Measures of individual uncertainty for ecological models: Variance and entropy

Organisms on the move face uncertainty regarding the state of their environments, and characterizing the magnitude of this uncertainty is important because of its influence on organismal decision making. Two common measures of the uncertainty inherent in a distribution of possible outcomes are variance and entropy, yet there is currently no standard for when one measure should be used over the other. This paper explores this question using two models of resource uncertainty. For small numbers of discrete possible outcomes, variance is the better measure because it captures the spread between outcomes as well as their differential possibilities. However, variance can categorically fail as a measure of uncertainty when distributions are multimodal or discontinuous, in which case entropy should be used to characterize uncertainty.