Model Selection in Species Distribution Modelling with Minimum Message Length

Species distribution models help us understand where animals and plants live and why. Selecting the right model is key to identifying the most important environmental factors and the biological mechanisms at work. This is crucial for conservation, evolutionary biology, and protecting biodiversity. Current model selection methods for species distribution models are often inaccurate. This thesis develops a new, more accurate selection scheme based on minimum message length, making it efficient and robust against errors.