deposited at the NCBI sequence read archive
tween 2 and 5 g of roots per subsample). Nonsenescent fine roots under the accession numbers SRP056724 for October 2010, and
(< 2 mm in diameter), which were identified by their colour, tex- SRP062101 for July 2011, October 2011, July 2012 and Octo-
ture, turgor and branching structures, were washed over a 6-mm ber 2012.
sieve to remove adherent soil. By selecting these nonsenescent Operational taxonomic units (OTUs) were picked de novo
fine roots, sampling of contaminating roots from the herbaceous from sequences using the UCLUST algorithm at a 97% similarity
understory species was also avoided. Entire pools of selected roots level, and chimera checked using the UCHIME algorithm in de
were cut into 1-cm lengths, mixed, and a 0.5 g randomly selected novo mode (Edgar, 2010; Edgar et al., 2011). Taxonomic assign-
subsample was used for DNA extraction. Roots were initially ments were initially made using the UNITE database
exposed to mechanical lysis of two periods of 30 s at 30 hz using (27.08.2013 release from the UNITE project) using BLAST (with
a TissueLyser (Qiagen) before extraction using the PowerSoil an e-value > 0.001), before OTUs of >0.1% relative abundance

Ó 2018 The Authors New Phytologist (2018)

New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com
 New
4 Research Phytologist

underwent a further BLAST search against the continuously taxa), hydrophobicity and extramatrical exploration type in order
updated UNITE database and taxonomy reassigned if to analyse variation between sampling locations (df = 39) and
e-value > 0.001. OTUs with ≤3 reads assigned were also removed sampling times (df = 4), and post-hoc (Tukey’s honestly signifi-
to limit the effects of sequencing errors and artefacts (Huse et al., cant difference) tests were performed when significant differences
2007). were observed. Due to the high number of analyses performed on
After processing, 76.6% (92 595) of total reads remained, individual OTU data, P-values from individual OTU analyses
averaging 2341 reads per sample, across 931 OTUs (mean were adjusted using Benjami and Hochberg’s false discovery rate
length = 496.1, standard deviation = 22.5). Chao1 estimates were procedure in order to limit false-positives, using the p.adjust
calculated to estimate the theoretical maximum number of OTUs function within the STATS package of R (Benjamini & Hochberg,
per sampling location (Chao, 1987), revealing that 63.8% of 1995). Finally, general linear modelling (GLM) was used to
OTUs were detected after rarefaction and many rare OTUs were ascertain which temporal (time as weeks after sampling) and envi-
undetected. Given the issues McMurdie & Holmes, 2014 raise in ronmental parameters (soil pH, P, K, Mg, NO3 and SWC) sig-
performing community analyses with variable library sizes, the nificantly explained variance in the relative abundances of
rarefied OTU table underwent normalization using the DESEQ2 individual OTUs that were found to significantly change between
package as part of QIIME (Anders & Huber, 2010). Negative val- time points using repeated measure ANOVAs (Grafen and Hails,
ues were considered zeros before data were finally converted to 2002). GLM was performed in XLSTAT (Addinsoft, Paris, France)
relative abundances (as percentage of normalized read numbers) with the aim of disentangling the effects of sampling time from
as per Fernandez et al. (2017). The negative binomial model per- that of SWC.
formed within DESEQ2 can simultaneously account for variation Hellinger transformations downweight rare species and per-
in library sizes and biological variability better than simply sub- form well with high-throughput sequencing datasets in which
sampling to a uniform sequencing depth alone. there are many zero values, and thus were performed on the com-
Some members of the Ascomycota, Basidiomycota and munity data before Bray–Curtis similarity matrices were gener-
Zygomycota can form ECM associations or exist as root endo- ated (Bray & Curtis, 1957). Nonmetric multidimensional scaling
phytes, whereas others act as saprophytes or pathogens. There- analysis (nMDS) was performed from this matrix to visualize
fore, a likely life strategy (i.e. nutritional mode) was determined trends in community similarity. Variation within the overall
for OTUs in a two-step process. Initially an automated approach community and each life strategy group of the root-associated
was taken through parsing the OTU table through the fungi separately was quantified via PERMANOVA (using the
FUNGuild database (v.1.0) (Tedersoo et al., 2014; Nguyen et al., ADONIS function within R) against temporal (sampling year and
2016), before a secondary round of assignments were made on a season), environmental properties (pH, NO3, Mg, K and P) and
manual basis using previous literature findings (Supporting Infor- geographical distance (between sampling locations) using sam-
mation Table S1). These were performed on OTUs with family pling time as the strata (as October 2010, July 2011, October
level or greater specificity of taxonomic assignments and consen- 2011, July 2012 or October 2012). As the order of explanatory
sus of life strategy within the taxonomic unit. Groups considered variables affects the outcome of PERMANOVA analyses, vari-
were ECM fungi, endophytes, pathogens and saprotrophs, whilst ables were placed in order of the largest proportion of variation
other fungi that were very low in OTU richness such as lichen explained before running the final analysis. The nMDS analyses
symbionts and yeasts, or those that were not taxonomically and PERMANOVA were performed using the VEGAN package of
assigned at genus level, were not included within the life strategy R (v.2.4-2; Oksanen et al., 2007), whilst nMDS and volcano
analyses. plots were created using the package GGPLOTS2 (v.2.2.1; Wick-
In order to further understand the ECM fungal response to ham, 2016).
extreme rainfall, ECM fungal OTUs also were divided by explo-
ration type (Agerer, 2001) and by their mycelial hydrophobicity
Results
(Lilleskov et al., 2011), with extramatrical exploration type
divided into contact-short, contact-medium and medium-long
Edaphic properties
groupings as performed by Fernandez et al. (2017).
Biogeochemical values were calculated as the mean of the four
subsamples at each sampling location, before undergoing two-
Statistical analyses
way repeated measures ANOVA that used distance across transect
Repeated measures one-way ANOVAs were performed on cli- as a factor, and sampling time (as October 2010, July 2011,
matic data (soil moisture, precipitation and temperature) to test October 2011, July 2012 and October 2012) as a repeated mea-
for significant differences between years, whilst two-way sure. There were strong significant gradients across the transects
ANOVAs with sampling time (repeated measure) and distance for pH (df = 7, F = 14.45, P < 0.001), available P (df = 7,
across transects were performed for assessing significant spatio- F = 3.51, P < 0.001), available K (df = 7, F = 8.59, P < 0.001)
temporal variation in soil properties (Oksanen et al., 2007). and Mg (df = 7, F = 2.74, P = 0.024), all of which were signifi-
Repeated measures one-way ANOVAs were performed on the cantly greater in sampling locations 1–3 relative to sampling loca-
OTU data for phylum, family, genus, species, life strategy tions 5–8 at all time points (data not shown). Between sampling
(known ectomycorrhizal, pathogen, saprophyte and endophyte times, pH (df = 4, F = 21.61, P < 0.001) was significantly raised

New Phytologist (2018) Ó 2018 The Authors

www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust
New
Phytologist Research 5

in 2011 and 2012, available P decreased significantly after 2010

(df = 4, F = 16.07, P < 0.001), whilst NO3 showed significant
seasonal variation (df = 4, F = 13.95, P < 0.001), with substan-
tially higher quantities in soil in July, relative to October tran-
sects. Available K also varied over time, peaking in October 2010
and July 2012 time points (df = 4, F = 4.14, P = 0.030)
(Fig. S1a–e).

Climatic data
The average daily temperature at the field site varied throughout
the years (df = 10, F = 14.1, P < 0.001), from c. –1°C in Decem-
ber 2010 to 17°C in July 2012 (Fig. S2a); however, temperature
did not vary between years (df = 2, F = 0.325, P = 0.725) despite
2011 having a warmer spring and autumn compared to the other
years. Precipitation did not consistently differ between months
(Fig. S2b), yet it differed strongly between years (df = 2, F = 3.86,
P = 0.036), with 2012 being substantially wetter in April
(140 mm per month) and June (130 mm per month), which was
over double that of the previous year’s averages. SWC was only
monitored between June and November, but reflected rainfall
data, with 2012 significantly higher than 2010 and 2011 in all
months (df = 2, F = 5.47, P = 0.020), but with no apparent con-
sistent monthly differences between years (df = 4, F = 0.669,
P = 0.628; Fig. S2c). The average monthly SWC across the entire
monitoring period was 0.313 m3 m 3; however, for 2012 this fig-
ure was 0.406 m3 m 3 and reached as high as 0.484 m3 m 3 in
both June and July 2012. Saturation tests within the laboratory
suggested that complete saturation can occur from
0.510 m3 m 3, and thus in June and July 2012 soil was 94.9% of
minimum saturation (Fig. S3). Furthermore, observations in the Fig. 1 Nonmetric dimensional scaling showing clustering based on
similarity of the root-associated fungal communities between sampling
field showed a considerable presence of standing surface water
time points.
throughout this period (Fig. S4).
significant effect of seasonality (i.e. sampling in either July or
October) on community structure over the 3-yr sampling period.
Investigating spatial, temporal and environmental factors
Neither geographical distance between sampling locations nor
regulating root-associated fungal assemblage
the remaining soil properties affected community assembly.
In order to understand the spatio-temporal variation of the root- Given the community shift that occurred between 2011 and
associated fungal community composition over the sampling 2012, community data were partitioned into pre- (October 2010,
period, an initial Bray–Curtis similarity matrix was created from
the sequencing data and visualized using multidimensional scal- Table 1 The relative importance of time, geographical distance (between
ing (limited to two dimensions, stress = 0.145, 999 iterations). samples) and soil properties for the root-associated fungal community of
Samples taken in 2010 and 2011 grouped together, and were dis- willow as revealed by PERMANOVA
tinct from those taken in 2012 (Fig. 1). There was no evidence
Degrees of
for seasonal changes in community composition for either the Parameter freedom F-value R2 P-value
2011 or 2012 time points, as there was no grouping of October
or July time points within the nMDS ordinations. Season 1 1.338 0.026 0.155
PERMANOVA was subsequently performed against temporal Year 3 3.640 0.211 0.001
pH 1 5.586 0.108 0.001
data (sampling year and season as factors), environmental proper- Distance 1 0.772 0.015 0.736
ties (pH, NO3, Mg, K and P) and geographical distance (as loca- Nitrate (NO3) 1 0.859 0.017 0.612
tion on transect), whilst sampling time point was used as the Magnesium (Mg) 1 0.759 0.015 0.771
strata (Table 1). In this analysis, sampling year (2010, 2011 or Potassium (K) 1 1.045 0.020 0.350
2012) had the greatest effect on the community composition, Phosphorus (P) 1 1.356 0.026 0.153
Residuals 29 0.562
explaining 21.1% of variation, whilst pH explained a further Total 39 1.000
10.8% of variation (Fig. S5), with a total of 31.9% total commu-
nity variation explained by measured parameters. There was no Bold indicates significance (P < 0.05).

Ó 2018 The Authors New Phytologist (2018)

New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com
 New
6 Research Phytologist

July 2011 and October 2011) and post-extreme rainfall (July between time points (F = 5.04, P = 0.004), from 22.3% in the
2012 and October 2012), and reanalysed for seasonal, environ- pre-extreme rainfall samples to 33.7% of the post-extreme rainfall
mental and edaphic effects that may have been masked by the samples. The Zygomycota were the least abundant and diverse
weather event (Fig. S6; Table S2). The two separate phylum, but OTU richness increased (F = 16.45, P < 0.001) from
PERMANOVAs revealed that the pre-extreme rainfall communi- an average of 1.4 per sampling location in the pre-extreme to 4.2
ties were regulated only by pH (21.6%), whilst seasonality (as per sampling location in the post-extreme rainfall samples. Fur-
July or October) and geographical distance (as location on the thermore, relative abundance of Zygomycota reads increased sig-
line transect) had no effect. The post-extreme rainfall community nificantly (F = 13.06, P < 0.001) from just 0.4% in the pre-
remained entirely unexplained by all measured parameters. extreme to 6.2% in the post-extreme rainfall samples.
There were a number of OTU that could only be assigned to
the fungal kingdom, or had no taxonomic information at all.
Determining the taxonomic shifts associated with the
They accounted for between 9.3% and 21.2% of reads (F = 5.48,
change in community composition in 2012
P = 0.002, respectively) and were lowest in July 2012 and highest
Total OTU richness varied from 80 to 299 OTUs between sam- in October 2012. These averaged 30.8 OTUs per sample and
ples, and there was no significant difference between sampling remained stable over time in OTU richness (F = 2.68,
times (F = 0.98, P = 0.431). However, the Basidiomycota, which P = 0.052). There also was an additional 3.3% average abun-
represented the largest phylum by OTU richness, differed signifi- dance of reads that could not be assigned even at the kingdom
cantly between sampling times (F = 4.24, P = 0.008), with level (varying between 2.5% and 3.8% between sampling times),
between 116.4 and 98.6 OTUs per sampling location recorded accounting for an average of 9.6 OTUs per sample (ranging from
across the 2010 and 2011 communities (known as the pre- 7.5 and 12.1) which were retained in measures of richness and
extreme rainfall samples hereafter) and only 68.0 and 69.2 in the compositional analyses.
2012 communities (known as the post-extreme rainfall samples
hereafter; Fig. 2). The relative abundance of Basidiomycota
Determining the functional shifts associated with the
declined significantly from 65.4% in the pre-extreme to 37.7%
change in community composition in 2012
in the post-extreme rainfall samples (F = 12.27, P < 0.001).
The Ascomycota were the second most OTU-rich and abun- Relative abundance of ECM fungi remained stable throughout
dant phylum. Whilst OTU richness ranged from 34.6 to 62.0, the pre-extreme rainfall samples, varying from 52.0% to 56.1%.
there was no significant change over time (F = 2.46, P = 0.068). However, there was a significant decline to 23.5% in the samples
However, the Ascomycota varied significantly in abundance after the extreme rainfall event (F = 7.71, P < 0.001) (Fig. 3a).

(a)

(b)
Fig. 2 Average (a) relative abundance and
(b) operational taxonomic unit (OTU)
richness of the willow root-associated fungal
community separated at the phylum level
over a 3-yr sampling period. October 2010,
July 2011 and October 2011 were pre-
extreme rainfall at the site, whilst July 2012
and October 2012 were post-extreme
rainfall. Error bars represent  1 SD of the
mean. Capital letters in parentheses
represent the taxonomic groups that differ
between sampling times and different lower
case letters indicate the associated significant
differences, by Tukey’s test (a = 0.05).

New Phytologist (2018) Ó 2018 The Authors

www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust
New
Phytologist Research 7

(a)

(b)

Fig. 3 Average (a) relative abundance and

(b) operational taxonomic unit (OTU)
richness of differing life strategies of root-
associated fungi extracted from willow roots
over a 3-yr sampling period. October 2010,
July 2011 and October 2011 were pre-
extreme rainfall at the site, whilst July 2012
and October 2012 were post-extreme
rainfall. Error bars represent  1 SD of the
mean. Capital letters in parentheses
represent the taxonomic groups that differ
between sampling times and different lower
case letters indicate the associated significant
differences, by Tukey’s test (a = 0.05).

There was also a significant decline in ECM fungal OTU rich- total relative abundance, whilst OTU richness dropped from
ness in the post-extreme rainfall samples relative to pre-extreme 65.5 to 27.5.
rainfall samples (F = 20.14, P < 0.001), falling from 79–95 to ECM fungi were divided by hydrophobicity and exploration
43–45 OTUs (Fig. 3b). type, and dynamics analysed over time. The hydrophobic ECM
Pathogens were the second most abundant and OTU-rich life fungi were present in considerably greater relative abundance and
strategy, accounting for an average of 3.7% of relative abundance. OTU richness compared to hydrophilic types (Fig 4). Hydropho-
Pathogen relative abundance differed over time (F = 6.93, bic ECM fungi declined significantly in both relative abundance
P < 0.001), from 3.1% pre-extreme to 4.7% post-extreme rain- (F = 6.49, P < 0.001) and OTU richness (F = 6.20, P < 0.001)
fall. The number of pathogen OTUs within the community var- over time, from averages of 15.7% and 39.1 OTUs pre-extreme
ied significantly over time (F = 3.38, P = 0.022), with 5.3 OTUs rainfall to just 4.1% and 13.4 OTUs, respectively, post-extreme
pre-extreme and 6.6 OTUs post-extreme rainfall. rainfall. The relative abundance and OTU richness of
Saprophytes accounted for a further 3.2% of relative abun- hydrophilic ECM fungi did not change significantly over time
dance and also varied significantly over time (F = 2.94, (F = 1.67, P = 0.180 and F = 1.74, P = 0.163, respectively).
P = 0.038), ranging from 2.7% and 5.7% in the pre-extreme and There also were significant differences in the contact type of
post-extreme rainfall communities, respectively. Saprophyte ECM fungi over the sampling period (Fig. S7). The contact-
OTU richness increased significantly (F = 5.84, P = 0.013) from medium type ECM fungi were the most abundant, and
5.0% pre-extreme rainfall to 7.1% afterwards. declined significantly in relative abundance (F = 3.26,
Endophytes were consistently found in low abundance and P = 0.023) but not in OTU richness (F = 6.49, P = 0.131),
low OTU richness. Endophyte OTU richness varied significantly from averages of 31.6% and 64.1 OTUs pre-extreme rainfall
between sampling points, ranging from an average of 2.5 OTUs to 8.4% and 25.5 OTUs, respectively, post-extreme rainfall.
pre-extreme rainfall to 1.9 OTUs post-extreme rainfall (F = 3.08, The contact-short ECM fungi also declined in relative abun-
P = 0.032), but did not vary in relative abundance, with an aver- dance (F = 9.35, P < 0.001), from an average of 5.1% pre-
age of just 0.56% of total abundance (F = 1.59, P = 0.205). extreme rainfall to 2.8% post-extreme rainfall, whilst OTU
Closer inspection of ECM fungal families also revealed decli- richness also declined significantly over this period (F = 6.27,
nes in both OTU richness and abundance in nearly every family. P < 0.001), from averages of 6.4 to 4.6 OTU. The long-range
The overall trend of decline in ECM fungal abundance post- ECM fungi were present in too low abundance and OTU
extreme rainfall was driven primarily by changes in the Cortinari- richness to be meaningfully analysed, with an average richness
aceae (Table S3) whose abundance fell from 32.3% to 9.8% of of just 0.1 OTU and relative abundance of 0.1%.

Ó 2018 The Authors New Phytologist (2018)

New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com
 New
8 Research Phytologist

(a)

(b) Fig. 4 Average (a) relative abundances and

(b) operational taxonomic unit (OTU)
richness of the ectomycorrhizal (ECM) fungi
according to hydrophobicity. ECM fungi
were extracted from willow roots over a 3-yr
sampling period. October 2010, July 2011
and October 2011 were pre-extreme rainfall
at the site, whilst July 2012 and October
2012 were post-extreme rainfall. Error bars
represent  1 SD of the mean. Capital letters
in parentheses represent the taxonomic
groups that differ between sampling times
and different lower case letters indicate the
associated significant differences, by Tukey’s
test (a = 0.05).

42 OTUs, and accounted for the largest percentage of variation

Investigating the key fungal OTUs within the pre- and
of these OTUs, 20.3% on average. However, SWC correlated
post-extreme rainfall root-associated communities
with more of these 42 OTUs than time, with a total of 33 OTUs
We further analysed shifts after the extreme rainfall event by per- correlating with SWC and an average of 16.7% of variation.
forming repeated measure ANOVAs for individual OTUs over Whilst every OTU correlated with either time or SWC, or both,
the sampling period and visualized this variation using volcano soil pH, P, K and Mg correlated with relatively few OTU abun-
plots (Fig. S8; Newbold et al., 2017). In total, 42 OTUs were dances, with just two, six, five, four and nine OTUs, respectively.
shown to vary significantly between time points (q < 0.001), with
41 increasing in abundance and only one decreasing in the post-
Discussion
extreme rainfall samples (Table S4). Despite the overall signifi-
cant decline in ECM fungal relative abundance and OTU rich- This work demonstrates that a root-associated fungal community
ness, no single ECM fungal OTU declined significantly post- structure was relatively consistent between October 2010 and
extreme rainfall at the q < 0.001 level, although nearly all showed October 2011, before a dramatic taxonomic and functional tran-
a numerical decline in 2012. Meanwhile four pathogens (two sition took place between October 2011 and July 2012, which
Truncatella angustata OTUs, a Plectosphaerella OTU and a coincided with an extreme rainfall event that left soils saturated
Pilidium OTU) and five saprophytic OTU (four Mortierellaceae and likely under anoxic soil conditions. As part of this transition
OTUs and a Sporormiaceae OTU) increased significantly in in the root-associated fungal communities, there was a substantial
abundance. A further four soil yeasts (All Cryptococcus OTUs) decline in the relative abundance and richness of the ectomycor-
and two lichenized fungi (Venturiaceae OTU and Verrucaria rhizal (ECM) fungal community, and a concomitant increase in
andesiatica OTU) also increased in relative abundance, in addi- the relative abundance and richness of pathogens and sapro-
tion to 26 OTUs with an unknown life strategy. The only OTU phytes. These changes persisted until at least October 2012.
to significantly decrease in relative abundance within the post- The driver for the community transition between 2011 and
extreme rainfall samples could not be assigned to a life strategy. 2012 was not directly elucidated. A strong gradient in soil prop-
Finally, GLM was performed on the 42 OTUs that varied sig- erties existed across transects, which affected composition of the
nificantly between sampling times, with individual OTU abun- root-associated fungi (pH) and may have overshadowed any sea-
dances analysed against soil pH, P, K, Mg, NO3, SWC and time sonal changes in composition, which have previously been
(weeks) (Table S4). As expected, time correlated with 23 of these observed in studies of root-associated fungi (Bohrer et al., 2004;

New Phytologist (2018) Ó 2018 The Authors

www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust
New
Phytologist Research 9

Dumbrell et al., 2011), and which were not detected here. Previ- 2012 was associated with losses in richness and relative abun-
ous studies have demonstrated extensively that soil nutrients dance of hydrophobic but not hydrophilic ECM fungi, with
heavily regulate the composition of root-associated fungal com- losses from both the contact-short and contact-medium explo-
munities (Klamer et al., 2002; van der Gast et al., 2011). The ration types, providing further support that the transitions in
field site underwent coppicing in the spring of 2010, alongside community structure which we detected could be the result of
fertilization, compost application and liming, which was the most soil water saturation in 2012.
likely cause of the reduction in phosphorus (P) availability Reduced ECM fungal abundance under flooding could have
between 2010 and the later time points. However, major soil negative implications for plant vitality for a number of reasons:
physico-chemical properties did not differ significantly in 2012 (1) Flooding has been linked to increased nutrient availability to
compared to 2010 and 2011, and although a delayed response to plants (Mendoza et al., 2005), and under these conditions, the
the intense management practices of spring 2010, 2 yr earlier, ECM fungal continuum suggests that ECM fungi could become
cannot be ruled out, there were no farm management practices slightly parasitic to their hosts (Karst et al., 2008). (2) ECM fungi
imposed during the sampling period, such as crop harvest, fungi- show functional diversity, particularly in soil resource uptake and
cide application or tillage that could have directly or indirectly contributions to host nutrition, and thus major community tran-
altered root-associated fungal biota in such a sudden manner. sitions effecting a wide phylogenetic range of ECM fungi, such as
However, importantly, climate records show that although that observed here, could impact upon host nutrient uptake
temperature was comparable between 2010, 2011 and 2012, (Leake et al., 2004). (3) Increased diversity and relative abun-
there were major differences in rainfall between the 2010/2011 dance of root pathogens were associated with the rainfall event.
and 2012 growing seasons. Heavy rainfall in early 2012 resulted Mycorrhizal fungi may play a vital role in protecting plants from
in peak soil moisture content (SWC) field measurements attack by pathogens, and increased colonization by pathogens
throughout June and July, and a considerable amount of surface could be a consequence of reductions in the diversity and compo-
water was present at the field site throughout this period sition of ECM fungi (Nagy & Fossdal, 2013).
(Fig. S4). Nationally, 2012 was the wettest summer, and second Changes in the abundance and composition of ECM fungi of
wettest year in the UK since records began in 1910 (Parry et al., the magnitude observed in the post-extreme rainfall samples
2013), whilst locally it was the highest rainfall year since records could represent shifts in the pathway by which carbon (C) passes
began in 1947 (Waddington monitoring station, MetOffice, UK; from the plant to the soil, given that ECM fungi are large sinks
c. 14 km away from study site). When general linear modelling of C within forest ecosystems (Durall et al., 1994; H€ogberg et al.,
(GLM) were used to partition variation of the temporally variable 2010). Additionally, the increased diversity and relative abun-
operational taxonomic units (OTUs), SWC was the best predic- dance of saprophytes could also indicate the increased impor-
tor of OTU abundance, correlating with more OTU than time tance of assimilate reaching soil decomposer pathways following
and all other environmental variables. The taxonomic and func- the rainfall event, through increased decomposition of root hairs,
tional transition in the root fungal community in 2012 was there- roots or other root-associated fungi in response to flooding
fore associated with an extreme weather event, and the extreme (Sauter, 2013; Jaiphong et al., 2016).
rainfall clearly resulted in the prolonged soil saturation we It should be noted that the shift in the ECM fungal commu-
observed during the early summer of 2012. nity persisted for the duration of the growing season of 2012.
Soil saturation can result in hypoxic or anoxic soil conditions Clearly the timescales and extent of community recovery will be
which can result in changes in soil microbial communities and of fundamental importance in determining the ecosystem-level
death of biomass, dependent on the extent and longevity of O2 implications and legacy of the extreme weather. This study was
depletion (Wagner et al., 2015). Relative to the pre-extreme rain- conducted within an homogenous willow plantation. There is
fall samples, in the post-extreme rainfall samples ECM fungal evidence that ECM fungal richness increases with plant richness
OTU richness halved and relative abundance declined by nearly (Johnson et al., 2005; Tedersoo et al., 2014), and the extent to
two-thirds, driven primarily by basidiomycete ECM fungi, and which the resistance and resilience responses of willow-associated
the Cortinariaceae in particular. Laboratory studies have shown fungal communities reflects that of more diverse natural ecosys-
that ECM fungus hyphal growth and ECM formation are tems is unclear.
reduced at high soil moisture contents (Theodorou, 1978; Lodge, Extensive research into the biotic and abiotic factors regulating
1989; Boucher & Malajczuk, 1990) and within experimental root-associated fungi has been performed (Tedersoo et al., 2014),
plots, extra-matrical hyphae of ECM fungi have been shown to yet the weather-induced shift in taxonomic and functional com-
die under waterlogged conditions (Coutts & Nicoll, 1990). The position has, to the best of the authors’ knowledge, not been
mycelia produced by fungal species, particularly ECM fungi, observed before within environmental samples. Studies of the
range from hydrophobic to hydrophilic. This may be an adapta- effects of disturbances on soil fungal community composition
tion to widely contrasting soil moisture availability between habi- generally have been limited to persistent changes, such as fertiliza-
tats and across seasons, with hydrophobicity an adaptation to tion (Brearley et al., 2007; Gosling et al., 2013) and pollution
retain water in dry soil and prevent inundation in wet soil. Fun- (van der Gast et al., 2011; Thion et al., 2012). However, precipi-
gal species with hydrophilic mycelium may be better able to tation has been linked to changes in AM fungal community com-
resist, or show preference for, water saturated soil (Unestam & position (Hazard et al., 2013) and sporulation (Lovelock et al.,
Sun, 1995). Importantly, we found that the extreme rainfall in 2003), and ECM fungal richness (Tedersoo et al., 2012), in the

Ó 2018 The Authors New Phytologist (2018)

New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com
 New
10 Research Phytologist

field. Studies of the effects of manipulation of soil moisture con- Anders S, Huber W. 2010. Differential expression analysis for sequence count
tent on fungal communities are limited to just a handful of stud- data. Genome Biology 11: R106.
Baar J, Bastiaans T, van de Coevering M, Roelofs J. 2002. Ectomycorrhizal root
ies. Whilst soil fungal communities had considerable resilience to development in wet Alder carr forests in response to desiccation and
droughts (Evans & Wallenstein, 2012; Barnard et al., 2013; eutrophication. Mycorrhiza 12: 147–151.
Amend et al., 2016), increasing soil water content has had mixed Bardgett RD, Freeman C, Ostle NJ. 2008. Microbial contributions to climate
results, reducing fungal diversity in a study by Hawkes et al. change through carbon cycle feedbacks. ISME Journal 2: 805–814.
(2011) but having a nominal effect in a study by Fry et al. (2016). Bardgett RD, van der Putten WH. 2014. Belowground biodiversity and
ecosystem functioning. Nature 515: 505–511.
These results together with those presented in our study indicate Barnard RL, Osborne CA, Firestone MK. 2013. Responses of soil bacterial and
that ECM fungal communities are sensitive to precipitation, with fungal communities to extreme desiccation and rewetting. ISME Journal 7:
pronounced effects at very high soil water contents. 2229–2241.
Extreme weather may therefore be an often overlooked regula- Barnes CJ, Burns CA, van der Gast CJ, McNamara NP, Bending GD. 2016a.
tor in microbial community development over annual time Spatio-temporal variation of core and satellite arbuscular mycorrhizal fungus
communities in Miscanthus giganteus. Microbial Symbioses 7: 1278.
frames, potentially overshadowing the more frequently observed Barnes CJ, van der Gast CJ, Burns CA, McNamara NP, Bending GD. 2016b.
seasonal community variations (Bell et al., 2009; Dumbrell et al., Temporally variable geographical distance effects contribute to the assembly of
2011; Gilbert et al., 2012). With projections suggesting increases root-associated fungal communities. Microbial Symbioses 7: 195.
in the frequency of extreme weather patterns associated with cli- Bell CW, Acosta-Martinez V, McIntyre NE, Cox S, Tissue DT, Zak JC. 2009.
mate change (Hansen et al., 2012), and extreme weather already Linking microbial community structure and function to seasonal differences in
soil moisture and temperature in a chihuahuan desert grassland. Microbial
known to impact C cycling, leading to losses in terrestrial C Ecology 58: 827–842.
stocks (Reichstein et al., 2013), there is a pressing need to under- Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate: a practical
stand the impacts of extreme weather events on soil fungal pro- and powerful approach to multiple testing. Journal of the Royal Statistical Society
cesses and their environmental significance. The decreasing costs B (Methodological) 57: 289–300.
of performing large-scale metagenomic and metatranscriptomic Berendsen RL, Pieterse CMJ, Bakker PAHM. 2012. The rhizosphere
microbiome and plant health. Trends in Plant Science 17: 478–486.
studies should allow the comprehensive long-term environmental Bohrer KE, Friese CF, Amon JP. 2004. Seasonal dynamics of arbuscular
monitoring studies urgently required to understand climate- mycorrhizal fungi in differing wetland habitats. Mycorrhiza 14: 329–337.
rhizosphere feedbacks to be performed. Bonkowski M. 2004. Protozoa and plant growth: the microbial loop in soil
revisited. New Phytologist 162: 617–631.
Boucher NL, Malajczuk N. 1990. Effects of high soil moisture on formation of
Acknowledgements ectomycorrhizas and growth of karri (Eucalyptus diversicolor) seedlings
inoculated with Descolea maculata, Pisolithus tinctorius and Laccaria laccata.
This project was part of the Carbo-Biocrop project funded by the New Phytologist 114: 87–91.
Natural Environment Research Council (NERC). We acknowl- Bray JR, Curtis JT. 1957. An ordination of the upland forest communities of
edge NERC Centre for Ecology and Hydrology National Capa- southern Wisconsin. Ecological Monographs 27: 325–349.
bility funding through project NEC03487, Ross Morrison for Brearley FQ, Scholes JD, Press MC, Palfner G. 2007. How does light and
phosphorus fertilisation affect the growth and ectomycorrhizal community of
providing meterological data and the landowner, Jonathan two contrasting dipterocarp species? Plant Ecology 192: 237–249.
Wright for access to the commercial plantation. Cai W, Borlace S, Lengaigne M, van Rensch P, Collins M, Vecchi G,
Timmermann A, Santoso A, McPhaden MJ, Wu L et al. 2014. Increasing
frequency of extreme El Nino events due to greenhouse warming. Nature
Author contributions Climate Change 4: 111–116.
Camy C, Dreyer E, Delatour C, Marcßais B. 2003. Responses of the root rot
G.D.B., C.J.v.d.G. and N.P.M. conceived the study; G.B., C.J.B., fungus Collybia fusipes to soil waterlogging and oxygen availability. Mycological
N.P.M. and C.J.v.d.G. designed the sampling strategy; C.J.B. per- Research 107: 1103–1109.
formed the sampling and analysis; C.J.v.d.G. advised on spatial sta- Chakravarty P, Unestam T. 1987. Differential influence of ectomycorrhizae on
tistical methods; R.R. and N.P.M. performed soil saturation tests plant growth and disease resistance in Pinus sylvestris seedlings. Journal of
and provided assistance in the field; C.J.B and G.D.B. wrote the Phytopathology 120: 104–120.
Chao A. 1987. Estimating the population size for capture-recapture data with
manuscript; and all authors contributed to revisions. unequal catchability. Biometrics 43: 783–791.
Coughlan AP, Dalpe Y, Lapointe L, Piche Y. 2000. Soil pH-induced changes in
root colonization, diversity, and reproduction of symbiotic arbuscular
ORCID mycorrhizal fungi from healthy and declining maple forests. Canadian Journal
Christopher J. Barnes X http://orcid.org/0000-0001-6800-4233 of Forest Research 30: 1543–1554.
Coutts MP, Nicoll BC. 1990. Waterlogging tolerance of roots of Sitka spruce
clones and of strands from Thelephoraterrestris mycorrhizas. Canadian Journal
of Forest Research 20: 1894–1899.
Daniell TJ, Husband R, Fitter AH, Young JPW. 2001. Molecular diversity of
References arbuscular mycorrhizal fungi colonising arable crops. FEMS Microbiology
Agerer R. 2001. Exploration types of ectomycorrhizae. Mycorrhiza 11: 107–114. Ecology 36: 203–209.
Amend AS, Martiny AC, Allison SD, Berlemont R, Goulden ML, Lu Y, Dean R, Van Kan JaL, Pretorius ZA, Hammond-Kosack KE, Di Pietro A,
Treseder KK, Weihe C, Martiny JBH. 2016. Microbial response to simulated Spanu PD, Rudd JJ, Dickman M, Kahmann R, Ellis J et al. 2012. The Top
global change is phylogenetically conserved and linked with functional 10 fungal pathogens in molecular plant pathology. Molecular Plant Pathology
potential. ISME Journal 10: 109–118. 13: 414–430.

New Phytologist (2018) Ó 2018 The Authors

www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust
New
Phytologist Research 11

Dumbrell AJ, Ashton PD, Aziz N, Feng G, Nelson M, Dytham C, Fitter AH, Heimann M, Reichstein M. 2008. Terrestrial ecosystem carbon dynamics and
Helgason T. 2011. Distinct seasonal assemblages of arbuscular mycorrhizal climate feedbacks. Nature 451: 289–292.
fungi revealed by massively parallel pyrosequencing. New Phytologist 190: H€ogberg MN, Briones MJI, Keel SG, Metcalfe DB, Campbell C, Midwood AJ,
794–804. Thornton B, Hurry V, Linder S, N€a sholm T et al. 2010. Quantification of
Durall DM, Jones MD, Tinker PB. 1994. Allocation of 14C-carbon in effects of season and nitrogen supply on tree below-ground carbon transfer to
ectomycorrhizal willow. New Phytologist 128: 109–114. ectomycorrhizal fungi and other soil organisms in a boreal pine forest. New
Edgar RC. 2010. Search and clustering orders of magnitude faster than BLAST. Phytologist 187: 485–493.
Bioinformatics 26: 2460–2461. Husband R, Herre EA, Turner SL, Gallery R, Young JPW. 2002.
Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. 2011. UCHIME Molecular diversity of arbuscular mycorrhizal fungi and patterns of host
improves sensitivity and speed of chimera detection. Bioinformatics 27: 2194– association over time and space in a tropical forest. Molecular Ecology 11:
2200. 2669–2678.
Evans SE, Wallenstein MD. 2012. Soil microbial community response to drying Huse SM, Huber JA, Morrison HG, Sogin ML, Welch DM. 2007. Accuracy and
and rewetting stress: does historical precipitation regime matter? quality of massively parallel DNA pyrosequencing. Genome Biology 8: R143.
Biogeochemistry 109: 101–116. Jaiphong T, Tominaga J, Watanabe K, Nakabaru M, Takaragawa H, Suwa R,
Fernandez CW, Nguyen NH, Stefanski A, Han Y, Hobbie SE, Montgomery Ueno M, Kawamitsu Y. 2016. Effects of duration and combination of drought
RA, Reich PB, Kennedy PG. 2017. Ectomycorrhizal fungal response to and flood conditions on leaf photosynthesis, growth and sugar content in
warming is linked to poor host performance at the boreal-temperate ecotone. sugarcane. Plant Production Science 19: 427–437.
Global Change Biology 23: 1598–1609. Johnson D, Ijdo M, Genney DR, Anderson IC, Alexander IJ. 2005. How do
Finzi AC, Abramoff RZ, Spiller KS, Brzostek ER, Darby BA, Kramer MA, plants regulate the function, community structure, and diversity of mycorrhizal
Phillips RP. 2015. Rhizosphere processes are quantitatively important fungi? Journal of Experimental Botany 56: 1751–1760.
components of terrestrial carbon and nutrient cycles. Global Change Biology 21: Karst J, Marczak L, Jones MD, Turkington R. 2008. The mutualism–parasitism
2082–2094. continuum in ectomycorrhizas: a quantitative assessment using meta-analysis.
Fougnies L, Renciot S, Muller F, Plenchette C, Prin Y, de Faria SM, Bouvet JM, Ecology 89: 1032–1042.
Sylla SN, Dreyfus B, B^a AM. 2007. Arbuscular mycorrhizal colonization and Kiers ET, Duhamel M, Beesetty Y, Mensah JA, Franken O, Verbruggen E,
nodulation improve flooding tolerance in Pterocarpus officinalis Jacq. seedlings. Fellbaum CR, Kowalchuk GA, Hart MM, Bago A et al. 2011. Reciprocal
Mycorrhiza 17: 159–166. rewards stabilize cooperation in the mycorrhizal symbiosis. Science 333: 880–
Fry EL, Manning P, Macdonald C, Hasegawa S, De Palma A, Power SA, Singh 882.
BK. 2016. Shifts in microbial communities do not explain the response of Klamer M, Roberts MS, Levine LH, Drake BG, Garland JL. 2002. Influence of
grassland ecosystem function to plant functional composition and rainfall elevated CO2 on the fungal community in a coastal scrub oak forest soil
change. Soil Biology and Biochemistry 92: 199–210. investigated with terminal-restriction fragment length polymorphism analysis.
Garcia K, Delteil A, Conejero G, Becquer A, Plassard C, Sentenac H, Applied and Environmental Microbiology 68: 4370–4376.
Zimmermann S. 2014. Potassium nutrition of ectomycorrhizal Pinus pinaster: Kubartova A, Ranger J, Berthelin J, Beguiristain T. 2008. Diversity and
overexpression of the Hebeloma cylindrosporum HcTrk1 transporter affects the decomposing ability of saprophytic fungi from temperate forest litter. Microbial
translocation of both K+ and phosphorus in the host plant. New Phytologist Ecology 58: 98–107.
201: 951–960. Laird DA, Fleming P, Davis DD, Horton R, Wang B, Karlen DL. 2010. Impact
Garcıa I, Mendoza R, Pomar MC. 2008. Deficit and excess of soil water impact of biochar amendments on the quality of a typical Midwestern agricultural soil.
on plant growth of Lotus tenuis by affecting nutrient uptake and arbuscular Geoderma 158: 443–449.
mycorrhizal symbiosis. Plant and Soil 304: 117–131. Last FT, Mason PA, Ingleby K, Fleming LV. 1984. Succession of fruitbodies of
Gardes M, Bruns TD. 1993. ITS primers with enhanced specificity for sheathing mycorrhizal fungi associated with Betula pendula. Forest Ecology and
basidiomycetes – application to the identification of mycorrhizae and rusts. Management 9: 229–234.
Molecular Ecology 2: 113–118. Leake J, Johnson D, Donnelly D, Muckle G, Boddy L, Read D. 2004. Networks
van der Gast CJ, Gosling P, Tiwari B, Bending GD. 2011. Spatial scaling of of power and influence: the role of mycorrhizal mycelium in controlling plant
arbuscular mycorrhizal fungal diversity is affected by farming practice. communities and agroecosystem functioning. Canadian Journal of Botany 82:
Environmental Microbiology 13: 241–249. 1016–1045.
Geisseler D, Scow KM. 2014. Long-term effects of mineral fertilizers on soil Lilleskov EA, Bruns TD, Horton TR, Taylor D, Grogan P. 2004. Detection of
microorganisms – a review. Soil Biology and Biochemistry 75: 54–63. forest stand-level spatial structure in ectomycorrhizal fungal communities.
Gilbert JA, Steele JA, Caporaso JG, Steinbruck L, Reeder J, Temperton B, Huse FEMS Microbiology Ecology 49: 319–332.
S, McHardy AC, Knight R, Joint I et al. 2012. Defining seasonal marine Lilleskov EA, Hobbie EA, Horton TR. 2011. Conservation of ectomycorrhizal
microbial community dynamics. ISME Journal 6: 298–308. fungi: exploring the linkages between functional and taxonomic responses to
Gosling P, Mead A, Proctor M, Hammond JP, Bending GD. 2013. Contrasting anthropogenic N deposition. Fungal Ecology 2: 174–183.
arbuscular mycorrhizal communities colonizing different host plants show a Liu J, Maldonado-Mendoza I, Lopez-Meyer M, Cheung F, Town CD, Harrison
similar response to a soil phosphorus concentration gradient. New Phytologist MJ. 2007. Arbuscular mycorrhizal symbiosis is accompanied by local and
198: 546–556. systemic alterations in gene expression and an increase in disease resistance in
Grafen A, Hails R. 2002. Modern statistics for the life sciences. Oxford, UK: the shoots. Plant Journal 50: 529–544.
Oxford University Press. Lodge DJ. 1989. The influence of soil moisture and flooding on formation of
Hansen J, Sato M, Ruedy R. 2012. Perception of climate change. Proceedings of VA-endo-and ectomycorrhizae in Populus and Salix. Plant and Soil 117: 243–
the National Academy of Sciences, USA 109: E2415–E2423. 253.
Hartmann A, Rothballer M, Schmid M. 2008. Lorenz Hiltner, a pioneer in Lovelock CE, Andersen K, Morton JB. 2003. Arbuscular mycorrhizal
rhizosphere microbial ecology and soil bacteriology research. Plant and Soil communities in tropical forests are affected by host tree species and
312: 7–14. environment. Oecologia 135: 268–279.
Hawkes CV, Kivlin SN, Rocca JD, Huguet V, Thomsen MA, Suttle KB. 2011. Mallakpour I, Villarini G. 2015. The changing nature of flooding across the
Fungal community responses to precipitation. Global Change Biology 17: central United States. Nature Climate Change 5: 250–254.
1637–1645. McMurdie PJ, Holmes S. 2014. Waste not, want not: why rarefying microbiome
Hazard C, Gosling P, van der Gast CJ, Mitchell DT, Doohan FM, Bending data is inadmissible. PLoS Computational Biology 10: e1003531.
GD. 2013. The role of local environment and geographical distance in Mendoza R, Escudero V, Garcıa I. 2005. Plant growth, nutrient acquisition and
determining community composition of arbuscular mycorrhizal fungi at the mycorrhizal symbioses of a waterlogging tolerant legume (Lotus glaber Mill.) in
landscape scale. ISME Journal 7: 498–508. a saline-sodic soil. Plant and Soil 275: 305–315.

Ó 2018 The Authors New Phytologist (2018)

New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com
 New
12 Research Phytologist

Nagy NE, Fossdal CG. 2013. Host responses in Norway spruce roots induced to Wilson JS, Baldwin DS, Rees GN, Wilson BP. 2011. The effects of short-
the pathogen Ceratocystis polonica are evaded or suppressed by the term inundation on carbon dynamics, microbial community structure and
ectomycorrhizal fungus Laccaria bicolor. Plant Biology 15: 99–110. microbial activity in floodplain soil. River Research and Applications 27:
Newbold LK, Burthe SJ, Oliver AE, Gweon HS, Barnes CJ, Daunt F, van der 213–225.
Gast CJ. 2017. Helminth burden and ecological factors associated with Yang H, Koide RT, Zhang Q. 2016. Short-term waterlogging increases
alterations in wild host gastrointestinal microbiota. ISME Journal 11: 663–675. arbuscular mycorrhizal fungal species richness and shifts community
Nguyen NH, Song Z, Bates ST, Branco S, Tedersoo L, Menke J, Schilling JS, composition. Plant and Soil 404: 373–384.
Kennedy PG. 2016. FUNGuild: an open annotation tool for parsing fungal Yang H, Zang Y, Yuan Y, Tang J, Chen X. 2012. Selectivity by host
community datasets by ecological guild. Fungal Ecology 20: 241–248. plants affects the distribution of arbuscular mycorrhizal fungi:
Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O’Hara RB, evidence from ITS rDNA sequence metadata. BMC Evolutionary Biology
Simpson GL, Solymos P, Stevens MHH, Wagner H. 2007. The vegan 12: 50.
package. [WWW document] URL http://cran.r-project.org/.
Parry S, Marsh T, Kendon M. 2013. 2012: from drought to floods in England
and Wales. Weather 68: 268–274. Supporting Information
Phillips CJ, Marden M, Suzanne LM. 2014. Observations of root growth of
young poplar and willow planting types. New Zealand Journal of Forestry Additional Supporting Information may be found online in the
Science 44: 15. Supporting Information tab for this article:
Reichstein M, Bahn M, Ciais P, Frank D, Mahecha MD, Seneviratne SI,
Zscheischler J, Beer C, Buchmann N, Frank DC et al. 2013. Climate extremes Fig. S1 Boxplots of nutrient content for each sampling time
and the carbon cycle. Nature 500: 287–295.
Rivest D, Lorente M, Olivier A, Messier C. 2013. Soil biochemical properties
points pH, (b) P, K, Mg and NO3 kg 1.
and microbial resilience in agroforestry systems: effects on wheat growth under
controlled drought and flooding conditions. Science of The Total Environment Fig. S2 Monthly average temperature, monthly precipitation and
463–464: 51–60. monthly soil water content from the willow site over the 3-yr
Sauter M. 2013. Root responses to flooding. Current Opinion in Plant Biology 16: sampling period.
282–286.
Shepherd TG. 2015. Climate science: the dynamics of temperature extremes.
Nature 522: 425–427. Fig. S3 Boxplot showing distribution of soil VMC which corre-
Smalla K, Wieland G, Buchner A, Zock A, Parzy J, Kaiser S, Roskot N, Heuer H, sponded with full saturation.
Berg G. 2001. Bulk and rhizosphere soil bacterial communities studied by
denaturing gradient gel electrophoresis: plant-dependent enrichment and seasonal Fig. S4 Evidence of surface water found within July 2012 that
shifts revealed. Applied and Environmental Microbiology 67: 4742–4751.
Smit E, Leeflang P, Glandorf B, van Elsas JD, Wernars K. 1999. Analysis of
was not present in any of the other sampling events.
fungal diversity in the wheat rhizosphere by sequencing of cloned PCR-
amplified genes encoding 18S rRNA and temperature gradient gel Fig. S5 Nonmetric dimensional scaling showing clustering based
electrophoresis. Applied and Environmental Microbiology 65: 2614–2621. on similarity of the root-associated fungal communities and pH
Smith SE, Read DJ. 2010. Mycorrhizal symbiosis. London, UK: Academic Press. shown by colour.
Swaty RL, Gehring CA, Van Ert M, Theimer TC, Keim P, Whitham TG.
1998. Temporal variation in temperature and rainfall differentially affects
ectomycorrhizal colonization at two contrasting sites. New Phytologist 139: Fig. S6 Nonmetric dimensional scaling plots with clustering
733–739. based on similarity of the root-associated fungal communities,
Tedersoo L, Bahram M, P~olme S, K~oljalg U, Yorou NS, Wijesundera R, Ruiz partitioned into before and after extreme rainfall.
LV, Vasco-Palacios AM, Thu PQ, Suija A et al. 2014. Global diversity and
geography of soil fungi. Science 346: 1256688.
Tedersoo L, Bahram M, Toots M, Diedhiou AG, Henkel TW, Kjøller R,
Fig. S7 Average relative abundances and OTU richness of the
Morris MH, Nara K, Nouhra E, Peay KG et al. 2012. Towards global patterns ECM fungi divided by exploration type, sampled from willow
in the diversity and community structure of ectomycorrhizal fungi. Molecular roots over the 3-yr period.
Ecology 21: 4160–4170.
Tedersoo L, Suvi T, Jairus T, Ostonen I, P~olme S. 2009. Revisiting Fig. S8 Volcano plots of change in relative abundance of each
ectomycorrhizal fungi of the genus Alnus: differential host specificity, diversity
and determinants of the fungal community. New Phytologist 182: 727–735.
OTU between seasons plotted against -log10(q) values generated
Theodorou C. 1978. Soil moisture and the mycorrhizal association of Pinus from repeated measure ANOVAs (using data from all sampling
radiata D. don. Soil Biology and Biochemistry 10: 33–37. times).
Thion C, Cebron A, Beguiristain T, Leyval C. 2012. Long-term in situ dynamics
of the fungal communities in a multi-contaminated soil are mainly driven by Table S1 List of taxonomic assignments and references, in addi-
plants. FEMS Microbiology Ecology 82: 169–181.
Unestam T, Sun Y-P. 1995. Extramatrical structures of hydrophobic and
tion to P-values (and subsequent q-values) produced via repeated
hydrophilic ectomycorrhizal fungi. Mycorrhiza 5: 301–311. measure ANOVAs for each individual OTU
Unger IM, Kennedy AC, Muzika R-M. 2009. Flooding effects on soil microbial
communities. Applied Soil Ecology 42: 1–8. Table S2 Relative importance of season, geographical distance
Visser S. 1995. Ectomycorrhizal fungal succession in jack pine stands following (between samples) and soil properties on the root-associated fun-
wildfire. New Phytologist 129: 389–401.
Wagner D, Eisenhauer N, Cesarz S. 2015. Plant species richness does not
gal communities of willow as revealed by PERMANOVA, with
attenuate responses of soil microbial and nematode communities to a flood the pre-shift communities (October 2010, July 2011 and Octo-
event. Soil Biology and Biochemistry 89: 135–149. ber 2012) and post-shift communities (July 2012 and October
Wickham H. 2016. ggplot2: elegant graphics for data analysis. New York, NY, 2012) also analysed again in separate analyses
USA: Springer.

New Phytologist (2018) Ó 2018 The Authors

www.newphytologist.com New Phytologist Ó 2018 New Phytologist Trust
New
Phytologist Research 13

Table S3 Average relative abundance and OTU richness of ECM these OTU to correlate their abundance against time and envi-
families within SRC willow root-associated fungal communities ronmental properties
sampled over the 3-yr period
Please note: Wiley Blackwell are not responsible for the content
Table S4 Table of OTUs which significantly varied in relative or functionality of any Supporting Information supplied by the
abundance over time (as found with repeated measure authors. Any queries (other than missing material) should be
ANOVAs), whilst GLM models were also performed for each of directed to the New Phytologist Central Office.

New Phytologist is an electronic (online-only) journal owned by the New Phytologist Trust, a not-for-profit organization dedicated
to the promotion of plant science, facilitating projects from symposia to free access for our Tansley reviews and Tansley insights.

Regular papers, Letters, Research reviews, Rapid reports and both Modelling/Theory and Methods papers are encouraged.
We are committed to rapid processing, from online submission through to publication ‘as ready’ via Early View – our average time
to decision is <26 days. There are no page or colour charges and a PDF version will be provided for each article.

The journal is available online at Wiley Online Library. Visit www.newphytologist.com to search the articles and register for table
of contents email alerts.

If you have any questions, do get in touch with Central Office (np-centraloffice@lancaster.ac.uk) or, if it is more convenient,
our USA Office (np-usaoffice@lancaster.ac.uk)

For submission instructions, subscription and all the latest information visit www.newphytologist.com

Ó 2018 The Authors New Phytologist (2018)

New Phytologist Ó 2018 New Phytologist Trust www.newphytologist.com