Time series modelling and trophic interactions: rainfall, vegetation and un gulate dynamics

Research output: Contribution to journalArticle

Abstract

Abstract Time series analysis is a tool that is now
commonly used when analysing the states of natural populations.
This is a particularly complicated task for ungulates,
since the data involved usually contain large
observation errors and span short periods of time relative to
the species’ life expectancies. Here we develop a method
that expands on previous analyses, combining statistical
state space modelling with biological mechanistic modelling.
This enables biological interpretability of the statistical
parameters. We used this method to analyse African
ungulate census data, and it revealed some clarifying patterns.
The dynamics of one group of species were generally
independent of density and strongly affected by rainfall,
while the other species were governed by a delayed density
dependence and were relatively unaffected by rainfall
variability. Dry season rainfall was more influential than
wet season rainfall, which can be interpreted as indicating
that adult survival is more important than recruitment in
governing ungulate dynamics.

Details

Authors
Organisations
Research areas and keywords

Subject classification (UKÄ) – MANDATORY

  • Biological Sciences

Keywords

  • Population dynamics, Mechanistic model, Kalman filter, Time series data
Original languageEnglish
Pages (from-to)287-296
JournalPopulation Ecology
Volume49
Issue number4
Publication statusPublished - 2007
Publication categoryResearch
Peer-reviewedYes