Meeting Minutes for February 28, 2025

Hurricane Forecasting & HAFS Ensemble Modeling – Meeting Notes

Overview

This meeting focused on improving hurricane forecasts using NOAA’s Hurricane Analysis and Forecast System (HAFS) and ensemble modeling to assess uncertainty and enhance predictions.

🔗 Useful Links:

• NOAA HAFS Overview: wpo.noaa.gov/the-hurricane-analysis-and-forecast-system-hafs

• HAFS Ensemble Products (EMC): emc.ncep.noaa.gov/HAFS/HAFSEPS/tcall.php

Key Discussion Points

1. HAFS Model & Deterministic Forecasting

• HAFS = HAFS-A + HAFS-B, NOAA’s next-gen hurricane forecast system.

• Deterministic model – Produces the same forecast if given the same initial conditions.

• Key Variables:

  • 🌪️ Peak Surface Wind – Measures storm intensity but only at a single point.

  • 📉 Pressure Gradients – Better predictor of storm evolution and wind intensity.

• Forecasting complexity: Many model parameters ("knobs & dials") impact accuracy.

• Uncertainty matters: "Never trust a single number" – always consider forecast variability.

2. Major Hurricane Forecast Models

Currently, five key hurricane forecast models are used:
✅ HWRF – Hurricane Weather Research & Forecast Model
✅ HMON – Hurricane Multiscale Ocean-coupled Non-hydrostatic Model
✅ GFS (NOAA Global Model) – Global Forecast System
✅ European Model (ECMWF) – Highly regarded global forecast model
✅ CTCX (DoD Model) – U.S. Navy & Air Force collaboration

3. HAFS Ensemble Forecasting & Consensus Models

• Ensemble models combine multiple forecasts to improve reliability.

• NOAA, Microsoft, and AWS partnership allows HAFS to run in the cloud.

• 31 different versions of HAFS run in parallel on Azure, updating every 12 hours.

• Goal: Identify the most reliable ensemble members and assess forecast uncertainty.

4. Data Access & Technical Considerations

• HAFS Ensemble data is stored on Azure.

• Accessing data requires:

  • Azure cloud data tools

• Retrieving large datasets may incur costs, so usage must be managed efficiently.

5. Project Goals & Next Steps

Mission: Improve Forecaster Tools

• Develop graphical products to better visualize forecast uncertainty.

• Assess storm structure & intensity using:

  • Vortex tilt analysis – Tracking storm organization.

  • Wind radius uncertainty – Estimating storm size and impact zones.

• Data clustering methods – Identify meaningful groupings of ensemble forecasts.

Meeting Minutes for March 19, 2025

Notes from 04/11/2025

LINK Provided by Lew:

• /work2/noaa/aoml-hafs1/lgramer/ocean/HEP/gp_tools.py 

•