Improving the GFDL/GFDN Operational Tropical Cyclone Models at NOAA/NCEP and Navy/FNMOC

Biju Thomas,
Isaac Ginis

This project is funded by the NOAA Joint Hurricane Testbed Program.

The forecast operations of both NOAA’s National Hurricane Center and the U.S. Navy’s Joint Typhoon Warning Center require more accurate GFDL and GFDN models for improved model guidance and as integral parts of ongoing multi-model ensemble forecast efforts. This project is aiming to (1) further improve the performance of both the operational GFDL model at NOAA’s National Centers for Environmental Prediction (NCEP) and the operational GFDN model at the U.S. Navy’s Fleet Numerical Meteorology and Oceanography Center (FNMOC) and (2) provide assistance to NCEP and FNMOC in transitioning the model upgrades to operations.

Specific tasks for 2013-2015:

  1. Upgrade the GFDL and GFDN atmospheric model resolution and physics
  2. Upgrade the GFDL and GFDN ocean model resolution, physics, domain configuration, and initializaion using URI’s newly-developed MPIPOM-TC ocean model
  3. Upgrade the GFDL and GFDN air-sea exchange parameterization

This work is conducted at URI and GFDL in close collaboration with our NCEP/EMC and FNMOC colleagues, building directly upon the successes of our previous work to improve and operationally implement GFDL and GFDN model upgrades. We also collaborate with scientists at ESRL towards the implementation of the ESRL sea-spray parameterization. All data and results obtained are shared fully with NOAA and the Navy.

In collaboration with NOAA’s GFDL we are developing a high-resolution version of the GFDL and GFDN models with upgraded physics. In the new model configuration, the area of the innermost nest with highest resolution remains the same as the present version, but with an increased horizontal resolution from 1/12° to 1/18° grid spacing. This 1/18° grid spacing is the highest resolution physically justified for the GFDL model because it is hydrostatic. In order for this high-resolution model to be operationally feasible, open MP computer software is being incorporated to improve efficiency. Also, several modifications to the model physics have been made to maximize the benefit of the increased resolution. A preliminary version of the model has been run for selected cases from the 2010, 2011, and 2012 Atlantic hurricane seasons. Results indicate a significantly improved storm structure for most storms, a reduced negative intensity bias for intense storms, and a neutral intensity impact for weak storms, probably because of the physics modifications being made. Finally, the track forecast skill has improved, as illustrated in the Hurricane Nadine (2012) comparisons in Fig. 1.

Figure 1. GFDL simulated tracks for Hurricane Nadine (2012) at four initial times. Each panel includes current operational GFDL (green), high-resolution GFDL (blue), and observed (black).