Oral Presentation NCGRT/IAH Australasian Groundwater Conference 2019

The Decision Support Groundwater Modelling Project: An Industry Perspective (559)

Keith Brown 1
  1. Rio Tinto Iron Ore, Perth, WA, Australia

Global miners such as Rio Tinto (RT) and BHP rely on the utility of groundwater models to inform water related risk in support of mining.  However, many of the models that, we use lack flexibility to meet the demands of the dynamic work environment we operate in. We are not alone on this issue.  The hydrogeological community has raised similar concerns, about currently accepted practices in groundwater modelling. These ‘inconsistencies’, while widely recognised across the global groundwater modelling community have to this point largely been ignored. 

As joint co-funders in a tripartite consortium RT and BHP and the NCGRT, under the Groundwater Modelling Decision Support Initiative (GMDSI).  By entering into this external partnering relationship it is RT and BHP’s intent to help ‘lift the bar’, with greater emphasis on technical integrity in hydrogeological science particularly in the governance and use of groundwater modelling. By offering its support RT and BHP are utilising its expertise and position to drive development and the application of groundwater models while at the same time build robustness and strengthen its own internal groundwater modelling processes.

The Decision Support Groundwater Modelling Project is a global initiative aimed at advancing mathematical models that will better inform groundwater related business decisions through the development of a framework for decision making and on-going software development. 

GMDSI brings together the world’s leading model ‘thinkers’ and developers who will collectively play a key role in spotlighting and addressing the inconsistencies in groundwater modelling on a global platform. The fundamental premise of decision-support modelling is that decision-pertinent predictions are accompanied by estimates of predictive uncertainty. The modelling process is most likely to be successful in quantifying and reducing the uncertainty of a prediction of management interest if the model construction methodology are designed for this purpose.