Modelling Environmental Sensitivity

In response to the global threat of climate change, it becomes increasingly important to understand which geographic areas are most likely to be impacted by climate change. A PhD student at the University of Canberra is using stream invertebrates (caddis-, stone-, and mayflies) to understand how temperatures experienced over evolutionary timescales affects the tolerance to global warming. 

Climates vary in daily, seasonal, and annual cycles, therefore the local species must tolerate all three in order to survive in a particular location. But which temperature regime is most important for understanding species’ thermal tolerance?

One hypothesis states that annual temperature range drives temperature tolerance: locations with a greater difference between the warmest (summer) and coldest days (winter) have species with greater thermal tolerances. As a consequence of this theory, organisms in thermally stable ‘tropical’ regions are predicted to have a narrower thermal tolerance and be more sensitive to warming compared to organisms from more variable ‘temperate’ regions. 

The UC researcher used support from Intersect eResearch Analyst, Kyle Hemming, to compare models that tested the different theories using a variety of temperature sensitivity metrics and three taxonomic insect orders (caddis-, stone-, and mayflies). The model selection framework was developed using AIC scores (Akaike’s Information Criterion). Akaike’s Information Criterion compares the fit of multiple models to the same data – in this case, how well different temperature data link with community composition of the flies. 

The AIC scores for the models are then ranked according to how well the temperature models fit the fly data, with the best model having the lowest AIC score. Importantly, AIC penalises more complex models, so that the model that describes the most amount of variation in the data, while also remaining relatively simple, gets the lowest (and best) scores.

As a result of this model fitting framework, climate zone, elevation, and body size, as well as the temperature measures ‘annual temperature range’ and ‘annual maximum temperature’ were the best predictors for thermal tolerance for the three groups of stream insects. 

The results of this analysis suggest that communities in stable, warm regions may be more sensitive to increases in temperature compared to communities in more variable temperate regions.