Cross-Continental Analysis Shows That Disturbance Effects on Reptile Body Condition Do Not Predict Abundance Responses

Abstract

Ecological disturbances are discrete events that alter or transform the physical, chemical, or biological characteristics of ecosystems. Disturbance can cause animal populations to decline and, according to the risk-disturbance hypothesis and population collapse framework, these declines can be predicted by declines in animal body condition. However, no research has empirically examined the general relationship between body condition and abundance, nor their relationship in response to disturbance. We used a combined dataset representing 33 studies and > 42,000 observations of 75 species from Australia, New Zealand, Spain and the United States of America to test predictions relating to the relationship between reptile body condition and abundance. We first investigated the relationship at the site level and then used meta-analytical models to test whether populations showed linked changes in abundance and body condition in response to disturbance. We further tested whether key environmental and species traits influenced this relationship and whether there was a time-lagged effect of body condition responses on abundance. We found a positive relationship between mean reptile body condition and abundance at the site level. However, the relationship was largely lost when investigating population responses to disturbance. As such, our results provided no support for the risk-disturbance hypothesis and limited support for the population collapse framework. Therefore, the impacts of disturbance on reptile body condition cannot be assumed to reflect or predict abundance responses. We provide a new conceptual framework that shows how disturbances can modify or uncouple the relationship between abundance and body condition by influencing underlying drivers, such as predation, competition and resource availability. Monitoring programs that infer population impacts based on changes in body condition should first confirm the relationship between these two variables in the relevant study system.