Smart Watershed Network Management

By Arup

· 1 min. read

Smart Watershed Network Management
Source: Arup

The EPA estimates 10 trillion gallons of untreated stormwater runoff containing raw sewage and toxins enter U.S. waterways annually from city sewer systems, polluting drinking water and intensifying flood risk. Arup and The Nature Conservancy launched the Smart Watershed Network Management project to address this crisis using artificial intelligence and machine learning. The pilot focuses on Florida's Indian River Lagoon, a biologically rich estuary threatened by runoff pollution.

Key points

  • EPA reports 10 trillion gallons of untreated stormwater enters U.S. waterways yearly. This runoff contains raw sewage, trash, and toxins that degrade both environmental and drinking water quality.
  • Smart Watershed project integrates AI and machine learning for stormwater optimization. Arup leads the digital systems component, applying remote sensing, data analytics, and predictive algorithms to Florida's Indian River Lagoon.
  • The system makes real-time adjustments to stormwater infrastructure based on weather. Before major storms, the AI can proactively lower water levels in ponds to create capacity and reduce flooding and pollution.
  • Graph neural networks enable the system to understand watershed relationships. The technology represents interconnected ponds, pipes, and outlets as a network, allowing the AI to reason about how upstream decisions affect downstream conditions hours later.
  • Arup uses simulation to test strategies before applying them to physical infrastructure. The "world-model-style" approach allows thousands of operating scenarios to be tested digitally, eliminating liability risks of trial-and-error on live systems.
  • Open-source tools will enable global scaling of smart watershed solutions. The project is developing compatible software and AI approaches designed to work with any sensor or IoT device, allowing adoption "across the US and globally."

Summarized by MalakyeAI

Links Cited on this Page