NSW Stats Society October Meeting: Zammit Mangion -- Answering important questions about the environment using multivariate space-time modelling

This month’s Stats Society NSW talk is this evening (Tuesday 18 October) at UTS and is
on multivariate space-time modelling with environmental applications.  Details are
below, all are welcome, refreshments are served from 6pm.  

Michael 

=== 

Time: 
6:00pm - 6:30pm: Refreshments 
6:30pm - 7:30pm: Lecture 
7:45pm onwards: Dinner (at a nearby restaurant) 

Venue: CB07.03.010B G.  Level 3, University of Technology, Sydney - Building 7, 638
Jones Street, Ultimo, NSW 2007 

Dr Andrew Zammit Mangion 

NIASRA, University of Wollongong, Australia 

Answering important questions about the environment using multivariate space-time
modelling The natural environment is made up of systems that vary in both space and time
and that interact with one another in complicated ways.  Frequently, the analysis of
just one physical phenomenon, in isolation of the others, is insufficient to answer
important questions about the environment such as “What is the cause for the net
observed increase in Carbon Dioxide?” In this talk I will outline the principles
behind the statistical theory of multivariate space-time modelling, and proceed with
discussing its use in two important applications.  The first will consider the
assessment of Antarctica’s contribution to sea-level rise in the past decade using
satellite data.  The second will consider the problem of localising methane sources in
the UK from ground station data.  Practical implementations and modelling assumptions
will be discussed, while a distributed computation algorithm used to render the
inference tractable highlighted.  This is joint work between the University of
Wollongong, Australia, and the University of Bristol, UK.  

Biography of Dr Andrew Zammit Mangion Andrew Zammit Mangion is a Senior Research Fellow
in the National Institute for Applied Statistics Research Australia (NIASRA) at the
University of Wollongong, Australia.  He obtained his PhD in 2012 from the University of
Sheffield, UK, on approximate inference for spatio-temporal models, and has subsequently
applied his work in several contexts within the social and environmental sciences.  In
2013 Andrew was awarded a National Academy of Sciences prize for his work on the
modelling and prediction of armed conflicts.