Poor modeling of hurricane boundary layer turbulence in computer models is a key obstacle to improving hurricane intensity forecasts. This study uses a recently developed modeling framework based on a large-eddy simulation, or LES, (where model grids are small enough to resolve turbulence) to evaluate the pros and cons of four different planetary boundary layer (PBL) schemes in hurricane conditions. Using this insight, we recommend suitable PBL schemes for hurricane modeling and propose solutions to address issues in these schemes.
Fig. 1 Vertical profiles of effective eddy viscosity and wind velocity flowing toward or away from the center from the LES (black), and tests using four different PBL schemes (colored lines).
Summary: The part of the atmosphere closest to the earth’s surface (the planetary boundary layer or PBL) is where the energy from the warm water below a tropical cyclone is transferred into the atmosphere fueling the storm. This transfer is done by turbulent features or eddies; these eddies are small (less than 100 m across) and are comprised of random and continuously changing wind. Poor understanding and modeling of this turbulence in hurricane conditions in computer models is a key obstacle to improving hurricane intensity forecasts.
Computer models forecast the weather on a regularly spaced grid, but the size of turbulent eddies is much smaller than the space between forecast grid points in our current models. Thus, the effects of turbulent eddies are not actually forecast in the models, but are estimated using what we call parameterization, using techniques known as PBL schemes. Current PBL schemes are designed for non-hurricane conditions, and we don’t know whether they can properly represent turbulence in the hurricane boundary layer, especially since observations of turbulence in the hurricane boundary layer over the open ocean are very rare.
There are computer models where the grid spacing is small enough to forecast turbulence, but they require too much computer power to make forecasts using current computers. Using a recently developed modeling framework, we can run these LESs over a very small area about the same size as the grid spacing in our current computer models and compare the average LES results over its entire domain to the performance of PBL schemes to see how well these schemes do. This study evaluated four different PBL schemes in hurricane conditions based on LESs with 10-m distance between grid points. The findings provide valuable guidance for the development of better PBL schemes in high-wind conditions that can further improve hurricane forecasts.
■ Important Conclusions:
- Comparison of the four different PBL schemes with the LES results shows that most schemes are inherently flawed in hurricane conditions.
- Some PBL schemes produce radial inflow (wind that is toward the hurricane center) near the surface that is too strong (see the blue line in Fig. 1b), which is due to large eddy viscosity (a measure of the ability of turbulence to move momentum vertically, Fig. 1a). Some schemes produce inflow that is too weak (e.g., see the green line in Fig. 1). This inflow is important because it brings the warm, moist air that fuels the hurricane toward its center.
- The scheme that performs best, known as MYNN3.0, performs well due to the sophisticated way it estimates what is happening in the boundary layer; note in Fig. 1 how the orange line (MYNN3.0) is closest to the LES (black). Further analysis of MYNN supports a “three-layer” strategy for the parameterization of mixing length (the average size of turbulent eddies) in the hurricane boundary layer (HBL): 1) the surface layer (bottom 10% of the HBL) where the size of turbulent eddies increases with height, 2) the lower-to-middle HBL where turbulence is primarily generated by wind shear (the change in wind speed and direction with height), and the average size of turbulent eddies can be used in the PBL scheme, and 3) the upper HBL and above where turbulence generated by buoyancy (where warm air rises into cool air above) becomes important.
For more information, contact email@example.com. The full study can be found at https://doi.org/10.1029/2022MS003088. The author would like to acknowledge high-performance computing support from Cheyenne (doi:10.5065/D6RX99HX) provided by NCAR’s Computational and Information Systems Laboratory, sponsored by the National Science Foundation. The work was supported by award NA21OAR4320190 to the Northern Gulf Institute at Mississippi State University from NOAA’s Office of Oceanic and Atmospheric Research, U.S. Department of Commerce. Xiaomin Chen was also supported by an NRC Research Associateship award.
All news and articles are copyrighted to the respective authors and/or News Broadcasters. eWeatherNews is an independent Online News Aggregator
Read more from original source here…