Trending:
AI & Machine Learning

VicWater wants Bureau of Meteorology data access for flood AI models

Victoria's 18 water corporations are pushing for cross-government data sharing to train AI systems on flood management and leak detection. They want Bureau of Meteorology rainfall data - but haven't formally asked yet. The timing aligns with Victoria's AI Mission promising $30B economic impact, while data centres demand water equivalent to 100,000 homes.

VicWater wants Bureau of Meteorology data access for flood AI models

The Ask

VicWater, representing Victoria's 18 public water corporations, wants access to Bureau of Meteorology data to improve AI-based flood modelling. Chief executive Jo Lim told iTnews there's "strong appetite" for cross-government data sharing, particularly meteorological data that would help predict floods like the 2022 events that cost $24 million in emergency relief.

Notably, VicWater hasn't formally approached the Bureau yet. Some members, including Melbourne Water, have sought access for specific floodplain modelling projects. The gap between stated interest and actual proposals matters here - government data access typically requires demonstrating security controls and clear use cases.

What's Already Happening

The water corporations are already using AI through VicWater's Intelligent Water Networks program. Current trials include processing sensor data for leak detection and using AI to accelerate development approvals during the housing shortage. The association has run sector-wide hackathons using public datasets to get members comfortable with data sharing.

This is significant because it shows implementation before advocacy - they're asking for more data because they're already using what they have.

The Tension

The push comes as Victoria projects AI could add $30 billion to state economic output over ten years, backed by 40+ data centres from AWS, Microsoft, and others. The state's $5.5 million Sustainable Data Centre Action Plan aims to unlock $25 billion in private investment.

But here's the trade-off VicWater is quietly flagging: those data centres need 3,000 litres per second - equivalent to water for 100,000 homes. Twenty-five new hyperscale facilities are proposed, each consuming 100-500 litres per second.

VicWater is urging data centre developers to engage early on water needs and site facilities near recycling plants. Translation: the same government AI mission creating demand for better water management is also straining the resource those systems are meant to protect.

Three Things to Watch

  1. Whether VicWater formalises its Bureau of Meteorology data request with specific technical and governance proposals
  2. How the state balances data centre growth against water utility sustainability concerns
  3. If Melbourne Water's floodplain modelling project becomes a template for broader government data access

The real question: can Victoria's AI ambitions and water constraints coexist, or is this an early warning of infrastructure limits?