New Results

Comparing trait syndromes between Taiwanese subtropical terrestrial and epiphytic ferns and lycophytes at the species and community level

Kenny Helsen, Tsung-Yi Lin, David Zelený
doi: https://doi.org/10.1101/2021.09.06.459074

Summary

  • While functional trait-trait and trait-environment relationships are well studied in angiosperms, it is less clear if similar relationships, such as the leaf economics spectrum (LES), hold for ferns and lycophytes. Similarly, studies exploring potential differences in trait-trait and trait-environment relationships between terrestrial and epiphytic fern communities in a given ecosystem are largely lacking.

  • We measured nine leaf traits for 76 terrestrial and 43 epiphytic fern and lycophyte species across 59 vegetation plots along an elevation gradient in the subtropical forest of Northern Taiwan. We explored trait-trait and trait-environment relationships at both the species- and community level for both species groups.

  • Epiphytes differed from terrestrial ferns and lycophytes in species- and community-level trait values, mainly reflecting responses to higher drought and nutrient stress. The angiosperm LES was reflected in the trait-trait correlations of terrestrial ferns, but not of epiphytes. This suggests that epiphytic trait patterns are mainly shaped by water, rather than nutrient availability. Trait-environment relationships were nonetheless more-or-less similar for several drought-related traits across both species’ groups.

  • This study illustrates that ferns and lycophyte trait patterns are not equivalent for epiphytic and terrestrial species or communities, and should not be extrapolated across species groups or between the species- and community-level.

Introduction

Functional traits, i.e. morphological, physiological or phenological features that indirectly impact an organism’s fitness (Violle et al., 2007), often dictate under what environmental conditions plants will be able to thrive. This environmental trait filtering has been extensively studied and has increased our mechanistic understanding of community-level species composition (Lavorel & Garnier, 2002; Kraft et al., 2015; Bruelheide et al., 2018). Several functional traits furthermore covary systematically across plant species. The leaf economics spectrum (LES), for example, links several leaf traits that govern a species’ leaf investment strategy, ranging from highly resource acquisitive, productive species characterized by high specific leaf area (SLA) and leaf nitrogen and phosphorous content, to more resource conservative, slow-growing species with high leaf dry matter content (LDMC) and leaf lifespan (Wilson et al., 1999; Wright et al., 2004; Díaz et al., 2016). The LES is even retained when scaling up species-level traits to the community level (Bruelheide et al., 2018). The discovery of the LES, and the links between LES traits and plant responses to several environmental stressors, such as low nutrient availability (Ordoñez et al., 2009; Hodgson et al., 2011) and low temperatures (Wright et al., 2005; Dong et al., 2020), have greatly increased our understanding of plant functioning and plant-environment interactions.

Most of this trait work has been performed on seed plants (spermatophytes) though, and it is unclear if the same trait-trait and trait-environment relationships hold for older vascular plant lineages, such as ferns (Polypodiophyta) and lycophytes (Lycopodiophyta) (Karst & Lechowicz, 2007). Since both ferns and lycophytes form separate evolutionary lineages from seed plants (Smith et al., 2006), morphological and physiological traits are not guaranteed to be functionally equivalent across these groups (Peppe et al., 2014). For example, recent work has suggested that the evolution of photosynthesis-related trait co-regulations and trade-offs, such as the LES, have been more strongly constrained by leaf water transport capacity in ferns (Sessa & Givnish, 2014; Zhang et al., 2014). This seemed to be caused by the more limited potential for controlling evaporation, low water-use efficiency and xylem hydraulic limitations in ferns (Brodribb & Holbrook, 2004; Zhang et al., 2014). This makes them more prone to drought stress and might explain the overall lower photosynthetic capacity and nutrient concentrations of fern leaves compared to angiosperm leaves (Tosens et al., 2016). A few studies have nevertheless found some support for LES trait relationships across a limited number (i.e. 11-35 species) of terrestrial fern and lycophyte species of temperate (Karst & Lechowicz, 2007; Sessa & Givnish, 2014; Tosens et al., 2016) and (sub)tropical ecosystems (Campany et al., 2019; Lin et al., 2020). However, strong evidence for global LES trade-offs across fern species, similar to those known for seed plants, is currently still lacking. Our limited understanding of trait-trait relationships in ferns and lycophytes (cf. Kessler et al., 2016) is nonetheless surprising, considering that they contain more than 10,000 species in the world, second only to flowering plants (angiosperms) in terms of plant species diversity (Smith et al., 2006; Tosens et al., 2016).

Almost one third (29%) of all fern species are epiphytic (Dubuisson et al., 2009). Especially in (sub)tropical regions, ferns form the predominant vascular epiphytic group, together with orchids (Schellenberger Costa et al., 2018; Campany et al., 2021). The beforementioned constrain of photosynthesis-related traits by water relationships are expected to be even more severe for epiphytic species, since they are likely to experience more frequent and extended drought periods than terrestrial species (Dubuisson et al., 2009; Zhang et al., 2014; Aros-Mualin et al., 2021). Epiphytes are additionally confronted with low nutrient availability, since they have no access to the soil (Watkins Jr et al., 2007). On average, epiphytic ferns indeed have higher water-use efficiency (δ13C isotope ratio) and lower SLA, stomatal density and leaf N than terrestrial species in tropical forests, strongly suggesting the importance of water and nutrient stress in shaping their traits (Watkins Jr et al., 2007; Nitta et al., 2020; Campany et al., 2021). It is consequently unclear if similar LES-related trait relationships can be expected for epiphytic and terrestrial ferns. A recent study on epiphytic plant communities along elevation in Tanzania, for instance, found markedly different trait-environment correlations for epiphytic ferns and epiphytic flowering plants (Schellenberger Costa et al., 2018).

In ecosystems where both terrestrial and epiphytic ferns co-occur, such as (sub)tropical rain- and cloud forests, the environmental filtering of their respective communities might be governed by very different drivers. The proneness to drought of epiphytes suggests that their community composition will be more strongly filtered by water availability through precipitation, relative air humidity and ground fog cover (Dubuisson et al., 2009; Zhang et al., 2014; Aros-Mualin et al., 2021). Additionally, woody species composition directly affects epiphyte substrate quality and availability, thus further impacting epiphyte community composition. Terrestrial species, on the other hand, will also be filtered by soil variation, through differences in soil nutrient and water availability, while the impact of temperature might be important for both lifeforms. Some research has, for example, found that several morphological leaf traits vary in response to temperature (elevation) and humidity gradients for both tropical terrestrial and epiphytic fern communities (Kessler et al., 2007; Kluge & Kessler, 2007; Salazar et al., 2012; Petter et al., 2016). Nevertheless, studies separately exploring the importance of different environmental drivers on community-level trait composition of terrestrial and epiphytic ferns simultaneously for a given ecosystem are largely lacking (however see Wegner et al., 2003).

In this study, we measured nine functional leaf traits, including four LES-related traits, that are known to respond to nutrient, water and cold stress in seed plants, for 76 terrestrial and 43 epiphytic fern and lycophyte species across 59 10 m × 10 m plots. These plots were spread along an elevation gradient from 870 to 2130 m a.s.l. in the subtropical forests of Northern Taiwan. Next to the measured species-level traits, we also calculated community mean trait values for each plot. Using this dataset, we explored the following research questions:

  • - Are traits of terrestrial and epiphytic species and communities systematically different?

  • - Do similar trait-trait correlations occur for terrestrial and epiphytic species and communities, and do these relationships mirror known trait patterns of seed plants, such as the LES?

  • - Do we find similar trait-environment relationships for the nine measured traits and elevation, ground fog frequency and heat load, for terrestrial and epiphytic ferns?

Materials and Methods

Study design

The study was conducted in the Wulai district of New Taipei City, in northern Taiwan. In this area we established 59 10 m × 10 m vegetation plots in natural, undisturbed vegetation along an elevational gradient ranging from Mount Meilu (870 m a.s.l., 24.85°N 121.53°E) to Mount Taman (2130 m a.s.l., 24.71°N 121.45°E) (Supporting Information Fig. S1). Vegetation along this gradient ranged from lowland Pyrenaria-Machilus subtropical winter monsoon forest to lower cloud zone Quercus montane evergreen broad-leaved cloud forest and Chamaecyparis montane mixed cloud forest (Li et al., 2013). The vegetation plots were evenly spread across six elevation zones (with centres at 850, 1100, 1350, 1600, 1850 and 2100 m a.s.l. ±50 m), with ten plots per elevation zone spaced at least 50 m apart (except for the 1850 m zone, where only nine plots were located due to logistic constraints). At each elevation zone, plots were positioned along a topographic gradient across the ridge, ranging from the southwest facing (leeward) to northeast facing slope (windward).

The climate in the study region is classified as a humid subtropical climate, with monthly mean temperature in January and July ranging from 12.8□ and 25.6□ (near Mt. Meilu) to 8.7□ and 21.1□ (near Mt. Taman), respectively (obtained from on-site microclimatic measurements in a subset of plots; Zelený et al., unpublished). Mean annual precipitation ranges from 2033 to 3396 mm along the gradient, with no obvious dry season (Lalashan and Fushan weather stations, Taiwan Central Weather Bureau). Microclimatic loggers installed at three sites at each elevation zone furthermore indicated very high relative humidity for most plots (median between 97.5 and 100% RH; Zelený et al., unpublished). Parent materials of the study area mainly consists of Miocene and Oligocene sandstone and slate (Central Geological Survey, MOEA). The soils covering this material are strongly acidic and have very high organic matter content, especially in higher elevation and on the ridges. Soils located near ridges have higher degrees of podzolization than those located on steeper slopes, due to lower degrees of soil layer disturbance on the flat microtopography of the ridges (Lin et al., 1988).

Species composition and functional traits

In each plot, both the understory (terrestrial) and epiphytic fern and lycophyte species composition was recorded (presence-absence data), in May to October 2017 or 2018. Epiphytic species within reach were collected directly, while those out of reach were identified using binoculars and sampled for trait measurements using a 11m long telescopic knife. We measured nine functional leaf traits for 48 understory and 34 epiphytic fern and lycophyte species (respectively 63% and 79% of all recorded understory and epiphytic fern and lycophyte species) (Supporting Information Table S2). For each of these species, we collected leaves from, on average, 5 (range 1-12) individuals, directly in our plots, or within a 50 m radius. Only mature individuals were selected, preferentially those bearing sporangia when present. For species occurring across different elevation zones, we attempted to collect individuals across this full range. From each collected individual, we selected 1-6 leaves (fronds) for trait measurements, using the following criteria: 1) for dimorphic species, we selected the sterile tropophyll, rather than the fertile sporophyll; 2) the frond should be fully expanded and matured without visible herbivore or parasite-induced damage. For non-dimorphic species, traits were measured on sori-containing fronds. Collected fronds were stored in wet, sealed plastic bags at low temperatures (< 10°C) for a minimum of 12h before trait measurement, to allow full rehydration (cf. Pérez-Harguindeguy et al., 2013).

The measured traits consisted of leaf dry matter content (LDMC, mg g-1), specific leaf area (SLA, mm2 mg-1), leaf nitrogen content (leaf N, mg g-1), area-based leaf chlorophyll content (SPAD units), leaf area (cm2), leaf thickness (mm), equivalent water thickness (EWT, mg mm-2), leaf 13C/12C stable isotope ratio (δ13C, ‰) and leaf 15N/14N stable isotope ratio (δ15N, ‰). EWT is sometimes referred to as ‘succulence’ (Mantovani, 1999; Féret et al., 2019). The first four of these traits are part of the LES for vascular plants (cf. Wilson et al., 1999; Wright et al., 2004), and thus position species along a gradient from resource acquisitive (high SLA and leaf N) to resource conservative (high LDMC and area-based leaf chlorophyll). The next four traits are expected to relate to drought stress, with small leaf area and high leaf thickness, EWT and δ13C indicative of drought tolerance, with the latter trait a proxy of long-term water use efficiency (Farquhar et al., 1982; Medeiros et al., 2019; Maréchaux et al., 2020). δ15N reflects the nitrogen source used by a plant, with values around 0 ‰ usually indicating nitrogen fixation, while values around -2 and -6 ‰ indicating plant nitrogen acquisition through arbuscular and ectomycorrhiza, respectively (Craine et al., 2015). Trait measurements largely followed standard protocols (Pérez-Harguindeguy et al., 2013), but were partly adapted for fern leaves (Supporting Information Appendix S3 for all details). For four epiphytic ferns, leaves were too small to measure leaf chlorophyll content (Supporting Information Table S2).

For all traits, except leaf area, values with a Z-score larger than 2.5 at the species level were considered outliers caused by measurement error and were removed from the final dataset (<1% of data points). Next, all leaf level traits were averaged to the species level, resulting in a 48 species × 9 traits matrix for the terrestrial species and a 34 species × 9 traits matrix for the epiphytic species. Trait values were also translated to the plot-level by calculating community mean (CM) trait values, i.e. the average trait value across all species in the plot. Since no abundance data was collected, CM values were not abundance-weighted. Leaf area was logarithmically transformed before CM calculation for both species groups.

Climate proxies

In each plot we recorded exact elevation using a topographic map combined with the GPS coordinates (GPSMAP 64st, Garmin, USA), aspect using a compass (SILVA, Sweden) and slope using a clinometer (SUUNTO PM-5/360 PC Clinometer, SUUNTO, Finland). We then calculated heat load from the aspect folded on the prevailing wind direction (45°) and slope with equation 2 of McCune & Dylan (2002). We also used the ground fog frequency raster map for Taiwan of Schulz et al. (2017) (250 m per pixel resolution) to extract average annual ground fog frequency for each plot. Variation inflation factors indicated that the three climate proxies, namely elevation, fog frequency and heat load, were largely independent of each other (VIF < 1.5) (Zuur et al., 2010).

Species-level analysis

Species-level trait-trait relationships were visualized through three principal component analyses (PCA) on the standardized species × trait matrices (i.e. zero mean, unit standard deviation), one for all species combined and one for either the terrestrial and epiphytic species separately. For all three ordinations, the first two PC axes together contained around 61-64% of the total trait variation (Supporting Information Table S4). Using the first three axes of the PCA performed on the full dataset, we constructed separate trait hypervolumes for the epiphytic and terrestrial species with the ‘hypervolume’ R package (Blonder et al., 2014; Blonder, 2018), using the protocol described in Helsen et al. (2020). These hypervolumes were used to quantify the trait space size and overlap for both species’ groups. Since PCA does not allow missing trait values, the missing leaf chlorophyll content values of four epiphytic species were replaced by mean chlorophyll content across all epiphytic species. All other analyses were performed without replacement of missing trait values. Pairwise species-level trait-trait correlations were tested separately for terrestrial and epiphytic species using Spearman rank correlations. Differences in average trait values between epiphytic and terrestrial species were assessed with a Mann-Whitney U test for each trait.

Community-level analysis

Community-level trait-trait relationships were visualized through three PCA analyses on the standardized plot × CM trait matrices, one for all species combined and one each for the terrestrial and epiphytic species separately. For all three ordinations, the first two PC axes together contained around 72-84% of the total CM trait variation (Supporting Information Table S5). CM trait space size and overlap for both species’ groups was again assessed through trait hypervolume construction, based on the first three axes of the PCA performed on the full plot × CM trait matrix. Pairwise community-level trait-trait correlations were tested separately for terrestrial and epiphytic species using Spearman rank correlations. Differences in average trait values between epiphytic and terrestrial species were additionally assessed with Wilcoxon signed-rank tests for each trait. For each trait, we additionally performed a Spearman rank correlation between the plot-level CM trait of the epiphytic species and that of the terrestrial species.

Lastly, to explore relationships between the measured traits and the three climate proxies (elevation, fog frequency and heat load), we used the fourth-corner approach, which calculates (weighted) Pearson correlations between a single trait from the standardized species × trait matrix and a single environmental variable from the plot × environmental data matrix, weighted by the plot × species matrix (Legendre et al., 1997; Dray & Legendre, 2008; ter Braak et al., 2018). We computed both the original fourth-corner correlation coefficient and the Chessel fourth-corner correlation coefficient, following the recommendations of Peres-Neto et al. (2017). The latter is the fourth-corner correlation coefficient divided by its maximum attainable value, and thus provides a relative measure of how well the trait-environment correlation explains the species distribution (Peres-Neto et al., 2017). The fourth corner correlations were tested by the ‘max test’ proposed by ter Braak et al. (2012), which overcomes Type-I error inflation issues often observed during trait-environment correlations (ter Braak et al., 2012, 2018). The fourth-corner approach was performed separately for the epiphytic and terrestrial species/traits datasets. Before analyses, heat load was squared and leaf area logarithmically transformed for both species groups.

Significance levels of all trait-trait correlations and average trait comparisons between epiphytic and terrestrial species at both the species- and community-level were adjusted for Type-I error inflation using the false discovery rate method with the ‘p.adjust’ function in the ‘stats’ R base package (Benjamini & Hochberg, 1995). Ordinations were performed with the ‘vegan’ R package (Oksanen et al., 2017), and fourth-corner analyses with the ‘weimea’ R package (Zelený, 2018; https://github.com/zdealveindy/weimea) all using R version 4.0.5 (https://www.r-project.org/).

Results

Species-level

Comparing the traits of the 48 terrestrial and 34 epiphytic fern and lycophyte species in our study showed that, on average, epiphytic species have lower SLA, leaf area, δ15N and leaf N and higher EWT and δ13C than terrestrial species (Fig. 1, Supporting Information Table S6). This was reflected in the PCA trait space, where epiphytic and terrestrial species were largely segregated (Fig 2A, Supporting Information Fig. S7), and thus both had a relatively large proportions of unique trait space (69.8 and 58.1%, respectively), following the hypervolume construction. The epiphytic species trait hypervolume was furthermore 28.1% larger than the terrestrial species trait hypervolume, despite containing less species. This seemed to be mainly caused by a larger variation in LDMC, EWT, leaf thickness and leaf chlorophyll values among epiphytic species, compared to terrestrial species (Fig. 1).

Figure 1.
Figure 1. Violin plots for each trait separately, at both the species and community (CM) level.

Separate plot for epiphytic (dark purple) and terrestrial species (light green). Asterix indicates significant difference between epiphytic and terrestrial species (Supporting Information Table S6).

Figure 2.
Figure 2. Biplots for the performed principal component analyses on A. the full species × trait matrix, B. the epiphytic species × trait matrix, C. the terrestrial species × trait matrix, D. the plot × community mean (CM) trait matrix using all species, E. the plot × epiphytic CM trait matrix and F. the plot × terrestrial CM trait matrix.

Species visualized as codes (see Supporting Information Table S2), plots visualized as points, traits visualized as vectors. Light green = terrestrial species, dark purple = epiphytic species. Chl = leaf chlorophyll content, δ13C = the leaf 13C/12C stable isotope ratio, δ15N = the leaf 15N/14N stable isotope ratio, EWT = equivalent water thickness.

When comparing trait patterns separately for epiphytic and terrestrial species, different patterns emerge. For epiphytic species, LDMC seems fully disconnected from the LES, showing no relationship to leaf chlorophyll or leaf N, and even a slightly positive correlation with SLA (Fig 1B, Supporting Information Fig. S8). SLA is also more strongly related to EWT and leaf thickness for epiphytic species, than to the other leaf economics traits (Supporting Information Fig. S8). Epiphytes seem to cluster in three more-or-less distinct groups, one group with high-LDMC species, one group with high values for EWT and leaf thickness and one group with less extreme trait values, that largely overlaps the terrestrial species trait space (Fig. 2A&B).

For terrestrial species, the four measured leaf economics traits show the theoretical expected links (Fig. 2C), although these pairwise correlations are rather weak (Supporting Information Fig. S8). Overall, terrestrial species seem to be loosely spread along a LES axis, with high SLA species on one side and high LDMC species on the other side. An additional group of terrestrial species exhibits large EWT and leaf thickness, with intermediate SLA and LDMC values (Fig. 2C). This latter group is differentiated from the epiphyte species group with high EWT/ leaf thickness, by higher leaf area and leaf N levels (Fig. 2A).

Community-level

Community-level trait patterns largely mirror those observed at the species-level. Both the differentiation between the terrestrial and epiphytic species trait space (86.6% and 92.8% unique trait space for terrestrial and epiphytic species, respectively) and the difference in hypervolume size between the species groups (46.5% larger for epiphytes), are more pronounced at the community-level. Note that the constructed hypervolumes are larger than the convex hulls visualized in Fig. 2D, due to their probabilistic nature, explaining the slight overlap in hypervolumes but not convex hulls. Besides reflecting the species-level differences in SLA, leaf area, EWT, δ13C, δ15N and leaf N between epiphytic and terrestrial species, community-level trait patterns also show higher CM LDMC and leaf thickness, and lower CM leaf chlorophyll for epiphytic species (Fig. 1, Supporting Information Table S6).

Not surprisingly, several pairwise trait correlations observed at the species-level were also present at the community-level for both species’ groups. However, community-level patterns did not completely mirror those at the species-level. For example, while δ13C and δ15N showed almost no relationships to any other trait at the species-level for epiphytes (except for a link between δ15N and leaf N), both traits were each related to four other traits at the community-level. LMDC was also more strongly linked to several other traits at the community-level than at the species-level, for both epiphytic and terrestrial species (Appendices S8 & S9). For instance, the positive correlation between species-level SLA and LDMC for epiphytes became even stronger at the community-level (Fig. 2, Appendices S8 & S9).

Community-level traits were furthermore only positively correlated between epiphytic and terrestrial species across the vegetation plots for LDMC, leaf thickness and EWT (Fig. 3, Supporting Information Table S10). For the other traits, CM values were not related, while CM δ13C was even negatively correlated across lifeforms (Fig. 3, Supporting Information Table S10). Consequently, the results of the fourth-corner analysis showed that trait-environment relationships were not completely similar for both species groups. While elevation was negatively related to leaf thickness and EWT for both species groups, LDMC also showed a strong positive relationship with elevation only for terrestrial species, and δ13C was negatively related with elevation for epiphytic species (Table 1). Similarly, fog frequency was negatively related to EWT, and heat load was negatively related to LDMC, for both species’ groups. Fog frequency was additionally positively related to LDMC for terrestrial species, and heat load showed relationships with three additional traits for epiphytic species (Table 1).

View this table:
Table 1. Fourth-corner results for individual trait – environment relationships for epiphytic and terrestrial species separately.
Figure 3.
Figure 3. Scatterplots for pairwise correlations between plot-level terrestrial and epiphytic fern and lycophyte community mean (CM) trait values.

Regression line + SE presented for significant correlations (see Supporting Information Table S10). Each datapoint corresponds to one vegetation plot, with colours indicating plot elevation. δ13C = the leaf 13C/12C stable isotope ratio, δ15N = the leaf 15N/14N stable isotope ratio, EWT = equivalent water thickness.

Discussion

Trait differences between epiphytic and terrestrial species

Epiphytic species were functionally differentiated from terrestrial species in our study, with six out of the nine measured leaf traits significantly different, at both the species and community level. These differences are in agreement with fern studies in other (sub)tropical regions in Asia, Central America and Polynesia. Higher leaf thickness, EWT and δ13C of epiphytic species is usually attributed to the higher frequency and intensity of drought events experienced by these species, compared to terrestrial ferns (Watkins Jr et al., 2007; Campany et al., 2021). Lower leaf area for epiphytes has also been linked to water relations in a study in French Polynesia (Nitta et al., 2020), although several other studies found no significant difference in leaf area between epiphytic and terrestrial ferns (Zhang et al., 2014; Campany et al., 2021). The importance of drought stress for epiphytic ferns is further supported by their lower stomatal density observed in previous studies (Zhang et al., 2014; Campany et al., 2021).

Other trait differences, such as low SLA and leaf N of epiphytes compared to terrestrial species are also in agreement with previous studies (Watkins Jr et al., 2007; Zhang et al., 2014; Nitta et al., 2020; Campany et al., 2021), and likely reflect a shift along the LES towards more resource conservative strategies for epiphytes (Wright & Cannon, 2001; Wright et al., 2004), driven by the lower nutrient availability experiences by epiphytes compared to terrestrial species (Watkins Jr et al., 2007). SLA differences could, however, also been indirectly caused by water stress avoidance (Campany et al., 2021). The lower δ15N signature of epiphytes has been hypothesized to reflect the uptake of depleted, atmospherically derived N sources, through precipitation and fog (Watkins Jr et al., 2007; Craine et al., 2015). Although not different at the species level, the community-level higher LDMC and lower leaf chlorophyll for epiphytes might also reflect more resource-conservative strategies, and suggest that epiphytes with high LDMC and low leaf chlorophyll are more common at the community-level (Wright et al., 2004; Hodgson et al., 2011).

Trait variation was considerably larger among epiphytic species and communities than among terrestrial species and communities, reflecting patterns from previous studies (Watkins Jr et al., 2007; Nitta et al., 2020). This high variation seems to reflect the co-occurrence of three more-or-less distinct drought-coping trait syndromes for epiphytic ferns (Kessler et al., 2007). The first group consists of species with very high LDMC and mainly contain filmy ferns (Hymenophyllaceae) (Supporting Information Fig. S7). These species, such as Hymenophyllum badium and H. polyanthos, can completely dry out when under water stress, and rehydrate after a drought period (Garcés Cea et al., 2014). Kessler et al. (2007) classified this as a ‘poikilohydric’ drought strategy. The second group, termed ‘xeromorphic’ by Kessler et al. (2007), contains species that prevent desiccation through thick, fleshy leaves with high leaf thickness and EWT (e.g. Lemmaphyllum microphyllum and bird’s nest ferns such as Asplenium antiquum). The third group has more intermediate trait values, with seemingly no pronounced drought-related traits and contains both mesomorphic species that avoid drought by growing in less drought-prone microclimates (e.g. Lepidomicrosorium ningpoense and Vandenboschia auriculata) and drought-deciduous species that avoid water stress by shedding their leaves during drought. The latter species usually have succulent rhizomes and mainly occur in the Davalliaceae and Polypodiaceae families in our study (e.g. Arthromeris lehmannii, Davallia clarkei, Goniophlebium amoenum). This higher trait variation among epiphytic species has been attributed to the strong vertical gradients in light intensity, temperature and humidity occurring in forests (Petter et al., 2016). This allows epiphytes to sort vertically in different niches from the dark and humid understory to the sunny and dry outer canopy, thus presenting a more varied environment as experienced by terrestrial species, who can only sort horizontally along the forest floor (Hietz & Briones, 1998; Petter et al., 2016; Nitta et al., 2020).

Terrestrial species did not show similar distinct species groups as epiphytes, mainly due to the absence of species with very high LDMC (Poikilohydric group) and leaf thickness and EWT (xeromorphic group). Although only moderate compared to the xeromorphic epiphytes, some terrestrial species did nonetheless express higher leaf thickness and EWT values, thus likely presenting moderate xeromorphic adaptations (e.g. Deparia formosana, Diplazium donianum var. donianum, Polystichum integripinnum). Most terrestrial species, however, mainly spread along a LES-like axis, of which species on the conservative side (e.g. Diplopterygium glaucum, Polystichum parvipinnulum) fully overlap with the mesomorphic and drought-deciduous epiphytes in the trait space. The acquisitive side of this gradient contains several terrestrial (e.g. Hymenasplenium adiantifrons, Monachosorum henryi), but no epiphytic species. Overall, these trait patterns support the more severe drought (only poikilohydric and pronounced xeromorphic epiphytes) and nutrient limitation (only acquisitive terrestrial species) experienced by epiphytic compared to terrestrial species.

Trait-trait patterns

For terrestrial species, our data shows LES-related correlations between SLA, LDMC, leaf chlorophyll and leaf N. While these relationships were relatively weak at the species-level, they strengthened at the community-level. These results are in agreement with previous studies that observed similar LES relationships for terrestrial ferns as for angiosperms (Karst & Lechowicz, 2007; Sessa & Givnish, 2014; Tosens et al., 2016; Campany et al., 2019; Lin et al., 2020).

For epiphytic species, LES patterns were visible at neither the species, nor the community level. This is in agreement with previous work that found no correlation between SLA and leaf N (Campany et al., 2021) and SLA and maximum photosynthetic rate for epiphytic ferns (Zhang et al., 2014). These patterns are likely due to a combination of factors. Firstly, epiphytic species seem to consist of only relatively resource-conservative species, thus lacking representatives at the acquisitive side of the spectrum. Secondly, the strong divergent trait adaptations to drought across the epiphytic groups also affected traits traditionally aligned with the LES. For example, LDMC is usually negatively related to SLA across species and communities, but unique trait composition of Poikilohydric fern species resulted in a positive correlation between both traits.

Trait-trait correlations did not only differ strongly between both species’ groups, but also between the species and community level. This shows that trait patterns should not be extrapolated from the species to the community level, since species filtering at different sites can result in different trait – trait correlations at the community level. Similarly, trait patterns cannot be extrapolated from terrestrial to epiphytic fern species.

Trait-environment patterns

Despite the stark differences in mean traits between epiphytic and terrestrial species, several drought-related traits responded similarly to the climate proxies for both species’ groups. These responses, such as increased leaf thickness and EWT at low elevation and fog frequency and high heat load seem to support the importance of water availability for fern community and trait composition (Kluge & Kessler, 2007; Petter et al., 2016; Medeiros et al., 2019). The positive correlation of CM of leaf thickness and EWT across species groups further illustrates their similar environmental response. The negative relationship between elevation and δ13C for epiphytic species again suggests higher proneness to drought and associated higher water-use efficiency at low elevation (Farquhar et al., 1982; Maréchaux et al., 2020). This response was not found for terrestrial species, even resulting in an unexpected negative correlation for δ13C across species groups.

The LES traits SLA, leaf N and leaf chlorophyll did not respond to any of the measured climate proxies for either of the two species groups. Several studies nevertheless found a significant decrease in SLA with elevation for both terrestrial and epiphytic ferns (Kessler et al., 2007; Salazar et al., 2012; Nitta et al., 2020). The CMs of these traits were furthermore not correlated across species groups in our study, suggesting that their community-level variation is governed by different processes. Soil variation, for example, is expected to only directly impact leaf economics traits of terrestrial, but not epiphytic species (Watkins Jr et al., 2007). We indeed previously observed significant effects of soil composition on leaf N of terrestrial ferns in our study area (Helsen et al., 2021). LDMC, however, did show strong relationships with elevation and fog for terrestrial species and moderate relationship with heat load for both species’ groups. The increase of LDMC with elevation might reflect a shift to more conservative leaf traits for terrestrial species due to lower temperatures, as often observed for angiosperms (cf. Helsen et al. 2018), or due to decreased nutrient availability at higher elevation. The LDMC patterns, however, more likely reflect water availability patterns. The consistently strong negative correlation between LDMC and EWT at both the species and community level for both species groups, further supports this.

Conclusion

Epiphytic ferns and lycophytes differ in mean species-level and community-level trait values from terrestrial fern and lycophyte species along an elevation gradient in Northern Taiwan. These trait differences seem to occur because epiphytes experience higher drought and nutrient stress than terrestrial species. Trait-trait correlations traditionally associated with the LES were present for terrestrial species. They did, however, not occur for epiphytes, likely because epiphytic trait patterns are driven by stronger gradients in water than in nutrient availability. Trait–environment relationships were more-or-less similar for several drought-related traits across both species’ groups, while LES traits were not coordinated across species groups at the community level. Overall, these results illustrate that trait patterns are not equivalent for epiphytic and terrestrial species nor communities and should not be extrapolated across species groups or between the species- and community-level.

Author Contribution

TYL and DZ conceived the original idea and collected data. KH elaborated the idea, analyzed data and lead the manuscript writing, with contributions from TYL and DZ. All coauthors commented on the final version of the manuscript. Data were collected as the part of Master thesis conducted by TYL at the National Taiwan University.

Data Availability

Upon acceptance, the raw data will be made available on the Dryad repository.

Acknowledgements

We would like to thank all volunteers who contributed to fieldwork and trait measurements. This study was supported by Ministry of Science and Technology, Taiwan (106-2621-B-002-003-MY3 and 109-2811-B-002-644).

References

  1. ↵
    Aros-Mualin D, Noben S, Karger DN, Carvajal-Hernández CI, Salazar L, Hernández-Rojas A, Kluge J, Sundue MA, Lehnert M, Quandt D, et al. 2021. Functional diversity in ferns is driven by species richness rather than by environmental constraints. Frontiers in Plant Science 11: 615723.
    OpenUrl
  2. ↵
    Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society. Series B 57: 289–300.
    OpenUrlCrossRefWeb of Science
  3. ↵
    Blonder B. 2018. Hypervolume concepts in niche- and trait-based ecology. Ecography 41: 1441–1455.
    OpenUrlCrossRef
  4. ↵
    Blonder B, Lamanna C, Violle C, Enquist BJ. 2014. The n-dimensional hypervolume. Global Ecology and Biogeography 23: 595–609.
    OpenUrl
  5. ↵
    ter Braak CJF, Cormont A, Dray S. 2012. Improved testing of species traits–environment relationships in the fourth-corner problem. Ecology 93: 1525–1526.
    OpenUrlPubMed
  6. ↵
    ter Braak CJF, Peres-Neto PR, Dray S. 2018. Simple parametric tests for trait– environment association. Journal of Vegetation Science 29: 801–811.
    OpenUrl
  7. ↵
    Brodribb TJ, Holbrook NM. 2004. Stomatal protection against hydraulic failure: a comparison of coexisting ferns and angiosperms. New Phytologist 162: 663–670.
    OpenUrlCrossRefWeb of Science
  8. ↵
    Bruelheide H, Dengler J, Purschke O, Lenoir J, Jiménez-Alfaro B, Hennekens SM, Botta-Dukát Z, Chytrý M, Field R, Jansen F, et al. 2018. Global trait–environment relationships of plant communities. Nature Ecology and Evolution 2: 1906–1917.
    OpenUrl
  9. ↵
    Campany CE, Martin L, Watkins Jr JE. 2019. Convergence of ecophysiological traits drives floristic composition of early lineage vascular plants in a tropical forest floor. Annals of Botany 123: 793–803.
    OpenUrl
  10. ↵
    Campany CE, Pittermann J, Baer A, Holmlund H, Schuettpelz E, Mehltreter K, Watkins Jr JE. 2021. Leaf water relations in epiphytic ferns are driven by drought avoidance rather than tolerance mechanisms. Plant, Cell & Environment 44: 1741–1755.
    OpenUrl
  11. ↵
    Craine JM, Brookshire ENJ, Cramer MD, Hasselquist NJ, Koba K, Marin-Spiotta E, Wang L. 2015. Ecological interpretations of nitrogen isotope ratios of terrestrial plants and soils. Plant and Soil 396: 1–26.
    OpenUrl
  12. ↵
    Díaz S, Kattge J, Cornelissen JHC, Wright IJ, Lavorel S, Dray S, Reu B, Kleyer M, Wirth C, Colin Prentice I, et al. 2016. The global spectrum of plant form and function. Nature 529: 167–171.
    OpenUrlCrossRefPubMed
  13. ↵
    Dong N, Prentice IC, Wright IJ, Evans BJ, Togashi HF, Caddy-retalic S, Mcinerney FA, Sparrow B, Leitch E, Lowe AJ. 2020. Components of leaf-trait variation along environmental gradients. New Phytologist 228: 82–94.
    OpenUrl
  14. ↵
    Dray S, Legendre P. 2008. Testing the species traits-environment relationships: the fourth-corner problem revisited. Ecology 89: 3400–3412.
    OpenUrlCrossRefPubMedWeb of Science
  15. ↵
    Dubuisson J-Y, Schneider H, Hennequin S. 2009. Epiphytism in ferns: diversity and history. Comptes Rendus Biologies 332: 120–128.
    OpenUrlCrossRefPubMedWeb of Science
  16. ↵
    Farquhar GD, O’Leary MH, Berry JA. 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Australian Journal of Plant Physiology 9: 121–137.
    OpenUrlCrossRefWeb of Science
  17. Féret J-B, le Maire G, Jay S, Berveiller D, Bendoula R, Hmimina G, Cheraiet A, Oliveira JC, Ponzoni FJ, Solanki T, et al. 2019. Estimating leaf mass per area and equivalent water thickness based on leaf optical properties: Potential and limitations of physical modeling and machine learning. Remote Sensing of Environment 231: 110959.
    OpenUrl
  18. ↵
    Garcés Cea M, Claverol S, Castillo CA, Rabert Pinilla C, Bravo Ramírez L, Cea MG, Claverol S, Castillo CA, Pinilla CR, Ramírez LB. 2014. Desiccation tolerance of Hymenophyllacea filmy ferns is mediated by constitutive and non-inducible cellular mechanisms. Comptes Rendus Biologies 337: 235–243.
    OpenUrlCrossRef
  19. ↵
    Helsen K, Van Cleemput E, Bassi L, Graae BJ, Somers B, Blonder B, Honnay O. 2020. InterD and intraspecific trait variation shape multidimensional trait overlap between two plant invaders and the invaded communities. Oikos 129: 677–688.
    OpenUrl
  20. ↵
    Helsen K, Shen Y-C, Lin T-Y, Chen C-F, Huang C-M, Li C-F, Zelený D. 2021. Do forest over-and understory respond to the same environmental variables when viewed at the taxonomic and trait level? bioRxiv.
  21. ↵
    Helsen K, Smith SW, Brunet J, Cousins SAO, De Frenne P, Kimberley A, Kolb A, Lenoir J, Shiyu MA, Michaelis J, et al. 2018. Impact of an invasive alien plant on litter decomposition along a latitudinal gradient. Ecosphere 9: e02097.
    OpenUrl
  22. ↵
    Hietz P, Briones O. 1998. Correlation between water relations and within-canopy distribution of epiphytic ferns in a Mexican cloud forest. Oecologia 114: 305–316.
    OpenUrlCrossRefWeb of Science
  23. ↵
    Hodgson JG, Montserrat-Martí G, Charles M, Jones G, Wilson P, Shipley B, Sharafi M, Cerabolini BEL, Cornelissen JHC, Band SR, et al. 2011. Is leaf dry matter content a better predictor of soil fertility than specific leaf area? Annals of Botany 108: 1337–1345.
    OpenUrlCrossRefPubMed
  24. ↵
    Karst AL, Lechowicz MJ. 2007. Are correlations among foliar traits in ferns consistent with those in the seed plants? New Phytologist 173: 306–312.
    OpenUrlCrossRefPubMedWeb of Science
  25. ↵
    Kessler M, Karger DN, Kluge J. 2016. Elevational diversity patterns as an example for evolutionary and ecological dynamics in ferns and lycophytes. Journal of Systematics and Evolution 54: 617–625.
    OpenUrl
  26. ↵
    Kessler M, Siorak Y, Wunderlich M, Wegner C. 2007. Patterns of morphological leaf traits among pteridophytes along humidity and temperature gradients in the Bolivian Andes. Functional Plant Biology 34: 963–971.
    OpenUrl
  27. ↵
    Kluge J, Kessler M. 2007. Morphological characteristics of fern assemblages along an elevational gradient: patterns and causes. Ecotropica 13: 27–43.
    OpenUrl
  28. ↵
    Kraft NJB, Adler PB, Godoy O, James EC, Fuller S, Levine JM. 2015. Community assembly, coexistence and the environmental filtering metaphor. Functional Ecology 29: 592–599.
    OpenUrlCrossRef
  29. ↵
    Lavorel S, Garnier E. 2002. Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology 16: 545–556.
    OpenUrlCrossRefWeb of Science
  30. ↵
    Legendre P, Galzin R, Harmelin-Vivien ML. 1997. Relating behavior to habitat: solutions to the fourth-corner problem. Ecology 78: 547–562.
    OpenUrlWeb of Science
  31. ↵
    Li CF, Chytrý M, Zelený D, Chen MY, Chen TY, Chiou CR, Hsia YJ, Liu HY, Yang SZ, Yeh CL, et al. 2013. Classification of Taiwan forest vegetation. Applied Vegetation Science 16: 698–719.
    OpenUrl
  32. ↵
    Lin K-C, Chen Z-S, Chang J-M. 1988. Morphology, physicochemical properties, and classification of Lalashan podzolic soils of Northern Taiwan. Journal of the Chinese Agricultural Chemical Society 26: 503–514.
    OpenUrl
  33. ↵
    Lin D, Yang S, Dou P, Wang H, Wang F, Qian S, Yang G, Zhao L, Yang Y, Fanin N. 2020. A plant economics spectrum of litter decomposition among coexisting fern species in a sub-tropical forest. Annals of Botany 125: 145–155.
    OpenUrl
  34. ↵
    Mantovani A. 1999. A method to improve leaf succulence quantification. Brazilian Archives of Biology and Technology 42: 9–14.
    OpenUrlWeb of Science
  35. ↵
    Maréchaux I, Saint□André L, Bartlett MK, Sack L, Chave J. 2020. Leaf drought tolerance cannot be inferred from classic leaf traits in a tropical rainforest. Journal of Ecology 108: 1030–1045.
    OpenUrl
  36. ↵
    McCune B, Keon D. 2002. Equations for potential annual direct incident radiation and heat load. Journal of Vegetation Science 13: 603–606.
    OpenUrlCrossRefWeb of Science
  37. ↵
    Medeiros CD, Scoffoni C, John GP, Bartlett MK, Inman-Narahari F, Ostertag R, Cordell S, Giardina C, Sack L. 2019. An extensive suite of functional traits distinguishes Hawaiian wet and dry forests and enables prediction of species vital rates. Functional Ecology 33: 712–734.
    OpenUrl
  38. ↵
    Nitta JH, Watkins Jr JE, Davis CC. 2020. Life in the canopy: community trait assessments reveal substantial functional diversity among fern epiphytes. New Phytologist 227: 1885– 1899.
    OpenUrlCrossRef
  39. ↵
    Oksanen AJ, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, Hara RBO, Simpson GL, Solymos P, et al. 2017. vegan: community ecology package. R package version 2.4-2.
  40. ↵
    Ordoñez JC, van Bodegom PM, Witte J-PM, Wright IJ, Reich PB, Aerts R. 2009. A global study of relationships between leaf traits, climate and soil measures of nutrient fertility. Global Ecology and Biogeography 18: 137–149.
    OpenUrl
  41. ↵
    Peppe DJ, Lemons CR, Royer DL, Wing SL, Wright IJ, Lusk CH, Rhoden CH. 2014. Biomechanical and leaf-climate relationships: A comparison of ferns and seed plants. American Journal of Botany 101: 338–347.
    OpenUrlAbstract/FREE Full Text
  42. ↵
    Peres-Neto PR, Dray S, Braak CJF te. 2017. Linking trait variation to the environment: critical issues with community-weighted mean correlation resolved by the fourth-corner approach. Ecography 40: 806–816.
    OpenUrl
  43. ↵
    Pérez-Harguindeguy N, Diaz S, Garnier E, Lavorel S, Poorter H, Jaureguiberry P, Bret-Harte MS, Cornwell WK, Craine JM, Gurvich DE, et al. 2013. New handbook for standardised measurement of plant functional traits worldwide. Australian Journal of Botany 61: 167–234.
    OpenUrlCrossRefWeb of Science
  44. ↵
    Petter G, Wagner K, Wanek W, Sánchez Delgado EJ, Zotz G, Cabral JS, Kreft H. 2016. Functional leaf traits of vascular epiphytes: vertical trends within the forest, intra- and interspecific trait variability, and taxonomic signals. Functional Ecology 30: 188–198.
    OpenUrl
  45. ↵
    Salazar L, Homeier J, Leuschner C, Kessler M, Kluge J. 2012. Altitudinal change in biomass, productivity and leaf functional traits in the Ecuadorian Andes: Comparing terrestrial ferns with trees. In: Unraveling the causal links between ecosystem productivity measures and species richness using terrestrial ferns in Ecuador. PhD thesis, 89–111.
  46. ↵
    Schellenberger Costa D, Zotz G, Hemp A, Kleyer M. 2018. Trait patterns of epiphytes compared to other plant life-forms along a tropical elevation gradient. Functional Ecology 32: 2073–2084.
    OpenUrl
  47. ↵
    Schulz HM, Li C-F, Thies B, Chang S-C, Bendix J. 2017. Mapping the montane cloud forest of Taiwan using 12 year MODIS-derived ground fog frequency data. PLOS ONE 12: e0172663.
    OpenUrl
  48. ↵
    Sessa EB, Givnish TJ. 2014. Leaf form and photosynthetic physiology of Dryopteris species distributed along light gradients in eastern North America. Functional Ecology 28: 108–123.
    OpenUrlCrossRef
  49. ↵
    Smith AR, Pryer KM, Schuettpelz E, Korall P, Schneider H, Wolf PG. 2006. A classification for extant ferns. Taxon 55: 705–731.
    OpenUrlCrossRefWeb of Science
  50. ↵
    Tosens T, Nishida K, Gago J, Coopman RE, Cabrera HM, Carriquí M, Laanisto L, Morales L, Nadal M, Rojas R, et al. 2016. The photosynthetic capacity in 35 ferns and fern allies: mesophyll CO2 diffusion as a key trait. New Phytologist 209: 1576–1590.
    OpenUrl
  51. ↵
    Violle C, Navas MML, Vile D, Kazakou E, Fortunel C, Hummel I, Garnier E. 2007. Let the concept of trait be functional! Oikos 116: 882–893.
    OpenUrlCrossRefPubMedWeb of Science
  52. ↵
    Watkins Jr JE, Rundel PW, Cardelús CL. 2007. The influence of life form on carbon and nitrogen relationships in tropical rainforest ferns. Oecologia 153: 225–232.
    OpenUrlCrossRefPubMed
  53. ↵
    Wegner C, Wunderlich M, Kessler M, Schawe M. 2003. Foliar C:N ratio of ferns along an Andean elevational gradient. Biotropica 35: 486–490.
    OpenUrl
  54. ↵
    Wilson PJ, Thompson K, Hodgson JG. 1999. Specific leaf area and leaf dry matter content as alternative predictors of plant strategies. New Phytologist 143: 155–162.
    OpenUrlCrossRefWeb of Science
  55. ↵
    Wright IJ, Cannon K. 2001. Relationships between leaf lifespan and structural defences in a low-nutrient, sclerophyll flora. Functional Ecology 15: 351–359.
    OpenUrlCrossRefWeb of Science
  56. ↵
    Wright IJ, Reich PB, Cornelissen JHC, Falster DS, Groom PK, Hikosaka K, Lee W, Lusk CH, Niinemets Ü, Oleksyn J, et al. 2005. Modulation of leaf economic traits and trait relationships by climate. Global Ecology and Biogeography 14: 411–421.
    OpenUrl
  57. ↵
    Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender-Bares J, Chapin T, Cornelissen JHC, Diemer M, et al. 2004. The worldwide leaf economics spectrum. Nature 428: 821–827.
    OpenUrlCrossRefPubMedWeb of Science
  58. ↵
    Zelený D. 2018. Which results of the standard test for community-weighted mean approach are too optimistic? Journal of Vegetation Science 29: 953–966.
    OpenUrl
  59. ↵
    Zhang S-B, Sun M, Cao K-F, Hu H, Zhang J-L. 2014. Leaf photosynthetic rate of tropical ferns is evolutionarily linked to water transport capacity. PLOS ONE 9: e84682.
    OpenUrlCrossRefPubMed
  60. ↵
    Zuur AF, Ieno EN, Elphick CS. 2010. A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution 1: 3–14.
    OpenUrlCrossRef
Back to top
Posted September 06, 2021.
Download PDF

Supplementary Material

Email
Comparing trait syndromes between Taiwanese subtropical terrestrial and epiphytic ferns and lycophytes at the species and community level
Kenny Helsen, Tsung-Yi Lin, David Zelený
bioRxiv 2021.09.06.459074; doi: https://doi.org/10.1101/2021.09.06.459074
Reddit logo Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area