Are species interactions important drivers of community dynamic stability?
-
1
Rio de Janeiro State University, Brazil
-
2
Federal University of Rio de Janeiro, Brazil
-
3
National Museum of Natural Sciences (MNCN), Spain
Land cover is a major driver of the structure of stream-dwelling fish assemblages both in temperate and tropical regions. Nonetheless, long-term patterns and subjacent ecological mechanisms that modulate temporal variation in pristine and anthropized areas are still largely unknown. Theory suggests that interspecific interactions and community structure have major effects on community dynamics. Experiments show that these attributes are not static in time and their variation have strong effects on local communities, which lacks a large empirical support from non-controlled systems. Therefore, we compared the community temporal dynamics of a forested and an unforested sites in a Neotropical micro-basin.
We carried out bimonthly to semesterly samples in two different sites in the Ubatiba River basin between July, 1994 to August, 2001 (n = 29). The Ubatiba River basin is a small coastal drainage in the Atlantic Forest. Local water level is mainly regulated by rainfall and run-off, and, to some extent, it exhibits a marked seasonality. Selected areas exhibit strong differences in riparian vegetation: Site 1 (S1) is covered by a secondary Pluvial Forest and site 2 has no vegetal cover since it is located in an extensive pasture area (S2). Samplings were carried out using electrofishing (AC – 100W, 2-3 A) with constant effort in all fishing sites and seasons.
First, we quantified fluctuating interspecific interactions using empirical dynamic modelling, which merges statistical tools that reconstruct nonlinear dynamic systems based on time-series data. We applied Convergent cross mapping (CCM) to determine causal relationships among fish species in each selected area. CCM allows the modelling of nonlinear dynamical systems by representing the system dynamics in a multivariate state space using time lags of observable variables as substitutes for unknown variables. If two variables are part of the same dynamical system, reconstructed state spaces of these variables will represent the same attractor and we may test the correspondence between them (i.e. cross mapping).
We selected fishes that occurred in ≥90% of sampling events since rarer species exhibit a high number of zero values in data matrix and are unsuitable for time-series analysis. We used abundance data of 6 fish species, which was further normalized to unit mean and variance. CCM is also sensitive to the choice of embedding dimension (E, i.e. the number of time-lag coordinates used in the state space reconstruction). We determined E evaluating predictability of different values of E varying from 1 to 6, which allows the influence of abundance values up to 1 or 3-years lag. We applied simplex projection to perform cross mapping and evaluated its predictability with a correlation coefficient (ƿ). Predictability increases with library length (i.e. number of points in the reconstructed state space of an effect variable) if two variables are causally coupled, i.e. the higher the input of data on the abundance of species x, the higher the predictability of the abundance of species y. The variation in interactions strength were calculated by a multivariate S-map, which are based on sequential Jacobian matrices.
Second, we estimate the community dynamic stability by the dominant eigenvalue of the time-varying interaction matrix (i.e. local Lyapunov stability). If we consider the matrix A (produced by the S-map) describing the fluctuating interactions in a given community, its eigenvalue will describe the rate of separation of causally related species’ abundances along time. The absolute values of the real part of the dominant eigenvalue of A will be related to the predictability of the system, more precisely to the return to a specific ecosystem state (or attractor) if interspecific strengths and self-regulation effects are kept constant. Dynamic stability values lower than 1 indicate faster recovery to initial states. We further calculated two network measures (mean interaction strength and weak interaction index) and one diversity measure (Simpson’s diversity) and applied CCM in order to estimate the predictability of dynamic stability by network and diversity measures.
We identified 3 and 4 interspecific interactions in pasture (S2) and forested (S1) sites, respectively. In the former, we identified two strong positive interactions and a weaker negative interaction, while in the latter, we identified two strong negative interactions and two weaker positive interactions. These interactions are not static in time. Generally, interaction strength between a pair of species vary over time, but maintains the interaction signal (positive or negative); but in some cases, the interaction strength varies around zero and eventually changes signal. In the forested (S1) site, for example, the interaction between two species of Astyanax is consistently negative, but in some periods are more negative than in other ones. In contrast, the interaction between the characids Astyanax hastatus and Mimagoniates microlepis is generally positive, but sometimes achieves negative values. These results highlight the context-dependency and temporal variability of interspecific interactions, which may have important consequences on community structure and dynamics.
Dynamics stability values were generally lower than 1 for both forested (S1) and pasture (S2) sites (Figure 1a and 2a), but the forested (S1) site exhibited values higher than 1 in some periods. This comparison indicates that the pasture (S2) exhibits a slightly higher potential to return to an ecosystem state than forested sites. Nevertheless, predictability of the dynamic stability was higher in the forested site. In the pasture, only weak interactions explained significantly dynamic stability (Figure 2b-d), but had a small importance predicting dynamic stability. In this case, the higher the dominance of weak interactions (lower values of the index), the lower the dynamic stability (higher values). In the forested (S1) site, both mean interaction strength and dominance of weak interactions were major drivers of dynamic stability (Figure 1b-d). Therefore, our data supports that community dynamics of forested sites are more predictable than pasture dynamics, and that weak interactions are stabilizing factors for both communities. Nevertheless, strong interactions have a large stabilizing role in the forested site. Loss of riparian vegetation causes strong changes in aquatic ecosystems, which may pose different environmental filters as main drivers of community dynamics instead of ecological interactions.
Acknowledgements
Cnpq - 301463/2017-4 to RM
FAPERJ - E-26/203.193/2015 to RM; E-26/201.583/2018 to RIR and E-26/200.758/2019 to BES
References
Santos FB, Ferreira FC, Esteves KE. 2015. Assessing the importance of the riparian zone for stream fish communities in a sugarcane dominated landscape (Piracicaba River Basin, Southeast Brazil). Environ Biol Fish
DOI 10.1007/s10641-015-0406-4.
Clark AT, Ye H, Isbell F, Deyler ER, Cowles J, Tilman GD, Sugihara G. 2015. Spatial convergent cross mapping to detect causal relationships from short time series Ecology, 96(5): 1174–1181.
McGowan JA, Deyle ER, Ye H, Carter ML, Perretti CT, Seger KD, Verneil A, Sugihara G. 2018. Predicting coastal algal blooms in southern California. Ecology, 98(5): 1419–1433.
Munch SB, Giron-Nava A, Sugihara G. 2018. Nonlinear dynamics and noise in fisheries recruitment: A global meta-analysis. Fish and Fisheries. 19: 964–973.
Keywords:
stream-dwelling fish,
Pluvial forest,
Neotropics,
Persistence,
Canopy
Conference:
XVI European Congress of Ichthyology, Lausanne, Switzerland, 2 Sep - 6 Sep, 2019.
Presentation Type:
Oral
Topic:
ECOLOGY AND LIFE CYCLES
Citation:
Mazzoni
R,
Caramaschi
EP,
Lobón-Cervi'A
J,
Iglesias-Rios
R and
Soares
BE
(2019). Are species interactions important drivers of community dynamic stability?.
Front. Mar. Sci.
Conference Abstract:
XVI European Congress of Ichthyology.
doi: 10.3389/conf.fmars.2019.07.00024
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
20 Jul 2019;
Published Online:
14 Aug 2019.
*
Correspondence:
Prof. Rosana Mazzoni, Rio de Janeiro State University, Rio de Janeiro, Brazil, mazzoni@uerj.br