Summer school on HIERARCHICAL MODELING FOR ENVIRONMENTAL PROCESSES, Bolzano, Italy

June, 20 – 24, 2011 – Bolzano/Bozen, Italy

Lecturer

Prof.  Alan Gelfand 
J.B. Duke Professor of Statistics and Decision Sciences
Department of Statistical Science, Duke University
Durham, NC, USA

Programme and registration details are available at www.mi.imati.cnr.it/conferences/abs11.html

Interested people are invited to contact the ABS11 Secretariat at abs11@mi.imati.cnr.it

This course is intended to expose the value of hierarchical modeling

within a Bayesian framework for investigating a range of problems in
environmental science. In particular, we focus on stochastic modeling
for such problems driven by the general hierarchical perspective,
[data | process, parameters][process | parameters][parameters]. This
specification is richer than it may appear, as the course will
demonstrate. More importantly, it allows the model development to
focus on the environmental process of interest, integrating the
sources of information that are available. Primary problems of
interest include assessment of environmental exposure, fusion of
environmental data from different sources, and assessing environmental
change and its potential impact on ecological processes.

The course will have a practical orientation, emphasizing model
development, computation and inference driven by real examples. The
course will begin with a brief review of Bayesian inference,
hierarchical modeling and Bayesian computation. Then, since most
environmental processes (including all of the ones we consider) are
observed over space and over time, we present foundational material on
spatial and space-time analysis including material on modeling for
point referenced data, areal unit data, and point pattern data. We
will also discuss multivariate processes, space-time processes, and
computation for large datasets. Real examples will include (i)
exposure assessment for particulate matter and ozone, (ii) data fusion
using monitoring station data and computer model output for ozone,
particulate matter and wet deposition of sulfates and nitrates, (iii)
inference regarding environmental extremes illustrated through
temperature data, (iv) relating environmental factors to species
distributions and climate change to plant performance, (v) distributed
lags in space-time regression with application to ozone formation.
Sessions devoted to implementing model fitting and inference using
Markov chain Monte Carlo methods will supplement the lectures and code
for illustrative examples will be provided.

The school will make use of lectures, practical sessions, software
demonstrations, informal discussion sessions and presentations of
research projects by school participants. The slides and background
reading material will be distributed to the students before the start
of the course.