The effect of spatiotemporal sampling strategies and data acquisition accuracy on the characterization of a dynamic ecological system

Abstract

Ecosystems generally have a great number of interacting processes, and their characteristics can vary rapidly across both space and time. With such complexity, it can be difficult to collect sufficient data to adequately understand the behaviour of the observed system. This paper describes a method being used to examine the relationship between sampling accuracy and classification fuzziness. A simulation model acts as a surrogate for a "real-world" resource-habitat system. Samples are collected from the running simulation according to different classification schemes, and the ability to map species habitat by resource levels from each sample are compared to the actual simulation behaviour. The balance between sampling accuracy and the vagueness within the classification scheme compared to the complexity of the modelled system is discussed.