Explaining species composition of local communities is a central theme in ecology
(Weiher and Keddy, 1995). A set of species which may potentially colonize a site is usually called the “species pool” of that site (Pärtel et al., 2011), and several mechanisms
have been proposed to filter a species pool into observed species composition of a local community. Three types of filtering mechanisms are usually mentioned, including dispersal filter, abiotic environmental filter and biotic interaction filter (Cadotte and Tucker, 2017) (Figure 2). Dispersal filter first excludes those species unable to arrive at the site due to limitations of dispersal abilities. Environmental filter excludes those species not capable of establishing and persisting in such environmental conditions (Bazzaz, 1991), and biotic interactions such as competition and predation may furthermore filter out some species (Hardy et al., 2012). Although these three types of filtering mechanisms are usually described as sequential and discrete processes, they actually interact with each other in complex ways in reality (Cadotte and Tucker, 2017).
doi: 10.6342/NTU201900489
8
Figure 2. Three types of filtering mechanisms
These three types of filtering mechanisms were proposed to explain species composition of a local community. Each symbol in this figure represents a species within species pool of the community. Dispersal filter excludes those species cannot disperse to the site, environmental filter excludes those species cannot persist in that abiotic environment, and interaction filter excludes those species cannot persist because of biotic interactions like competition. (modified from Cadotte and Tucker, 2017)
These filtering mechanisms can also be applied to several local communities distributed along an environmental gradient. In this case, environmental filter is usually considered most influential (Kraft et al., 2015). For gradual change of environmental conditions, the set of species which can pass environmental filter also changes. As a result,
doi: 10.6342/NTU201900489
9
a sequential change of species composition along the environmental gradient is expected (Laliberté et al., 2014). Besides directly filtering out those species not able to persist, the environmental gradient may also indirectly influence species composition by their effects on biotic interactions. For example, competition ability of a species usually change along the environmental gradient, and a species may be absent in a local community not because of failure to survive in that abiotic environmental condition but because of lower competition ability relative to other coexisting species (Cadotte and Tucker, 2017) (Figure 3a & 3b).
The concept of these filtering mechanisms were proposed to explain species composition of local communities, while they can be linked to the niche concept of species. Hutchinson (1957) described the fundamental niche of a species as a state of
abiotic environment which permits that species to exist, and can be viewed as an “n-dimensional hypervolume” of which each dimension represent one environmental
variable. However, he also pointed out that a species usually only utilize a subset of its
fundamental niche because of biotic interactions, and this subset was called “realized niche”. From this point of view, environmental filter actually filter out those species
whose fundamental niche do not encompass the environmental condition of the site, so environmental filter is also called a “niche-based” process (Püttker et al., 2015). The other way around, the distribution range of a species after considering the effects of
doi: 10.6342/NTU201900489
10
environmental filter reflects its fundamental niche, and the distribution range reflects its realized niche after considering the effects of both environmental and biotic interaction filter (Figure 3c).
Figure 3. Filtering mechanisms, species composition and the niche concept
(a) Population growth rate of a species can reflect its fitness or competition ability, and usually change along environmental gradients. Four symbols showed in the right
doi: 10.6342/NTU201900489
11
represent four different species. A, B and C represent three communities located along this environmental gradient. If population growth rate is less than 0 (below black dash line), that species cannot persist in that environment even if there is no competitor. (b) Species composition in three communities (A,B and C) are determined by both environmental and interaction filters. The environmental filter excludes those species with population growth rate less than 0 (e.g. yellow cross species in community A), while the interaction filter excludes those species with population growth rate larger than 0 but not large enough to compete with other species (e.g. orange square species in community A). (c) Distribution range of a species after considering environmental filter reflects its fundamental niche, while the range after considering both environmental and interaction filters reflect its realized niche. (partly modified from Cadotte and Tucker, 2017)
As mentioned in previous section, solar radiation, air humidity, temperature fluctuation and wind speed all change along vertical direction in a forest. All of these environmental variables are influential to physiology of plants. For example, net primary productivity of a leaf usually increases with the amount of photosynthetically active radiation (PAR) received until reaching saturation (Ö gren and Evans, 1993). Air humidity directly influences vapor pressure deficit between a leaf and the air, and hence has strong effects on transpiration rate (Lambers, 2008). A moderate increase in temperature usually causes an increase in primary productivity (Sage and Kubien, 2007), while usually accompanied by increasing vapor pressure deficit and transpiration rate (Lambers, 2008).
Therefore, these vertical environmental gradients are expected to have strong direct or
doi: 10.6342/NTU201900489
12
indirect effects on species composition. A sequential change of species composition along these gradients as well as a differentiation of niches occupied by each species are expected as a result.
Many previous studies have reported that different epiphyte species occupied
different height ranges on host trees just as the expectation above, and this phenomenon was called “vertical stratification” (Johansson, 1974; Nieder et al., 2000; Krömer et al.,
2007; Zotz, 2007; Parra et al., 2009). However, most of these studies were done in tropical forests, especially in Central and South America. Whether vertical stratification of epiphyte species also exists in subtropical forest in Taiwan is one of the questions that this study was aimed to answer.