@phdthesis{Ehrlich2019,
  author    = {Ehrlich, Elias},
  title     = {On the role of trade-offs in predator-prey interactions},
  doi       = {10.25932/publishup-43063},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-430631},
  school      = {Universit{\"a}t Potsdam},
  pages     = {192},
  year      = {2019},
  abstract  = {Predation drives coexistence, evolution and population dynamics of species in food webs, and has strong impacts on related ecosystem functions (e.g. primary production). The effect of predation on these processes largely depends on the trade-offs between functional traits in the predator and prey community. Trade-offs between defence against predation and competitive ability, for example, allow for prey speciation and predator-mediated coexistence of prey species with different strategies (defended or competitive), which may stabilize the overall food web dynamics. While the importance of such trade-offs for coexistence is widely known, we lack an understanding and the empirical evidence of how the variety of differently shaped trade-offs at multiple trophic levels affect biodiversity, trait adaptation and biomass dynamics in food webs. Such mechanistic understanding is crucial for predictions and management decisions that aim to maintain biodiversity and the capability of communities to adapt to environmental change ensuring their persistence. In this dissertation, after a general introduction to predator-prey interactions and tradeoffs, I first focus on trade-offs in the prey between qualitatively different types of defence (e.g. camouflage or escape behaviour) and their costs. I show that these different types lead to different patterns of predator-mediated coexistence and population dynamics, by using a simple predator-prey model. In a second step, I elaborate quantitative aspects of trade-offs and demonstrates that the shape of the trade-off curve in combination with trait-fitness relationships strongly affects competition among different prey types: Either specialized species with extreme trait combinations (undefended or completely defended) coexist, or a species with an intermediate defence level dominates. The developed theory on trade-off shapes and coexistence is kept general, allowing for applications apart from defence-competitiveness trade-offs. Thirdly, I tested the theory on trade-off shapes on a long-term field data set of phytoplankton from Lake Constance. The measured concave trade-off between defence and growth governs seasonal trait changes of phytoplankton in response to an altering grazing pressure by zooplankton, and affects the maintenance of trait variation in the community. In a fourth step, I analyse the interplay of different tradeoffs at multiple trophic levels with plankton data of Lake Constance and a corresponding tritrophic food web model. The results show that the trait and biomass dynamics of the different three trophic levels are interrelated in a trophic biomass-trait cascade, leading to unintuitive patterns of trait changes that are reversed in comparison to predictions from bitrophic systems. Finally, in the general discussion, I extract main ideas on trade-offs in multitrophic systems, develop a graphical theory on trade-off-based coexistence, discuss the interplay of intra- and interspecific trade-offs, and end with a management-oriented view on the results of the dissertation, describing how food webs may respond to future global changes, given their trade-offs.},
  language  = {en}
}
@article{vanVelzenGaedke2018,
  author    = {van Velzen, Ellen and Gaedke, Ursula},
  title     = {Reversed predator-prey cycles are driven by the amplitude of prey oscillations},
  series = {Ecology and evolution},
  volume    = {8},
  journal   = {Ecology and evolution},
  number    = {12},
  publisher = {Wiley},
  address   = {Hoboken},
  issn      = {2045-7758},
  doi       = {10.1002/ece3.4184},
  pages     = {6317 -- 6329},
  year      = {2018},
  abstract  = {Ecoevolutionary feedbacks in predator-prey systems have been shown to qualitatively alter predator-prey dynamics. As a striking example, defense-offense coevolution can reverse predator-prey cycles, so predator peaks precede prey peaks rather than vice versa. However, this has only rarely been shown in either model studies or empirical systems. Here, we investigate whether this rarity is a fundamental feature of reversed cycles by exploring under which conditions they should be found. For this, we first identify potential conditions and parameter ranges most likely to result in reversed cycles by developing a new measure, the effective prey biomass, which combines prey biomass with prey and predator traits, and represents the prey biomass as perceived by the predator. We show that predator dynamics always follow the dynamics of the effective prey biomass with a classic 1/4-phase lag. From this key insight, it follows that in reversed cycles (i.e., -lag), the dynamics of the actual and the effective prey biomass must be in antiphase with each other, that is, the effective prey biomass must be highest when actual prey biomass is lowest, and vice versa. Based on this, we predict that reversed cycles should be found mainly when oscillations in actual prey biomass are small and thus have limited impact on the dynamics of the effective prey biomass, which are mainly driven by trait changes. We then confirm this prediction using numerical simulations of a coevolutionary predator-prey system, varying the amplitude of the oscillations in prey biomass: Reversed cycles are consistently associated with regions of parameter space leading to small-amplitude prey oscillations, offering a specific and highly testable prediction for conditions under which reversed cycles should occur in natural systems.},
  language  = {en}
}
@misc{vanVelzenGaedke2018,
  author    = {van Velzen, Ellen and Gaedke, Ursula},
  title     = {Reversed predator},
  url       = {http://nbn-resolving.de/urn:nbn:de:kobv:517-opus4-411652},
  pages     = {13},
  year      = {2018},
  abstract  = {Ecoevolutionary feedbacks in predator-prey systems have been shown to qualitatively alter predator-prey dynamics. As a striking example, defense-offense coevolution can reverse predator-prey cycles, so predator peaks precede prey peaks rather than vice versa. However, this has only rarely been shown in either model studies or empirical systems. Here, we investigate whether this rarity is a fundamental feature of reversed cycles by exploring under which conditions they should be found. For this, we first identify potential conditions and parameter ranges most likely to result in reversed cycles by developing a new measure, the effective prey biomass, which combines prey biomass with prey and predator traits, and represents the prey biomass as perceived by the predator. We show that predator dynamics always follow the dynamics of the effective prey biomass with a classic ¼-phase lag. From this key insight, it follows that in reversed cycles (i.e., ¾-lag), the dynamics of the actual and the effective prey biomass must be in antiphase with each other, that is, the effective prey biomass must be highest when actual prey biomass is lowest, and vice versa. Based on this, we predict that reversed cycles should be found mainly when oscillations in actual prey biomass are small and thus have limited impact on the dynamics of the effective prey biomass, which are mainly driven by trait changes. We then confirm this prediction using numerical simulations of a coevolutionary predator-prey system, varying the amplitude of the oscillations in prey biomass: Reversed cycles are consistently associated with regions of parameter space leading to small-amplitude prey oscillations, offering a specific and highly testable prediction for conditions under which reversed cycles should occur in natural systems.},
  language  = {en}
}
@article{vanVelzenGaedke2018,
  author    = {van Velzen, Ellen and Gaedke, Ursula},
  title     = {Reversed predator},
  series = {Ecology and Evolution},
  journal   = {Ecology and Evolution},
  publisher = {Wiley},
  address   = {Hoboken},
  issn      = {2045-7758},
  doi       = {10.1002/ece3.4184},
  pages     = {1 -- 13},
  year      = {2018},
  abstract  = {Ecoevolutionary feedbacks in predator-prey systems have been shown to qualitatively alter predator-prey dynamics. As a striking example, defense-offense coevolution can reverse predator-prey cycles, so predator peaks precede prey peaks rather than vice versa. However, this has only rarely been shown in either model studies or empirical systems. Here, we investigate whether this rarity is a fundamental feature of reversed cycles by exploring under which conditions they should be found. For this, we first identify potential conditions and parameter ranges most likely to result in reversed cycles by developing a new measure, the effective prey biomass, which combines prey biomass with prey and predator traits, and represents the prey biomass as perceived by the predator. We show that predator dynamics always follow the dynamics of the effective prey biomass with a classic ¼-phase lag. From this key insight, it follows that in reversed cycles (i.e., ¾-lag), the dynamics of the actual and the effective prey biomass must be in antiphase with each other, that is, the effective prey biomass must be highest when actual prey biomass is lowest, and vice versa. Based on this, we predict that reversed cycles should be found mainly when oscillations in actual prey biomass are small and thus have limited impact on the dynamics of the effective prey biomass, which are mainly driven by trait changes. We then confirm this prediction using numerical simulations of a coevolutionary predator-prey system, varying the amplitude of the oscillations in prey biomass: Reversed cycles are consistently associated with regions of parameter space leading to small-amplitude prey oscillations, offering a specific and highly testable prediction for conditions under which reversed cycles should occur in natural systems.},
  language  = {en}
}
@article{vanVelzenThieserBerendonketal.2018,
  author    = {van Velzen, Ellen and Thieser, Tamara and Berendonk, Thomas U. and Weitere, Markus and Gaedke, Ursula},
  title     = {Inducible defense destabilizes predator-prey dynamics},
  series = {Oikos},
  volume    = {127},
  journal   = {Oikos},
  number    = {11},
  publisher = {Wiley},
  address   = {Hoboken},
  issn      = {0030-1299},
  doi       = {10.1111/oik.04868},
  pages     = {1551 -- 1562},
  year      = {2018},
  abstract  = {Phenotypic plasticity in prey can have a dramatic impact on predator-prey dynamics, e.g. by inducible defense against temporally varying levels of predation. Previous work has overwhelmingly shown that this effect is stabilizing: inducible defenses dampen the amplitudes of population oscillations or eliminate them altogether. However, such studies have neglected scenarios where being protected against one predator increases vulnerability to another (incompatible defense). Here we develop a model for such a scenario, using two distinct prey phenotypes and two predator species. Each prey phenotype is defended against one of the predators, and vulnerable to the other. In strong contrast with previous studies on the dynamic effects of plasticity involving a single predator, we find that increasing the level of plasticity consistently destabilizes the system, as measured by the amplitude of oscillations and the coefficients of variation of both total prey and total predator biomasses. We explain this unexpected and seemingly counterintuitive result by showing that plasticity causes synchronization between the two prey phenotypes (and, through this, between the predators), thus increasing the temporal variability in biomass dynamics. These results challenge the common view that plasticity should always have a stabilizing effect on biomass dynamics: adding a single predator-prey interaction to an established model structure gives rise to a system where different mechanisms may be at play, leading to dramatically different outcomes.},
  language  = {en}
}
@article{VelzenGaedkeKlauschies2022,
  author    = {Velzen, Ellen van and Gaedke, Ursula and Klauschies, Toni},
  title     = {Quantifying the capacity for contemporary trait changes to drive intermittent predator-prey cycles},
  series = {Ecological monographs : a publication of the Ecological Society of America},
  volume    = {92},
  journal   = {Ecological monographs : a publication of the Ecological Society of America},
  number    = {2},
  publisher = {Wiley},
  address   = {New York},
  issn      = {1557-7015},
  doi       = {10.1002/ecm.1505},
  pages     = {29},
  year      = {2022},
  abstract  = {A large and growing body of theory has demonstrated how the presence of trait variation in prey or predator populations may affect the amplitude and phase of predator-prey cycles. Less attention has been given to so-called intermittent cycles, in which predator-prey oscillations recurrently disappear and re-appear, despite such dynamics being observed in empirical systems and modeling studies. A comprehensive understanding of the conditions under which trait changes may drive intermittent predator-prey dynamics, as well as their potential ecological implications, is therefore missing. Here we provide a first systematic analysis of the eco-evolutionary conditions that may give rise to intermittent predator-prey cycles, investigating 16 models that incorporate different types of trait variation within prey, predators, or both. Our results show that intermittent dynamics often arise through predator-prey coevolution, but only very rarely when only one trophic level can adapt. Additionally, the frequency of intermittent cycles depends on the source of trait variation (genetic variation or phenotypic plasticity) and the genetic architecture (Mendelian or quantitative traits), with intermittency occurring most commonly through Mendelian evolution, and very rarely through phenotypic plasticity. Further analysis identified three necessary conditions for when trait variation can drive intermittent cycles. First, the intrinsic stability of the predator-prey system must depend on the traits of prey, predators, or both. Second, there must be a mechanism causing the recurrent alternation between stable and unstable states, leading to a "trait" cycle superimposed on the population dynamics. Finally, these trait dynamics must be significantly slower than the predator-prey cycles. We show how these conditions explain all the abovementioned patterns. We further show an important unexpected consequence of these necessary conditions: they are most easily met when intraspecific trait variation is at high risk of being lost. As trait diversity is positively associated with ecosystem functioning, this can have potentially severe negative consequences. This novel result highlights the importance of identifying and understanding intermittent cycles in theoretical studies and natural systems. The new approach for detecting and quantifying intermittency we develop here will be instrumental in enabling future study.},
  language  = {en}
}
@misc{YamamichiKlauschiesMineretal.2019,
  author    = {Yamamichi, Masato and Klauschies, Toni and Miner, Brooks E. and van Velzen, Ellen},
  title     = {Modelling inducible defences in predator-prey interactions},
  series = {Ecology letters},
  volume    = {22},
  journal   = {Ecology letters},
  number    = {2},
  publisher = {Wiley},
  address   = {Hoboken},
  issn      = {1461-023X},
  doi       = {10.1111/ele.13183},
  pages     = {390 -- 404},
  year      = {2019},
  abstract  = {Inducible defences against predation are widespread in the natural world, allowing prey to economise on the costs of defence when predation risk varies over time or is spatially structured. Through interspecific interactions, inducible defences have major impacts on ecological dynamics, particularly predator-prey stability and phase lag. Researchers have developed multiple distinct approaches, each reflecting assumptions appropriate for particular ecological communities. Yet, the impact of inducible defences on ecological dynamics can be highly sensitive to the modelling approach used, making the choice of model a critical decision that affects interpretation of the dynamical consequences of inducible defences. Here, we review three existing approaches to modelling inducible defences: Switching Function, Fitness Gradient and Optimal Trait. We assess when and how the dynamical outcomes of these approaches differ from each other, from classic predator-prey dynamics and from commonly observed eco-evolutionary dynamics with evolving, but non-inducible, prey defences. We point out that the Switching Function models tend to stabilise population dynamics, and the Fitness Gradient models should be carefully used, as the difference with evolutionary dynamics is important. We discuss advantages of each approach for applications to ecological systems with particular features, with the goal of providing guidelines for future researchers to build on.},
  language  = {en}
}