The fortnightly BRAG meeting will be held this Thursday 20/10 at 1 pm via Zoom/GP-Y801. This week we will have presentations by @Scott Forrest and @Mitchell O’Sullivan.
Zoom Link: https://qut.zoom.us/j/85315873893?pwd=bWlWci9lM2Z3RmFCdjc5bmppWG1qUT09
Password: brag@QUT (if prompted)
Scott’s Talk
Title: Comparing approaches for predicting species distributions from animal GPS tracking data.
Abstract: The step selection analysis framework is widely used to quantify animal-environment relationships from animal location data gathered with GPS devices. Unlike resource selection functions that assume location independence, generating species distribution predictions from step selection analyses is not as straightforward. Here we compare three approaches for generating species distribution predictions in novel areas from fitted step selection models; a resource selection function, an approach using a diffusion-taxis equation, and a simulation-based approach that aggregates biased correlated random walk (BCRW) trajectories. To fit the step selection model we use a Bayesian approach to incorporate the variability between individual animals, and test the prediction approaches on both synthetically-generated data and GPS-tracking data from Asian water buffalo (Bubalus bubalis) in Arnhem Land, Northern Territory. We use an out-of-sample cross validation approach to assess the predictions of the three approaches, using individual animals that were not included in model fitting. The simulation-based BCRW approach is more flexible and enables shorter-term forecasting, whereas the analytic methods are more computationally efficient. The goal of identifying an appropriate method for generating species distribution predictions from GPS-tracking data is part of a broader project led by CSIRO named ‘SpaceCows’, which aims to GPS-track feral herds in Northern Australia to increase our understanding of bovid spatial ecology and to predict the dynamic distribution of populations in un-sampled areas that can be used for decision making (e.g. harvest management or pest control).
Mitch’s Talk (Code Demo/Round Table)
Title: AdaptABC: A Julia Package for Novel Adaptive SMC-ABC Algorithms
Abstract: I’ll be running through some of the inner workings of the Julia package I’ve started as part of my PhD. I’ll demonstrate some never-before-seen features of the package and ask for feedback about the usage and API. If time permits, I’ll show off some of the results from my PhD that couldn’t make it into my confirmation.