Computer models developed to forecast the weather need to know what is currently happening in the atmosphere and in the ocean, what we call “initial conditions,” each time the model starts. The process of making these initial conditions is known as “initialization.” This study is about a new initialization method for both the atmosphere and ocean for hurricane forecasts.
To test models, scientists sometimes initialize them with what they call an “idealized” initial condition, made up of data not from the real world, but that resembles it. This allows scientists to have better control over how the atmosphere and ocean change in time. To create idealized initial conditions for a category-1 hurricane on the Saffir-Simpson hurricane intensity scale, a large number of aircraft measurements collected during the last decade are used and combined to obtain wind, temperature, moisture, and pressure analyses. Once these initial conditions are created, the model can be run with different characteristics, so a set of runs is collected.
The ocean is very important because it provides the fuel for the hurricanes, but the atmospheric initialization procedure is not easy to repeat there because ocean measurements near hurricanes are relatively rare. Therefore, a new method that doesn’t require observations is created.Once these initial conditions are created, the model can be run with different characteristics, so a set of runs is collected. The set of runs using just the atmospheric initialization is combined to calculate how the wind, temperature, moisture, and pressure are related to each other, and these relationships are used to guess at initial ocean characteristics. Therefore, real data are used to initialize the atmosphere, and statistics are used to initialize the ocean.
■ Important Conclusions:
- The speed at which a tropical cyclone moves is most important in controlling what is happening in the ocean below the storm.
- Other details of the tropical cyclone are less important in controlling what is happening in the ocean below the storm.
- These findings may be important for how future hurricane models are designed.
Read the article at http://onlinelibrary.wiley.com/doi/10.1002/2017MS000977/full.