The Herbaceous Understory Plant Community in the Context of the Overstory: An Overlooked Component of Tropical Diversity

: Lowland tropical rainforests harbor the most diverse plant communities in the world, but the herbaceous plants of the understory are often overlooked. To address this knowledge gap, we asked to what extent the understory herbaceous community contributes to the species richness and phylogenetic diversity of plant communities by surveying a neotropical rainforest at Los Tuxtlas, Mexico. We used Gentry transects to characterize the woody overstory community, and line-intercepts within the same transects to survey understory herbs and subshrubs. We also used published phylogenies to calculate community phylogenetic diversity with and without the understory stratum. We found that the understory contained a diverse (23 species, or 22.1% of all species surveyed) and phylogenetically distinct plant community dominated by aroids (13 species) and ferns (4 species). Inclusion of the understory stratum increased total species richness by 28.4% but increased phylogenetic diversity by 41.4%. Additionally, in contrast to temperate forests, the understory plant community was much less diverse than the overstory, which contained 81 species > 1 cm dbh (77.9% of all species surveyed). This survey adds to the hitherto small body of literature comparing understory and overstory strata in tropical rainforests and reveals previously overlooked patterns of floristic diversity.


Introduction
The diversity of tropical rainforest plant communities is unsurpassed by that of any other known terrestrial biome [1][2][3].Given the superlative nature of tropical plant diversity, attempts to explain the origins and maintenance of such diversity have been a central focus of ecological research [4,5].In documenting tropical diversity, substantial efforts have been made to conduct thorough surveys of plant communities for this purpose, especially since the advent of the 50-ha plot network coordinated by the Center for Tropical Forest Science [6].However, these and other comprehensive surveys typically overlook "small" plants, opting to use 1 cm diameter at breast height (dbh) in the case of CTFS, or 10 cm dbh in other surveys [7] as the minimum size cutoff for inclusion in forest censuses.In so doing, such surveys omit, in addition to the epiphytic species, the understory plant community, including terrestrial herbs, subshrubs, saplings and seedlings that comprise the "understory (sometimes referred to as the herbaceous) layer".
In temperate forests, herbaceous plant diversity typically exceeds that of sympatric tree communities [8].In tropical forests, the few available studies on floristic composition considering regional check lists (i.e., presence-absence data) suggest that herbs in the understory layer comprise a substantial component of the total local species diversity [9-12], but do not reach a proportional diversity similar to that of their temperate counterparts.However, species lists are simple measures that often provide useful information, but limited insight into forest ecology.To understand the ecology of the forest understories, quantitative assessments of plant community composition with ecological metrics of the species (abundance, frequency) comparing the different vegetation strata are necessary, particularly in tropical systems [13].
The use of standardized, spatially explicit sampling designs of woody plant communities such as "Gentry transects" [1,14], and high-resolution forest dynamics plots [6], have facilitated fine-grain characterization of overstory community composition in tropical forests.However, analogous methods have seldom been employed for understory plants in tropical rain forests (see [15,16]).Even more rare are studies simultaneously including this information on both the understory and overstory layers at the same site.
Exceptional Neotropical examples include: Gentry and Dodson [11] and Linares-Palomino et al. [17], both examining three forest types: wet evergreen, moist semi-deciduous, and dry deciduous in Ecuador and Bolivia respectively.Vázquez and Givnish [18], on the other hand, examined the herbaceous layer along an altitudinal transect within a dry deciduous forest in western Mexico.Collectively, these studies show that: (1) terrestrial herbaceous communities can be very diverse (proportional diversity ranged from 12% to 50% of total plant diversity), but do not reach the proportional diversity commonly found in temperate forests, and (2) diversity is highly variable, with no consistent pattern between forest types.A remarkable study [12] compared the richness (morpho-species) and composition of herbaceous assemblages across three continents, including family diversity, with an emphasis on the relationships between diversity and abiotic factors.Clearly, more quantitative studies are needed to assess the patterns of understory diversity in tropical forests.Furthermore, the contribution of the understory layer to the overall phylogenetic and functional diversity of tropical plant communities remains largely unknown.Given the eco-evolutionary and functional significance of phylogenetic diversity [19], it represents another important facet of tropical biodiversity that is critical to consider.
Here we address these lacunae by examining the woody overstory (encompassing all woody plants with dbh > 1.0 cm) and the herbaceous understory community in a Neotropical Forest.In this study, we present results from a systematic survey of both the overstory and understory plant communities of a lowland tropical rain forest, Los Tuxtlas, in Veracruz, Mexico.Using a combination of Gentry transects (for the overstory layer) and line intercept transects (for the understory), we test the prediction, based on the pattern insinuated by the floristic studies that include understory herbs [9][10][11], that understory diversity will be lower than that of the overstory.Furthermore, we leverage published phylogenetic information to examine the contribution of the understory stratum to overall community phylogenetic diversity.This type of systematic survey of the understory is critical, particularly in light of the current pulse of Anthropogenic impact on native forests, including the prevalence of defaunation in tropical understories, given that the absence of such animals (the "Empty Forest Syndrome" [20]) has a considerable potential to alter the diversity, structure and composition of understory communities.

Study Site
This study was conducted at Los Tuxtlas Tropical Biology Station administered by the Institute of Biology of the National Autonomous University of Mexico (UNAM).Los Tuxtlas is located in southeastern Veracruz on the Gulf Coast (18 • N, 95 • W, 47 m a.s.l.) and hosts the northernmost extent of lowland evergreen tropical rainforest on the continental Americas [21].The biological station is located within the Los Tuxtlas Biosphere Reserve and lies on the northeastern slope of the San Martin Volcano.The soils are volcanic and of relatively recent origin.The site receives more than 4 m of precipitation annually.A detailed description of the site can be found in González-Soriano et al. [22], and an analysis of the floristic composition and forest structure can be found in Bongers et al. [23] and Ibarra-Manríquez et al. [24].

Data Collection
To characterize the overstory and understory plant communities, we established 10 linear transects (50 m × 2 m each) in August 2014.Starting points for the transects were chosen by selecting two random points at least 100 m apart along each of five main trails within the preserve.The transects were laid out either perpendicular to the trail, or in cases of relatively steep topography, in the direction following the contour of the slope.The census of the overstory was performed using a modified Gentry protocol [1,14]: the diameter of each stem rooted within the transect, and with ≥1 cm dbh was measured, its position along the 50 m axis recorded, and the species identified.For climbing plants, basal diameter was used instead of dbh.A line intercept method, with a 50 m measuring tape, was used to approximate the relative abundance and distribution of understory species within each transect.For these species, we recorded the length (cm on the transect line) at which a terrestrial herbaceous plant intercepted the central 50 m transect line, and the species identified.Accurate identification to species was obtained for all individuals encountered in the census for both strata.Plant nomenclature follows Ibarra-Manríquez et al. [24].We recorded the number of intercepts (frequency) for each herbaceous species in each transect to calculate the relative frequencies.

Statistical Analyses
First, to establish whether sampling was complete, we constructed species accumulation curves for each transect and for both overstory and understory plants, separately.To determine whether sampling effort differed between the two strata, sample sums per transect were compared using a Mann-Whitney test [25].For each overstory species, we calculated stem density (total number of stems), total basal area (derived from dbh of all individuals), and the relative frequency of its presence (proportion of transects in which a species occurred).These measures were relativized, and their sum constitutes the Importance Value Index (IVI) [26].As in Kammesheidt [26], an index equivalent to the IVI was calculated for the plants for which basal area was not calculated (i.e., the understory plants in the present study).This analogue was calculated as the sum of the relative density (total number of intercepts), and relative frequency.The species' importance values were used to construct rank abundance curves for plants from both strata.
A time-calibrated community phylogeny was constructed using previously published phylogenetic data from Zanne et al. [27] for angiosperms and Rothfels et al. [28] for ferns.Phylogenetic diversity of the plant community was measured as the sum of branch lengths [29] and was calculated with and without the understory stratum.All statistics were computed using R (version 3.0.2) statistical computing language (www.r-project.org), and the Vegan package for R software [30].

The fifteen most important overstory and understory species arranged by decreasing order of IVI.
The phylogenetic tree of the studied plant community (Figure 3) portrays the qualitative descriptions described above, with three major clusters in the understory corresponding, from top to bottom, to ferns, aroids, and dicotyledonous herbs.The overstory highlights the prevalence of palms, and many taxa encompassing a multitude of lineages.
Total community phylogenetic branch length was 6191.8 without the understory stratum, and 8753.6 with the understory stratum (Figure 3).Thus, inclusion of understory plants resulted in a 41.4% increase in observed phylogenetic diversity.
The species accumulation curves based on each transect (Figure S1), show that the understory and overstory communities are not completely sampled according to the minimal area concept [31,32].This is typical of highly-diverse systems where reaching an asymptote in the species-area curve (i.e., a minimal area) would involve unfeasible sampling effort.However, we deployed roughly similar understory (56 ± 8.4 intercepts per transect [mean ± SE]) and overstory plant sampling efforts (43.7 ± 5.4 stems per transect [mean ± SE]), which represents an equivalent, statistically indistinguishable effort for both strata (Mann-Whitney U = 38.5, p = 0.41, Figure S1).

Figure 3.
Phylogeny of all species encountered in the study, derived from Zanne et al. [27] for angiosperms, and Rothfels et al. [28] for ferns.See text for details.[27] for angiosperms, and Rothfels et al. [28] for ferns.See text for details.

Discussion
We found that the contingent of plant species comprising the understory layer of a tropical rainforest contributes a substantial portion of total plant species richness and phylogenetic diversity.Inclusion of the understory layer in the present study resulted in a 28.4% increase in total plant species richness, and a 41.4% increase in total phylogenetic diversity.Given the common practice of omitting stems below 1 cm dbh in tropical plant surveys, this study confirms that such studies overlook an important component of forest biodiversity.Not only is the understory community diverse, but also phylogenetically distinct.Araceae, while absent from the overstory survey, contains 13 of the 23 (56.5%) understory species.It is worth noting that many Araceae, though core components of the understory layer, complete their life cycle as climbing plants, underscoring the importance of the overstory for tropical herbaceous plants.Pteridophyta also represents an important and unique component of the understory not represented in the overstory (except as epiphytic ferns).Indeed, although so called "tree ferns"(Cyatheaceae) are present in the Los Tuxtlas forests, they are more prominent in mid-elevation sites (e.g., cloud forests) and rather rare in the lowlands of this and other tropical regions [24].Our results also suggest that, at lower latitudes, herbaceous communities are typically less closely-related than woody communities, an observation in line with that of Massante et al. [13], who argue that most woody lineages evolved in the tropics, whereas herbaceous lineages mostly evolved in young habitat types at higher latitudes.
Increasing the completeness of plant surveys is an important goal aided by the inclusion of the understory stratum.Additionally, accounting for composition and dynamics of the understory stratum could be critical to understanding processes driving community composition in other strata.Herbaceous plants, including Pteridophytes and Araceae, have proven important in influencing forest dynamics in temperate regions by altering conditions for regenerating woody overstory species via shading [33].The strength of these negative effects has been shown to be species-specific, potentially influencing forest overstory composition [34].However, little has been done to elucidate the effect of understory vegetation on recruitment of larger plants to the overstory in tropical forests.This is despite the fact that many important studies focus on factors affecting seedling survival (i.e., negative density dependence), as seedling performance and filtering is presumed to be especially important in determining forest community composition at later ontogenetic stages [35][36][37].In addition, many understory plants (shrubby and herbaceous) are a crucial component of the functional diversity of the tropical rainforest, as many of them are important resources for plant consumers, including herbivores [38], pollinators [39], leaf pathogens [40] and dispersal agents [41].
Beyond the biodiversity perspective, the understory layer may be particularly relevant to conservation due to its susceptibility to disturbance and subsequent repercussions for the overstory [42].Historically, the community composition of the understory has been found to be more dynamic than that of the overstory, reflecting the relative sensitivity of these communities.For example, using the rate of species turnover in the fossil record, Levin and Wilson [43] estimated that the background extinction rate of herbaceous plants is three times that of woody species.In modern temperate forests, scientists have underscored the sensitivity of herbaceous understory plants to disturbance, and propose to use the understory stratum as an indicator of forest conservation value [44].According to Spice et al. [45], some characteristics of herbaceous plants, such as limited dispersal and shorter stature make forest herb communities more vulnerable than trees to most anthropogenic impacts such as land-use change, overabundant herbivores, invasive species, and climate change, which are strongly affecting forest herb communities [45].
Understory plants are particularly susceptible to changes in the terrestrial fauna [46].Terrestrial vertebrates often interact strongly with understory flora as dispersers, herbivores, and as agents of mechanical disturbance.Large terrestrial seed predators and herbivores are also frequently targeted as sources of bushmeat.Therefore, defaunation could have a disproportionate effect on understory plants in tropical forests.Indeed, Dirzo and Miranda [38] suggest that the understory vegetation at Los Tuxtlas may be highly affected by the recent local extinction of many of the large herbivores (see also [47]).Understory mammal exclosure studies in another Neotropical forest, Barro Colorado Island, reveal that defaunation may increase the abundance of rare understory herbs [48].However, a robust body of research has yet to emerge detailing the effects of defaunation and other current anthropogenic changes on tropical forest understory plant communities and its possible cascading consequences for ecosystem structure and functioning.
Given the disproportionate impact of current defaunation on the understory flora [38,46] and a historical trend of high turnover in herbaceous plant community composition [43], it is important for more tropical plant surveys to include the understory stratum.Perhaps more important is the study of interactions between resident understory plants and their counterparts in the overstory, including uni-and bi-directional processes of facilitation or competition.In combination, these research objectives promise to bolster our understanding of the potential impacts of global change on tropical plant community biodiversity.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/d14100800/s1,Table S1: Raw data on overstory and understory transects at Los Txutlas, Veracruz, Mexico.Figure S1: (A) Species accumulation curve for overstory stratum.(B) Species accumulation curve for understory stratum.Each curve shows the accumulation of species along five segments (every 10 m) of the transect.

Funding:
We gratefully acknowledge the Stanford Global Studies Division at Stanford University for support of field expedition costs under a grant to RD. RP was supported with a Marie Curie fellowship from the European Union (FP7-PEOPLE-2013-IOF-627450).
Institutional Review Board Statement: Not applicable.

Figure 1 .
Figure 1.(A) Total encounters (intercepts for understory plants and stem number for ov plants, per 100 m 2 transect).(B) Species richness in understory and overstory strata.

Figure 1 .
Figure 1.(A) Total encounters (intercepts for understory plants and stem number for overstory plants, per 100 m 2 transect).(B) Species richness in understory and overstory strata.

Figure 2 .
Figure 2. The fifteen most important overstory and understory species arranged by decreasin of IVI.

Figure 3 .
Figure3.Phylogeny of all species encountered in the study, derived from Zanne et al.[27] for angiosperms, and Rothfels et al.[28] for ferns.See text for details.

Table 1 .
Understory species encountered, and their corresponding Importance Value Index (IVI).