Enhanced species distribution models: a case study using essential population data from Actinotus helianthi (flannel flower)

Abstract

Species distribution models (SDMs) quantify the spatial configuration or change of suitable habitat for species. This makes SDMs indispensable for conservation planning and climate adaptation management. It is timely, therefore, to examine the underlying model assumptions more carefully. These models typically use known localities of individuals of the species as an indication of what environmental conditions the species will persist under. Each confirmed record is treated equally as an indication of suitability, thus assuming that all populations are equal. However, populations vary considerably in the number of individuals, ranging from a few individuals to many thousands with implications for how they might persist. The number of individuals may also indicate another dimension to how suitable the environment is to the growth, survival and reproductive success of that species. A second assumption is that within each population all individuals are equivalent in their requirements from the environment. My thesis focused on testing these assumptions by performing field and laboratory experiments, which incorporated population level data, collected from Actinotus helianthi plants, to determine whether populations are equivalent in their response to current or future environments. Specifically, I report on the variation in plant trait values associated with the reproductive niche of the species. I then incorporated some of the limiting soil environment factors identified in my experiments to produce a new and more accurate SDM prediction than an SDM built on climate alone. The choice of predictors in SDMs is crucial to their success, as the model assumes that all relevant factors are included. My thesis is an important foundation for experimentally testing the assumptions inherent in most SDMs, while at the same time, illustrating how these factors can be added to, or combined with, the initial model.