© Antje Herde
D06
Uncovering complex behavioural niche mechanisms from individual-level ecological time series
The automated collection of behavioural data, using technology such as GPS tracking or acceleration sensors, offers exciting opportunities for studying behavioural niche mechanisms.
State-switching models such as hidden Markov models (HMMs) are particularly suited to draw ecological inferences from various types of such time series data, e.g. GPS, video or acceleration data, as they allow not directly observable behavioural modes to be uncovered and related to individual-specific features as well as environmental covariates. As method development is still lagging behind the technological advancements, in this project we develop novel approaches for analysing such data. We will expand our research on statistical methodology relevant when studying individualised niches and how these are shaped by the interaction between individuals and their environment. The methodological projects focus on specific aspects of individualised niches, motivated from existing collaborations and interactions within the CRC. The project also supports empirical projects within the CRC, using state-of-the-art time series analysis approaches in order to fully exploit the potential of such data.
We aim to
- Develop new methodology for analysing behavioural time series data, which aims at i) formally distinguishing and investigating distinct behavioural phenotypes within a species, via mixtures of HMMs, and ii) inferring complex temporal niche mechanisms, including interactions with internal and external drivers as well as individual variation;
- Analyse data collected within the CRC, in particular the fur seal offspring data (pup production over time) collected within Project A01 (Hoffman & Gossmann), the mice lab data (movement between compartments of a cage) collected within Project A02 (Richter), the data related to the onset of metamorphosis in fire salamanders (capture-recapture) collected within Project A04 (Caspers), the Galápagos sea lion data (horizontal movement and diving) collected within Project B07 (Krüger), the beetle data (activity) collected within Project C01 (Kurtz), and the buzzard data (movement between habitats) collected within Project C03 (Chakarov & Krüger).
The outcome of this project will be a synthesis paper that presents and discusses the various options for modelling diel activity patterns and the sources of their variation across individuals, using several case studies to illustrate the methods.