Special Session
17th Biennial Congress on Modelling and Simulation (MODSIM07)
University of Canterbury, Christchurch, New Zealand, 10-13 December 2007

Description   Stochastic models have been used to describe a wide range of natural phenomena which evolve over time, and are they frequently used as a first step to parameter estimation, experimental design and hypothesis testing, and as models for simulation experiments. For example, they have been used to describe the evolution of populations, the spread of epidemics, competition between species, the transmission and prevalence of disease, and the evolution of DNA sequences. Over the past few years there has been renewed interest in developing statistical procedures for fitting stochastic models. For systems that be monitored continuously in time, the theory and practice of statistical inference is well developed. However, many natural systems can only be sampled at discrete time points. For example, an animal population might be sampled at the end of each breeding season, and this explains in part why discrete-time models are predominant in the ecological and applied population biology literature. Discrete sampling in the context of continuous-time models presents many challenging statistical problems: the construction of estimators, bias reduction, imputation for missing data, and optimal sampling.

This session, sponsored by the ARC Centre of Excellence for Mathematics and Statistics of Complex Systems (MASCOS), brought together researchers and practitioners who employ statistical and other stochastic methods in genetics, ecology, epidemiology, population biology and the environment.

Organizers   Dan Pagendam and Phil Pollett <pkp at maths dot uq stop edu period au>

Papers   All papers were refereed by two anonymous reviewers and one of two session editors. All are available online in the Electronic Proceedings.

Selected papers from this session appeared in Environmental Modeling & Assessment (Springer).

Submitting author	Presenter affiliation	Title of paper
Shahadat Chowdhury (presenter) and Patrick Driver	University of New South Wales	An ecohydrological model of waterbird nesting events to altered floodplain hydrology
Nitin Muttil (presenter) and K.W. Chau	Victoria University Melbourne	Revealing patterns in coastal water quality data using statistical analysis
Andrew Davey, Patrick Doncaster and Owen Jones (presenter)	University of Melbourne	A stochastic model for shelter use in mobile fish population: the effect of intraspecific competition
Patrick N.J. Lane (presenter), P.M. Feikema, C.B. Sherwin, M.C. Peel and A. Freebairn	University of Melbourne	Physically-based prediction of water yield from distributed water supply catchments
Jay W. Larson (presenter), E.T. Ong and C. Tokarz	Australian National University	The spheroidal data analysis library and toolkit: tools for climate model output analysis
Julia Piantadosi (presenter), Boland and Phil Howlett	University of South Australia	Generating synthetic rainfall on various timescales - daily, monthly and yearly
Daniel E. Pagendam (presenter) and Phil Pollett	University of Queensland	Optimal sampling and problematic likelihood functions in a simple population model
Phil Pollett	University of Queensland	Ensemble behaviour in population processes with applications to ecological systems
Joshua V. Ross (presenter) and Thomas Taimre	King's College Cambridge	On the analysis of hospital infection data using Markov models
L. Augusto Sanabria (presenter) and R.P. Cechet	Geoscience Australia	Monte-Carlo modelling of severe wind gust
Subana Shanmuganathan (presenter), Philip Sallis and W. Claster	Ritsumeikan Asia Pacific University	Statistical methods in ecological dynamics modelling
George Yu Sofronov (presenter), G.E. Evans, J.M. Keith and D.P. Kroese	University of Queensland	Identifying change-points in biological sequences via sequential importance sampling
R. Brian Webby (presenter), David Green and Andrew Metcalfe	University of Adelaide	Modelling water blending - sensitivity of optimal policies