Abstract:
Accurately predicting species’ responses to climate change is complex and requires the integration of multiple determinants of species-specific sensitivity and exposure (Williams et al. 2008). The two main approaches currently used to assess vulnerability to climate change are: a) general pattern-based correlative models or b) single-species, mechanistic models. Correlative models have been popular due to their ability to quickly generate predictions for multiple species (Thomas et al. 2004). However, detailed species-specific models incorporating comprehensive mechanistic data provide more precise predictions of how individual species will respond to changes in climate (Briscoe et al. 2016, Kearney et al. 2009). To bridge the gap between predictions provided by these species-specific mechanistic models and more rapid, generalised correlative models, we require novel modelling approaches to effectively and quickly assess species’ vulnerability to rising temperatures. In this thesis I sought to validate the usefulness of a simple behavioural index, ‘pant50’, as an indicator of high vulnerability to thermal physiological costs in birds inhabiting hot desert environments. This behavioural index is calculated at a species-specific level and is defined as the air temperature (Tair) at which 50 % of observed individuals of a particular bird species are engaged in panting behaviour, which augments rates of evaporative heat loss. To validate the relationship between pant50 and capacity to safely thermoregulate at high air temperatures, I tested predictions relating heat dissipation behaviours to underlying changes in physiology in a range of arid-zone bird species that exhibited wide variation in pant50. These pant50 values represent data collected from either free-living birds or those in semi-natural captivity in the southern Kalahari Desert.