Paper on improving model forecasts through understanding of the region near the ocean surface released online in Earth and Space Science – Hurricane Research Division

An important tool that forecasters use to predict the weather is a set of computer models that forecast things such as the wind, temperature, moisture, and pressure on a set of places called grid points.  In NOAA’s models to predict hurricanes, the distance between these points is about 1-3 km. Although hurricanes are hundreds of kilometers across and can be bigger than the state of Texas, the strongest wind occurs in gusts that are only a few meters across. So, the computer models cannot see these gusts.

These gusts are associated with swirling flow and mix the heat, moisture, and momentum (mass and velocity) between the ocean and the atmosphere nearby, and in the part of the atmosphere closest to the surface (what we call the boundary layer).  To forecast them, their average effects are included in the computer models using what we call parameterization schemes, especially near the ocean and land surface where hurricanes get heat and moisture that fuels their engine. However, different boundary parameterization schemes produce different forecasts, some of which could be far away from reality.  

We used the new Hurricane Analysis and Forecast System (HAFS), part of NOAA’s Unified Forecast system, and observations from NOAA Hurricane Hunter aircraft to understand why two different boundary layer parameterization schemes make different forecasts, and how well the schemes forecast what is actually happening in the boundary layer.

Important Conclusions:

●  The amount of mixing that the parameterization schemes produce is controlled by a variable known as the eddy diffusivity.  Small changes to the eddy diffusivity sometimes lead to large changes in forecasts which could deviate from reality.

●  The amount of mixing, and thus the eddy diffusivity, can be calculated based on observations from NOAA Hurricane Hunter aircraft.  Modified schemes using these values can improve forecasts. 

For more information, contact aoml.communications@noaa.gov. You can read the study at https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020EA001422.