### Abstract

Original language | English |
---|

Publisher | DTU Compute |
---|---|

Number of pages | 95 |

Publication status | Published - 2017 |

Series | DTU Compute PHD-2017 |
---|---|

Volume | 453 |

ISSN | 0909-3192 |

### Cite this

*Analysis of trait-based models in marine ecosystems.*DTU Compute. DTU Compute PHD-2017, Vol.. 453

}

*Analysis of trait-based models in marine ecosystems.*DTU Compute PHD-2017, vol. 453, DTU Compute.

**Analysis of trait-based models in marine ecosystems.** / Heilmann, Irene Louise Torpe.

Research output: Book/Report › Ph.D. thesis

TY - BOOK

T1 - Analysis of trait-based models in marine ecosystems.

AU - Heilmann, Irene Louise Torpe

PY - 2017

Y1 - 2017

N2 - The overarching theme for this thesis is spatial and temporal variations in ecosystems. The focus is on describing mechanisms that are responsible for generating the spatial and temporal patterns. The thesis contains two separate projects, each exploring a possible mechanism for pattern formation. In both projects, the model formulations result in partial integro-diﬀerential equations. The ﬁrst project in the thesis considers temporal patterns in a size structured population. Size structure is relevant for species that goes through signiﬁcant changes through their lifetime. The population’s response to regular temporal variations in the environment is investigated by introducing a periodic forcing in the system. This can for instance represent seasonal changes. The eﬀect of an imposed forcing is explored both when the underlying unforced system has a stable equilibrium and when it has stable oscillatory dynamics. The numerical solutions show regular cycles where the period is equal to, or an integer multiple of, the forcing period and where the population can have one or more pulses of reproduction in each cycle. Additionally, the numerical results indicate quasi-periodic or chaotic solutions, period doubling bifurcations and coexisting attractors. The bifurcation structure is similar to results for comparable unstructured population models in the literature. This indicates that size structure does not aﬀect the response to periodic forcing. The next project in the thesis considers spatio-temporal pattern formation in a predator–prey system where animals move towards higher ﬁtness. Reaction-diﬀusion systems have been used extensively to describe spatio-temporal patterns in a variety of systems. However, animals rarely move completely at random, as expressed by diﬀusion. This has lead to models with taxis terms, describing individuals moving in the direction of an attractant. An example is chemotaxis models, where bacteria are attracted to a chemical substance. From an evolutionary perspective, it is expected that animals act as to optimize their ﬁtness. Based on this principle, a predator–prey system with ﬁtness taxis and diﬀusion is proposed. Here, ﬁtness taxis refer to animals moving towards higher values of ﬁtness, and the speciﬁc growth rates of the populations are used as a measure of the ﬁtness values. To determine the conditions for pattern formation, a linear stability analysis is conducted. The analysis reveals that the ﬁtness taxis leads to mechanisms for pattern formation, which are based on the prey gathering together. It turns out, that in some cases the problem is not well-posed and an ultraviolet catastrophe occurs, i.e., perturbations with inﬁnitely short wavelength grow inﬁnitely fast. To prevent this, the population dynamics are revised with a spatial feeding kernel, that deﬁnes a spatial range wherein a predator consumes prey. A linear stability analysis for the revised system reveals the ultraviolet catastrophe is avoided and the basic mechanisms for pattern formation are unchanged. Numerical solutions to the revised system are computed to visualize the patterns. The solutions encompass stationary spatial patterns in addition to traveling waves, standing waves and irregular solutions that might be spatio-temporal chaos. The modeling approach of ﬁtness taxis presents a general way to express movement and it is concluded that the model provides a useful framework for describing generic mechanisms for pattern formation.

AB - The overarching theme for this thesis is spatial and temporal variations in ecosystems. The focus is on describing mechanisms that are responsible for generating the spatial and temporal patterns. The thesis contains two separate projects, each exploring a possible mechanism for pattern formation. In both projects, the model formulations result in partial integro-diﬀerential equations. The ﬁrst project in the thesis considers temporal patterns in a size structured population. Size structure is relevant for species that goes through signiﬁcant changes through their lifetime. The population’s response to regular temporal variations in the environment is investigated by introducing a periodic forcing in the system. This can for instance represent seasonal changes. The eﬀect of an imposed forcing is explored both when the underlying unforced system has a stable equilibrium and when it has stable oscillatory dynamics. The numerical solutions show regular cycles where the period is equal to, or an integer multiple of, the forcing period and where the population can have one or more pulses of reproduction in each cycle. Additionally, the numerical results indicate quasi-periodic or chaotic solutions, period doubling bifurcations and coexisting attractors. The bifurcation structure is similar to results for comparable unstructured population models in the literature. This indicates that size structure does not aﬀect the response to periodic forcing. The next project in the thesis considers spatio-temporal pattern formation in a predator–prey system where animals move towards higher ﬁtness. Reaction-diﬀusion systems have been used extensively to describe spatio-temporal patterns in a variety of systems. However, animals rarely move completely at random, as expressed by diﬀusion. This has lead to models with taxis terms, describing individuals moving in the direction of an attractant. An example is chemotaxis models, where bacteria are attracted to a chemical substance. From an evolutionary perspective, it is expected that animals act as to optimize their ﬁtness. Based on this principle, a predator–prey system with ﬁtness taxis and diﬀusion is proposed. Here, ﬁtness taxis refer to animals moving towards higher values of ﬁtness, and the speciﬁc growth rates of the populations are used as a measure of the ﬁtness values. To determine the conditions for pattern formation, a linear stability analysis is conducted. The analysis reveals that the ﬁtness taxis leads to mechanisms for pattern formation, which are based on the prey gathering together. It turns out, that in some cases the problem is not well-posed and an ultraviolet catastrophe occurs, i.e., perturbations with inﬁnitely short wavelength grow inﬁnitely fast. To prevent this, the population dynamics are revised with a spatial feeding kernel, that deﬁnes a spatial range wherein a predator consumes prey. A linear stability analysis for the revised system reveals the ultraviolet catastrophe is avoided and the basic mechanisms for pattern formation are unchanged. Numerical solutions to the revised system are computed to visualize the patterns. The solutions encompass stationary spatial patterns in addition to traveling waves, standing waves and irregular solutions that might be spatio-temporal chaos. The modeling approach of ﬁtness taxis presents a general way to express movement and it is concluded that the model provides a useful framework for describing generic mechanisms for pattern formation.

M3 - Ph.D. thesis

BT - Analysis of trait-based models in marine ecosystems.

PB - DTU Compute

ER -