text
stringlengths
0
6.44k
five sites. Core type and relative water level were also included in the best model for the number of
functional guilds (core type: F1,63 = 2.92, p = 0.1378; relative water level: F1,63 = 0.87, p = 0.3521), with an
18% greater number of guilds in peat compared to limestone core islands and an approximately 8%
decrease in the driest compared to wettest sites. When we examined the factors that influence
abundances of the three most common guilds from our dataset, the abundance of only one—plant
pathogen-wood saprotroph—could be explained by any of our measured variables. As the relative
water level decreased, the abundance of the plant pathogen–wood saprotroph guild increased (F1,68
= 5.47, p = 0.022; Figure 3B).
Figure 3. Factors involved in the distribution of fungal functional guilds. The number of functional
guilds was positively associated with understory plant community evenness (A; F1,61 = 10.12, p =
0.002). As the relative water level decreased, the abundance of the plant pathogen-wood saprotroph
guild increased (B; F1,61 = 10.12, p = 0.002). The lines are based on a linear model between variables
with the shaded area indicating the 95% confidence interval and each point representing a site on the
tree islands. Note that functional guild analyses are based on the approximately 20% exact sequence
variants (ESVs) that could be identified at the species level using UNITE.
4. Discussion
To better incorporate fungal communities into restoration planning in terms of both protecting
their diversity and utilizing these hidden players to improve restoration success, a more complete
understanding of which environmental factors and management decisions affect soil fungal
Figure 3. Factors involved in the distribution of fungal functional guilds. The number of functional
guilds was positively associated with understory plant community evenness (A; F1,61 = 10.12, p = 0.002).
As the relative water level decreased, the abundance of the plant pathogen-wood saprotroph guild
increased (B; F1,61 = 10.12, p = 0.002). The lines are based on a linear model between variables with the
shaded area indicating the 95% confidence interval and each point representing a site on the tree islands.
Note that functional guild analyses are based on the approximately 20% exact sequence variants (ESVs)
that could be identified at the species level using UNITE.
Diversity 2020, 12, 0324 9 of 17
4. Discussion
To better incorporate fungal communities into restoration planning in terms of both protecting
their diversity and utilizing these hidden players to improve restoration success, a more complete
understanding of which environmental factors and management decisions affect soil fungal
communities is required. Our study contributes to this goal by showing that (1) fungal diversity and
composition were affected by restoration decision factors that were manipulated on experimental
tree islands during their construction, and (2) variation in several important aspects of fungal
communities was explained by microhabitat differences in other environmental variables of interest
for management, including hydrology and properties of the naturally recruiting plant understory
community. Below, we address these results in more detail by first examining possible mechanisms
through which abiotic factors, and then biotic factors, can contribute to fungal community composition
and diversity. We conclude by briefly discussing directions for future research and the implications for
the Everglades.
4.1. Abiotic Factors: The Role of Hydrology, Island Core Type, and Light Environment in Fungal Diversity
and Composition
Abiotic stressors play important roles in limiting organismal niches [54], and thus abiotic
contexts often contribute substantially to the diversity and composition of communities in
nature [55]. Previous work in the Everglades has identified abiotic features that affect communities
of macro-organisms on tree islands [56,57]. For instance, the tree island plant community structure
is largely driven by water level, nutrient availability, and disturbance (e.g., fire, windstorms,
and drought) [30]. Our study found that the fungal community composition and/or diversity on tree
islands also appear to respond strongly to all three of the abiotic factors investigated—relative water
level, light environment, and core type. In particular, hydrology was important for the composition
and diversity of the soil fungal community, with greater fungal diversity at sites with higher water
levels. We saw a 34% increase in fungal diversity when comparing the five wettest sites to the five
driest sites. This shift in fungal diversity and composition is likely due to the sensitivity of some
species of fungi to dry conditions [58,59] or the increased dispersal of spores to wetter sites [38,60].
Hydrology is one of the most anthropogenically altered aspects of the Everglades, with much of the
historic sheet flow from central to southern Florida now diverted through approximately 1800 miles
of levees and canals [61]. This change has generally led to a change of inundation periods for tree
islands [62]. Given the substantial decrease in fungal diversity we found at dry sites in our study, it is
likely that these major alterations to tree island inundation are also having extensive consequences
for the fungal communities affected by water diversions. As a result of the importance of fungi in
ecosystem functions [1], a loss of fungal diversity on unnaturally dry tree islands may underpin
mechanisms causing tree island loss (e.g., reduced availability of the beneficial symbionts on which
trees depend, or a loss of high-quality decomposers required for nutrient recycling). These shifts in
diversity and composition also suggest the importance of hydrologic restoration targets that consider
fungal ecology, and tree island restoration and creation success may be assisted by a more careful
consideration of these important constituents of soil biota.
In addition, our investigation of abiotic factors demonstrated that early restoration decisions
can have abiding effects on fungal communities years later. In 2003, the experimental islands at
LILA were constructed with two core types—peat and limestone—to represent the natural variation
in Everglades tree islands. We found that differences in the choice of core type affected the overall
fungal community composition, including a significant difference between limestone core and peat
core communities along both axes of variation in community composition. Further research into the
indicator taxa identified for limestone versus peat core islands revealed functional differences among
these taxa. Indicator taxa of peat communities included ectomycorrhizal fungi [63,64], while those in
limestone communities included taxa described as putative wood-decayers and saprotrophs [65–67].
The distinctions in fungal communities on the two core types are presumably the result of the different
Diversity 2020, 12, 0324 10 of 17
soil environments they provide [68]. For instance, previous studies of the tree islands at LILA have
suggested increased water retention on peat islands compared to limestone [69,70]. The presence of
limestone may also change the mineral content of soils through inputs of calcium carbonate, which has
been associated with significant shifts in soil microbial communities in other systems [71]. In addition,
the difference in some of these indicator taxa could also be the result of the presence of the tree
Morella cerifera, which is the only surveyed species on the islands that has been reported to associate
with ectomycorrhizal fungi (in addition to its association with arbuscular mycorrhizal fungi) [72].
Morella cerifera had a somewhat lower survival rate on peat core islands, indicating they are more
stressed on these islands, which may result in an increased reliance on ectomycorrhizal fungi to
ameliorate stress [42]. Therefore, we examined whether the presence of M. cerifera was predictive of
the estimated abundance (total number of reads) of the ectomycorrhizal indicator taxa. While we did
not find support for a clear relationship between M. cerifera presence and these ectomycorrhizal fungi
(p > 0.05 in a distribution-free randomization test with the presence of M. cerifera and island identity in
the model), additional studies that investigate unmeasured aspects of M. cerifera biology, such as tree
size, and the root colonization of this tree may help elucidate if and how it influences ectomycorrhizal
abundance. More generally, the effect of core type in our study indicates that initial decisions in tree
island construction have cascading effects on the fungal community, even 15 years later.
We also found that as light availability increased (as indicated by canopy openness), fungal richness
also increased. While light environments can have consequences for the soil microbiome through effects
on soil temperature and moisture [73], light may act indirectly through its effects on the understory
plant community [74]. Further investigation of this possibility using an RDA (Redundancy Analysis)
with plant understory community composition data as the response and the same explanatory abiotic
and biotic factors used in the fungal analyses showed that canopy openness did not explain variation