European Ocean Biodiversity Information System

[ report an error in this record ]basket (0): add | show Print this page

Eco-evolutionary feedbacks in self-organized ecosystems
de Jager, M. (2015). Eco-evolutionary feedbacks in self-organized ecosystems. PhD Thesis. RUG: Groningen. ISBN 978-90-367-8153-4. 199 pp. https://hdl.handle.net/11370/4a1316a4-c764-4fae-8542-0e4776dc1fb2

Thesis info:

Available in  Author 

Author  Top 
  • de Jager, M.

Abstract
    Spatial patterns in natural systems may appear amazingly complex. Yet, they can often be explained by a few simple rules. In self-organized ecosystems, complex spatial patterns at the ecosystem scale arise as the consequence of actions of and interactions between organisms at a local scale. Aggregation into large-scale patterns may, however, also affect the survival and fitness of individuals within the ecosystem. As a consequence, pattern-producing behaviour in turn may have evolved as an adaptation to this self-generated environment in what is called an eco-evolutionary feedback process. Strikingly, both empirical and theoretical studies on eco-evolutionary feedbacks in self-organized ecosystems are rare. In this dissertation, I investigated the interplay between the ecological process of pattern formation and the evolution of two patterning-related traits: movement and attachment.I investigated the interplay between the evolution of self-organizing behaviour and the emergent large-scale patterns by performing both ecological experiments and eco-evolutionary computer simulations. For this purpose, I used mussel beds as my main model system. On intertidal sandbanks, young mussels move into labyrinth-like patterns after settlement. Mussels need sufficient neighbours in close proximity to decrease the risk of being dislodged by wave stress or predation. To accomplish this, mussels attach a glue-like substance called byssus threads to other individuals, and form dense clumps. However, gaps in between dense mussel clumps are needed to reduce competition for suspended algae. Because competition occurs over a larger range than attachment, self-organized patterns emerge in the mussel bed in the form of regularly spaced, labyrinth-like strings. The formation of labyrinth-shaped patterns increases the within-clump density of mussels while keeping the long-range mussel density low enough to prevent food competition. Two behavioural traits are mainly responsible for self-organization in mussel beds: movement and attachment. Without movement, mussels cannot search for conspecifics to aggregate with; without attachment (in the form of byssal threads), self-generated spatial patterns will not last very long, as unattached individuals are easily dislodged by waves. Investigating the eco-evolutionary feedback between mussel bed formation and the evolution ofmovement and attachment can provide us with interesting insights into eco-evolutionary feedbacks in self-organized ecosystems in general.Using movement trajectories recorded during mesocosm experiments, I observed that mussels use a particular movement strategy. Movement patterns of solitary mussels are similar to a Lévy walk, where many short steps are alternated with very long moves. Lévy walks are frequently observed in nature, yet theoretical models suggest that habitats in which Lévy walks are optimal are rare, as Lévy walks are only optimal when resources are scarce and heterogeneously distributed. In Chapter 2, I argue that the occurrence of Lévy-like movements in mussel beds is due to the eco-evolutionary feedback between self-organized pattern formation and mussel movement. To prove this hypothesis, I simulated mussel bed formation with an individual-based model, where I varied the movement strategy used by the virtual mussels between model runs. The results of these simulations show that a spatially patterned mussel bed is generated most efficiently when mussels make use of a Lévy walk. Further evolutionary analyses, where I test for the invasion success of mutant movement strategies in a mussel population in which all other individuals adopt a resident movement strategy, demonstrate that Lévy walks evolve in my simulated self-organized mussel beds. Because Lévy walks accelerate pattern formation and the spatial pattern in turn increases the survival of these Lévy walkers, my results suggest that, in mussel beds, Lévy walks evolve through an eco-evolutionary feedback between mussel movement and self-organized patterning. Although my model is specifically designed to simulate mussel movement in self-organized mussel beds, the conclusions drawn from this study may explain why Lévy walks are found under much broader conditions than is currently explained in mathematical models.Despite the increasing prevalence of observations of Lévy walks in nature, empiricists more and more notice that organisms might do a Lévy walk in one environment, but a Brownian walk in another. Lévy walks are frequently observed in the movement patterns of organisms that are searching for resources in resource-poor habitats, whereas their movements appear more Brownian-like, with more intermediate-sized steps and fewer large moves, in resource-rich areas. This phenomenon is often explained as an active switch in movement strategy tooptimize search efficiency in both environments. Opposing this view, I hypothesized in Chapter 3 that the intrinsic movement strategy does not change but rather that the observed movement pattern is the consequence of interactions with the environment. Following Einstein’s perspective on Brownian motion in atoms and molecules, I argued that collisions with other objects such as resources or conspecifics causes a move to be prematurely ended. In areas with few objects to encounter, an organism’s movement pattern would not be unrecognizably altered. In dense environments, however, the frequent occurrence of encounters transforms any movement strategy into a Brownian-like pattern. By analysing mussel movement in five different density treatments, I show that observed movement patterns become more Brownian-like with increasing mussel density. In Chapter 4, I found similar results for the movements of mud snails. I verified that this shift to Brownian motion is caused by collisions with conspecifics by disentangling truncated steps and moves into free space, demonstrating that the movement strategy does not change when only considering non-truncated steps. With individual-based model simulations, I showed that an active shift from Lévy to Brownian motion with increasing mussel density is unnecessary, as Lévy walks are equally efficient as Brownian movement in creating spatially patterned mussel beds at high mussel densities. Furthermore, I analytically confirmed the hypothesis that any movement strategy becomes more Brownian-like with increasing encounter rates using a simple argument. My results suggest that observed Brownian patterns in the movement trajectories of animals in their natural habitat can be the consequence of superdiffusive intrinsic movement that is altered by target density.Whether Lévy walks observed in nature are actual Lévy walks or the product of a mixture of different strategies (a ‘composite Brownian walk’) is currently under debate. Using traditional methods, one cannot distinguish between the two movement types. In Chapter 4, a novel technique is demonstrated that helps distinguishing between true Lévy walks and composite movement strategies, by examining whether clusters of small steps coincide with resource patches (which would be indicative of a composite Brownian walk). Using a mud snail experiment as an example, it was shown that local search clusters are not only produced in food patches but also on bare soil, demonstrating that true Lévy walksmay indeed exist in nature. The ability to extract intrinsic movement strategies from observed movement patterns (Chapter 3) and to distinguish between different movement strategies (Chapter 4) can have great implications for the representation of animal movement in ecological modelling: the use of Brownian motion as a default template for animal movement is not always justifiable and should be replaced by a more realistic, density-dependent type of movement template.Mussels, as well as many other organisms, actively aggregate into groups, where they cooperate with neighbouring conspecifics. Because cooperation can be exploited by individuals that do not contribute, the widespread occurrence of cooperation in nature remains puzzling. Theoretical studies have shown that the spatial structure of a population can promote the evolution of cooperation. However, these studies consider local dispersal to be the driving factor behind both the spatial patterning and the occurrence of cooperation, thereby disregarding the fact that many species disperse over a wide range and yet cooperate locally. In Chapter 5, I demonstrated how spatial population structure affects the evolution of investment into byssal thread attachments in spatially patterned mussel beds. Using a simple model, I showed that active aggregation into dense mussel clumps gives rise to the highest levels of cooperativeness over a wide range of environmental stress. These results suggest that active clustering can promote the evolution of cooperation even when offspring are widely dispersed.Cooperation and aggregative movement are two fundamental behaviours that form the foundation of self-organization in mussel beds. Without movement into clusters, mussels are unable to attach their byssus threads to neighbouring conspecifics, and without cooperation, movement into clusters would be a useless endeavour. Because movement and cooperative behaviour are quite dependent on one another, evolution of one of these traits is likely to affect evolution of the other and, subsequently, the spatial pattern that will be generated in the mussel bed. In Chapter 6 of this thesis, I showed that the joint evolution of cooperation and aggregative movement can result in differently patterned mussel beds than when only one of the two behaviours is allowed to evolve in isolation. In most evolutionary models, evolution of other than the one focal trait is habituallydisregarded; my results demonstrate that this may lead to drawing the wrong conclusions.The self-organized pattern that emerges from the individuals’ movement and cooperation in turn also affects the persistence of mussel clumps. With a simple field experiment, I showed that not only inadequately attached mussels can become dislodged by wave stress or predation, but that similarly, small mussel clumps are also more vulnerable to dislodgement than large clumps. Dislodgement often implies removal from the mussel beds into suboptimal habitats with high risk of predation and low food availability. Hence, mussel mortality is linked to the persistence of clumps formed by the self-organization process, and clump persistence thereby influences the selection of particular traits. Hence, a loop develops, where the ecological process of pattern formation adjusts selection processes acting upon the mussels, which than in turn alter the ecological process of pattern formation. Adding this group-level mechanism of selection to our model in Chapter 6 leads to a substantially higher occurrence of the emergence of labyrinth-like patterns than simulations with individual-level selection only. As these patterns are frequently observed in natural mussel beds, these results suggest that multi-level selection is of key importance in the eco-evolutionary feedback that leads to the formation of spatially patterned mussel beds.My findings demonstrate that eco-evolutionary feedbacks are of great importance for the evolution of traits that trigger spatial self-organization in ecological systems. At the individual level, self-organizing traits such as movement or attachment can evolve through the interplay between evolution of individual behaviour and the spatial complexity of the community. As large-scale, self-organized patterns are generated by the actions of and interactions between individuals, pattern formation is similarly affected by this eco-evolutionary feedback that often involves traits that modify the environment. In more general terms, an organism’s behaviour can affect its environment, which in turn influences the fitness of this individual and of others. The eco-evolutionary feedback that arises from the interplay between individual behaviour and spatial patterning can fundamentally alter the mechanisms that drive evolutionarychange by generating a group effect on survival, leading to an additional selection process affecting individual fitness. To truly understand ecological and evolutionary processes in nature, it is of key importance to study eco-evolutionary interactions as they develop in the complex settings of the natural world.

All data in the Integrated Marine Information System (IMIS) is subject to the VLIZ privacy policy Top | Author