Burrlioz is a software tool for ssd modelling that was developed over 20 years by statisticians at CSIRO. Given it forms a key plank of the ANZECC/ARMCANZ (2000) Water Quality Guidelines, it has become the de facto means by which default guideline values (DGVs) for toxicants in marine and freshwater systems in Australia and New Zealand are derived.
Although Burrlioz (and the underpinning Burr family of theoretical probability distributions) has served us well since that time, there is a growing sense that its ‘use-by’ date may have arrived.
In this talk, we discuss some of the pros and cons of moving to a more open, transparent computing environment that harnesses the computational power and richness of the R-computing system such as those employed by the French on-line tool ‘MOSAIC’ and the interactive, web-based shiny ssdtools developed by Canada. We also consider some of the statistical and computational challenges associated with ssd modelling and provide preliminary insights that have arisen from a joint Australian-Canadian collaboration that is aimed at enhancing mutual cooperation, research capacity, and knowledge sharing with respect to the statistical evaluation of ecotox data.