Projects

Selected work.

Selected engagements, and the modeling behind them.

01

Global coastal water level reanalysis

Flow Climate2026
ADCIRCGlobal domainReanalysisCompound flooding

A physics-based reanalysis of coastal water levels worldwide, built by simulating more than 350 historical events — supplying the coastal boundary for a compound flood risk framework.

  • Over 350 historical storm events simulated on a global domain, so water level comes from hydrodynamics rather than from interpolation between sparse gauge records.
  • ADCIRC throughout, with consistent forcing and configuration across every event, so results are comparable from one basin to the next rather than only internally valid.
  • The output supplies boundary conditions to a downstream compound flood risk framework, where coastal surge has to meet rainfall and river discharge on common terms.
Model skill

Measured against the water that actually arrived.

Three Florida landfalls hindcast on the public HSOFS mesh, scored against NOAA tide gauges.

ADCIRC skill on the HSOFS mesh against NOAA CO-OPS tide gauges
EventGaugesR²RMSE
Helene 2024Cat 4 · Florida Big Bend360.9610.23 m
Milton 2024Cat 5 · West-central Florida250.9530.30 m
Ian 2022Cat 5 · Southwest Florida350.9250.38 m

ADCIRC on the public HSOFS mesh, scored against NOAA CO-OPS tide gauges. RMSE is the mean error in water level; R² is the share of the observed signal reproduced. Both are means across the scored stations — a single number over a whole coast hides the stations where a model does worst, which are usually the ones that matter.

Want something like this?

Describe the decision you need to make and the deadline, and we will tell you what the modeling actually needs to be.