Seconds, Not Hours: AI Flood Model Guides Florida Flood Control and Long-Term Planning

New AI flood model gives water managers an up-to-the-minute decision tool

Florida's 2,175 miles of canals are the state's pressure valve during storms. The challenge: conditions shift fast, and conventional simulations can take close to an hour-too slow when gates, pumps and warnings need to move now.

A collaborative team at FIU has built an AI model that runs flood scenarios in seconds and suggests what to do next. It gives managers a live read on risk and a shortlist of actions before water levels turn into headlines.

What you'll learn

• An AI model delivers near-instant flood simulations plus actionable strategies for operations teams.
• Trained on a decade of environmental data, it supports both immediate response and long-term infrastructure planning.

Why this matters for leaders

Traditional physics-based models replicate canal behavior with high fidelity but demand heavy compute and time. During a fast-moving storm, that delay can stall decisions on gate settings, pump schedules and public alerts.

Real-time insight compresses the decision window. The sooner you see credible outcomes, the faster you can allocate crews, adjust infrastructure and coordinate with emergency management.

What FIU built

The FIU team developed an AI system that produces near-instant simulations and recommends specific operational moves. It's engineered to help eliminate or drastically reduce flood impact when minutes matter.

"Accuracy is obviously very important to us, because overestimation of water stages can cause false alarms and panic while underestimation can result in unexpected and dangerous flooding events," said Giri Narasimhan, professor at FIU's Knight Foundation School of Computing and Information Sciences. "We were able to create a tool that provides water managers with the information to either eliminate a flood event or drastically reduce it."

Using nearly a decade of historical data from the South Florida Water Management District, the model learned how rainfall, tides, groundwater and storm surge interact across the region. Historic storms-including Hurricane Irma (2017), Hurricane Sandy (2012) and Tropical Storm Isaias (2020)-were used to tune reliability, and the team tested the system on the Miami River.

The research is detailed in the Journal of Water Resources Planning and Management. For context on the journal, see ASCE's publication overview. Data sources include the South Florida Water Management District.

Operational impact you can act on

• Pre-storm drawdowns: Set canal targets with higher confidence and defend them as conditions shift.
• Gate and pump playbooks: Move from static rules to AI-informed sequences that adjust in real time.
• Staffing and logistics: Stage crews and equipment where the model shows the highest payoff.
• Public communication: Trigger alerts based on probabilistic thresholds, not guesswork.
• After-action reviews: Compare predicted vs. observed stages to tighten future response.

Planning the next 20-30 years

Study co-author Jayantha Obeysekera, director of FIU's Sea Level Solutions Center and former chief modeler at the South Florida Water Management District, points to broader impact: "The model also holds a lot of potential as a tool to help agencies make longer-term decisions," he said. "It could guide 20- or 30-year infrastructure investments, such as whether to build new pumps, reservoirs or levees by screening potential solutions efficiently."

In practice, that means faster screening of capital options, clearer trade-off analysis and fewer dead-end projects. Leaders get a portfolio view that connects today's operations to tomorrow's resilience.

How to get value fast

• Wire up live feeds: Rainfall, tide gauges, groundwater, storm surge forecasts and gate telemetry.
• Set decision thresholds: Define triggers for drawdowns, pump ramps and emergency messaging.
• Run worst-case drills: Use seconds-long simulations to stress-test your SOPs.
• Validate routinely: Benchmark AI outputs against physics models and field data after each storm.
• Establish governance: Document model versions, data lineage and approval workflows.
• Measure ROI: Track avoided damages, reduced response time and maintenance savings.

Upgrade team capability

If you're building leadership and operator skills around AI-driven operations, explore practical training for managers and teams at Complete AI Training - Courses by Job.