Bringing the QWMN Water Modelling Hive to life!

The idea for such a mechanism came from conversations with Lindsay Millard of seqwater at the 2018 QWMN Forum where we talked about the need for skills development in coding in python. Those conversations gradually evolved and gathered speed as we engaged with Daniel Botelho from BMT and David Hamilton from the Australian Rivers Institute – we were developing the idea that we could create a group that uses a water modelling workflow as a framework for dividing up what is actually involved in water modelling over time into different areas of capability.

These conversations culminated to date in a workshop held in September with 46 water modellers mostly from Queensland but spread across the States of Australia and even internationally.

To structure the conversations in the workshop, and more broadly the work of the Hive, Dan Botelho had bravely developed a conceptual framework of the key activities / tasks in a water modelling workflow. That framework is shown below.

The framework divides water modelling into the following activity/task areas:

• Calibration loop  – the set of activities involved in setting up, calibrating and validating models including wrangling input data and undertaking uncertainty and performance assessment
• Scenario loop – the set of activities involved in running and using models to assess the behaviour of different management options including doing nothing, preparing and analysing outputs and undertaking uncertainty and sensitivity analysis
• R&D loop – the set of activities involved in learning from previous modelling exercises, in improving models through new techniques, data and maybe even using different models
• Communication and consultation – the set of activities that involve communication and consultation / engagement with different stakeholders from commissioning clients (internal or external), data providers/owners, other modelling staff, decision makers and perhaps even those who could be impacted by decisions

The workshop participants listened to Daniel explain the overall framework, engaged in some discussion then were split into break out groups to identify the key knowledge, and skill development needs of water modellers in each of the 4 workflow areas.

What was identified? Perhaps not surprisingly with 46 modellers, a lot! Below is a brief summary of the top ones from each group:

• Calibration loop  – robustness analysis; understanding how to know the limits of applicability of a model as a predictive tool; best practice for calibration including standardisation of methods; data sourcing; data QA and data representativity assessment
• Scenario loop – problem definition – getting this right up front and knowing what it means for model design; how do we facilitate bringing together multiple skill sets / disciplines to modelling; knowing when to stop – when we have run enough scenarios and variants of scenarios
• R&D loop – model data management; improving knowledge of and skills to access open earth data; combining multiple data sets from different sources
• Communication and consultation – how to talk with decision makers; how to visually represent and communicate model results; producing multiple outputs for multiple audiences; how to explain model outputs

There was additional discussion on how to operationalise the Hive. This included broad agreement that activities / events were essential. There were some voices in favour of having a platform for sharing and others in favour of not having a platform, or at least not MS Teams!

With the diversity of voices and the breadth and volume of needs identified we have decided to run an informal event at the Ship Inn in Brisbane on Friday 19th November 3:30pm-5:30pm to help us determine what to focus on. If you can be in Brisbane then please register (free) and come along. If you are further away then please contact Brian on b.mcintosh@griffith,edu.au with your views on what is most important to you and why.