Spatial-statistical downscaling with uncertainty quantification in biodiversity modelling

Abstract

Accurate downscaling with uncertainty quantification and its inclusion in biodiversity models are crucial for accurate, valid inferences and predictions. We propose a two-stage protocol that propagates downscaling uncertainty to generalized linear models commonly used for biodiversity modelling. Simulation studies show that our two-stage protocol improves the quantification and propagation of uncertainty in biodiversity modelling when compared to existing methods. Finally, we apply the protocol to predict moss occurrence in East Antarctica.

This research is joint with Distinguished Professor Noel Cressie and Associate Professor Andrew Zammit-Mangion from the University of Wollongong, and Dr. David Clarke and Professor Melodie McGeoch from Monash University