ÌìÃÀ´«Ã½

Environmental Informatics (EI) uses environmental data to reveal, quantify, and validate scientific hypotheses. It does this with a panoply of tools from the Statistics, Mathematics, Computing, and Visualisation disciplines.

The Centre for Environmental Informatics (CEI)'s vision is to promote the University of ÌìÃÀ´«Ã½'s National Institute for Applied Statistics Research Australia (NIASRA) as an international leader in Environmental Informatics. CEI emphasizes statistical modelling and computation for big spatial and spatio-temporal environmental datasets that are incomplete, noisy, and generated by possibly non-linear, multi-scale, non-Gaussian, and multivariate processes. CEI has active research programs (e.g., in global remote sensing, CO₂ flux inversion, space-based spectroscopy, multivariate spatial stochastic modelling) that involve postgraduate PhD students, post-doctoral fellows, and visitors.

EI is a relatively young discipline; just as bioinformatics has grown and now includes biostatistics as a sub-discipline, EI has the potential to integrate itself into the environmental sciences and be much broader in scope than classical environmental-statistical methodology. An expository paper on EI is given by: Cressie, N., Burden, S., Shumack, C., Zammit-Mangion, A., and Zhang, B. (2017). . Wiley StatsRef: Statistics Reference Online, pp. 1-8 (doi:10.1002/9781118445112.stat07717.pub2).

There is a realization in the environmental sciences that uncertainty needs to be quantified, not only in the data, but also in the scientific processes and their parameters. A short article on uncertainty quantification: Cressie, N. (2017). . CEI consists of a group of researchers and support staff who are developing statistical, mathematical, computational and visualisation tools to interface with the environmental sciences. It emphasises uncertainty quantification based on a hierarchical statistical modelling (HM) paradigm, embedding it in all aspects of EI, from observations, to inference, to decisions. For example, the power of data mining to look for unusual patterns in a sea of big data is exploited, and we add to it by addressing such questions as, "Are the patterns real?" and "Unusual in relation to what?" A decision-theoretic framework within the HM paradigm gives environmental policy-makers a way to make rational decisions in the presence of uncertainty based on competing risks (i.e., probabilities) and competing benefits. Further, high-performance computing is essential for implementing HM, and parallelisation is at the core of CEI's new hardware and software developments.