Abstract
The Species-Sorting concept, one of the models developed to explain
patterns in metacommunity structure, suggests that relationships between
biological communities and environmental conditions is the basic means
of the species selection processes. A second concept is Neutral Theory,
and the idea of neutral dynamics underpinning metacommunity structure,
cannot be overlooked. The third mechanism is the Mass-Effect concept,
that focuses on the interaction between environmental condition and
neutral effects. In the present study, we partitioned fish communities
in streams between niche and neutral theory concepts, identifying the
best representation of metacommunity structure, and assessed if linear
and hydrographic distance were equivalent in the representation of
neutral processes. The result points to the importance of species
sorting mechanisms in structuring fish communities with neutral
processes best represented by the linear distances. These results are
important for the fish fauna conservation leading to three
considerations: (i) the variation of the landscape and habitat is
important for the stream fish, (ii) the natural barriers are an
important landscape component to be considered, and (iii) the artificial
barriers (dams and impoundments) need to be planned taking in account
the catchment basin as the landscape unit.
Keywords : Ichthyofauna, Connectivity, Stream network,
Mass-Effect, Species-Sorting, Environmental gradient.
INTRODUCTION
Fish stream communities favors the explicitly evaluation of current
metacommunity theories. On one side, a classical interpretation of the
relationships between environmental conditions and the composition of
fish communities is closely linked to Niche Theory, predicting that the
resources and conditions of a given environment dictate the distribution
of species over space and time (Hutchinson 1957). Studies of
relationships between species composition, environmental conditions
(e.g. water temperature, dissolved oxygen content, substrate type) and
resources (e.g. places of refuge, food) have had an important impact on
our understanding of how those systems work (Hutchinson 1957).
Considering the four general models currently used to explain
metacommunity patterns (Leibold et al. 2004), SpeciesSorting assumes
that the environment gradient is the only factor structuring the
communities (Cottenie 2005; Landeiro et al. 2011). On the other hand,
the explicitly spatial structure of riverine systems (Presley et al.
2010) and the existence of some complex models of linkage among aquatic
habitats within basin (e.g. river pulse and marginal lakes), favors the
view that dispersal within the metacommunity may account for a
signification portion of the explanation of species composition patterns
(Landeiro et al. 2011). This later explanation is usually referred as a
neutral framework (Leibold 2004). Even though both factors may interact,
we are still looking for a set of more general predictions from which it
is expected that each of the models can be better applied.
The effect of environmental conditions such as water velocity, dissolved
oxygen and water temperature successfully explain common patterns
observed on fish communities of tropical streams (Penczak et al. 1994;
Zeni and Casatti 2013). Nevertheless, the choice of environmental
descriptors and the quality of its statistical description is subject to
doubts. Often these conditions are measured as a single value (Aquino et
al. 2009; Esguícero 2011). In contrast, the use of repeated measurements
(Ferreira and Casatti 2006; Araújo and Tejerina-garro 2007; Takahashi et
al. 2013) may favor a better description of the natural stochasticity of
environmental conditions in streams. Better designs for environmental
conditions measurement may account for environmental heterogeneity (the
variation on the habitat conditions along the environmental gradient) of
the stream system, which is recognized a long time (Angermeier and
Schlosser 1989; Das et al. 2012), but frequently neglected.
Environmental heterogeneity can have important effects on biological
diversity (Vandermeer 1972; Whitaker Jr 1972), suggesting that variation
of in-stream conditions is a more important predictor of fish species
diversity (Angermeier and Schlosser 1989; Beisner et al. 2006; Das et
al. 2012), and other aquatic communities than average values (Fukaya et
al. 2014; Vieira et al. 2013). For example, the relationship between
aquatic macroinvertebrate diversity and the heterogeneity of substrates
in streams (Bond and Downes 2000; Landeiro et al. 2012) .
Dispersal-related process may also exert some challenges related to data
acquisition and interpretation. The Neutral Theory (Hubbell 2001)
proposes that all individuals of the same trophic guild, have the same
ability to compete (regardless of resources and present conditions at
the place), thus having constant growth in any environment. In such
scenario the author proposes that dispersal is the main source of
increase local community diversity at an ecological time framework.
Thus, it is expected that the spatial component (distance between
sampling sites) or geographical barriers (physical obstacles to
migration between locations) are the only mechanisms responsible for
structuring the community. In the studies of stream ecology, we observe
the frequent use of linear distance as a measurement of connectivity.
However, this distance may be not informative, since strict aquatic
organisms (such as fish) can move only through the hydrographic network.
Connectivity measures based on the stream network is beginning to be
used in the interpretation of patterns on estuarine fish communities
(Fullerton et al. 2010) and other aquatic organisms (Landeiro et al.
2012; Obolewski et al. 2009; Jacobson and Faust 2014). The main problem
in using linear distances is the comparison of streams present in
different hydrographic units. Usually, streams in the same basin (or
sub-basins) are more similar than streams in different basins even if
streams in different basins are geographically closer (linear distance)
than streams within the same basin (hydrographic distance; Landeiro et
al. 2012).
There is no simple dichotomy between communities structured by niche
processes and neutral processes in nature (Cottenie 2005), as many
natural communities are structured both by local conditions and
resources (niche-related processes), and species’ dispersion abilities
(Cottenie 2005; Thompson and Townsend 2006). The combined effects of
different processes on communities are united under the theory of
Mass-Effect (Leibold et al. 2004), which describes meta-community
processes as the product of both environmental gradients and the
dispersal ability of the species. The quantification of the relative
importance of niche and neutral processes can be evaluated using
variance partitioning techniques (Cottenie 2005; Beisner et al. 2006;
Peres-Neto and Legendre). These techniques allow to quantifying the
percentage explanation related only to environmental conditions and
resources, only to space, and to the interaction between the two sets
(Legendre and Legendre 2003). Obviously, the choice of descriptive
variables for both sets – environment and space – may have a
determinant effect on the results and on the interpretation of how both
explanations interact.
The evaluation of how space and environmental variables are related to
community composition allow to a more explicit analysis of the now
classic mechanisms for metacommunity patters developed by Leibold et al.
(2004). Communities that relate solely to the environmental gradient
would be structured by the Species Sorting mechanism, whereas
communities that relate only to space are structured by Neutral or Patch
Dynamics, and the interaction between space and the environmental
gradient is taken as communities where Mass-Effect processes are
predominant (Leibold et al. 2004; Cottenie 2005). Previous study have
suggested that the metacommunity considering all species pool, or
considering only detritivores, insectivores, or omnivores was structured
by a nested co-occurrence pattern with hyperdispersed species loss and a
mass-effect mechanism. However, core species (species pool without the
rare species) displayed a Clementsian pattern and were structured by a
Species Sorting mechanism (Vieira et al. 2020). So, our goals were: (i)
identify the best representation of local conditions (average values or
variance of conditions); (ii) evaluate the importance of linear and
hydrographic distance in neutral processes; and (iii) evaluate the
relative explanatory power of environmental and space variables to
determine which of the four models proposed by Leibold et al. (2004)
account for observed patterns in Brazilian Cerrado fish communities.