Centre for Biodiversity and Environment Research


Archive of CBER Events

CBER Research Talk - Professor Katriona Shea - 10 October 2017

Start: Oct 10, 2017 1:00:00 PM
End: Oct 10, 2017 2:00:00 PM

Title: Using multiple models to address uncertainty, value of information, and optimal control of disease outbreaks
Speaker: Professor Katriona Shea - Department of Biology, Pennsylvania State University (lab page)
Venue: H O Schild Pharmacology LT (map)
Professor Kate Jones (email)
Major disease outbreaks often generate multiple modeling efforts to assist with forecasting and management. Uncertainty about appropriate parameters, model structure and management intervention implementation can generate significant disagreements, which in turns hampers policy-making for animal and public health. Using examples from human (Ebola) and livestock (foot-and-mouth) diseases, I outline approaches to address uncertainty and learning to help improve epidemiological management.

CBER Research Talk - Dr Nick Isaac - 16 October 2017

Start: Oct 16, 2017 1:00:00 PM
End: Oct 16, 2017 2:00:00 PM

Title: Modelling biodiversity change from messy and biased data
Speaker: Dr Nick Isaac (Centre for Ecology & Hydrology)
Venue: Medawar Building G02 Watson LT (map)
Abstract: We have entered a new geological epoch, the Anthropocene, reflecting the pervasive impact of humans on our planet. One feature the Anthropocene is what ecologists refer to as the biodiversity crisis, or the “Sixth Mass Extinction”. Monitoring and understanding biodiversity change is critical in order to enact effective mitigation strategies, but there is a dearth of high quality data for this purpose. Occurrence records, such as those collected by Citizen Science projects, are a rich source of information: the Global Biodiversity Information Facility (GBIF) database now contains over 600 million records. However, occurrence records were not gathered in a systematic manner, leading to numerous biases. I will describe the application of hierarchical Bayesian occupancy-detection models to unstructured occurrence records, and show using computer simulation that the resultant trends are robust to known biases in the data. I will illustrate the use of these models using a suite of examples, including biodiversity indicators and measuring the impact of pesticides on beneficial insects.