Atmospheric modeling utilizes “parametrization” in a creating a weather or climate model to replace processes that are too small-scale, or too complex, to be physically represented, such as: turbulence in the atmospheric and oceanic boundary layers, the interactions of the circulation with small-scale-topography features, thunderstorms, cloud microphysics processes, etc. Parameterizations have to be designed to account for the large-scale influence of those processes not explicitly included.
Research excerpts from CCS Members studying Atmospheric Modeling:
- Understanding Tropical Cyclogenesis and Rapid Intensification Through Idealized, High-Resolution Simulations by David S. Nolan
- Hurricane Forecasting by Shuyi S. Chen
A complete understanding of the events leading to the formation of a tropical cyclone—a process known as tropical cyclogenesis—has remained elusive for tropical meteorologists. While track forecasts have improved considerably in recent years, and intensity forecasts have shown some improvement, our skill in forecasting tropical cyclogenesis remains very poor. This problem is due at least in part to our lack of understanding of the genesis process. The seemingly sudden appearance of Tropical Storm Katrina off the southeast coast of Florida in August of 2005, and its rapid transition to hurricane status as it traversed the Florida peninsula, highlights the need to understand this phenomenon.
Figure 1: Cyclogenesis and rapid intensification
Rain rate (colored) and surface winds vectors for an idealized simulation of tropical cyclogenesis. Rain rates are in mm/min; Max wind vector is 38 m/s. Values below 0.05 not shown. Times are shown at the top in days, hours, and minutes.
Rain rates are in mm/min; Max wind vector is 38 m/s. Values below 0.05 not shown.
Max wind vector is 38 m/s. Values below 0.05 not shown. Times are shown at the top in days, hours, and minutes. Every 4th vector is shown.
Fortunately, the steadily increasing ability of computer models to accurately simulate the atmosphere continues to bring new insight into the physical processes that precede the transition of a tropical disturbance from a seemingly disorganized cloud feature on a satellite image to a rapidly rotating storm. Our approach is to simulate this process in an idealized setting, one without external features such as sea surface temperature variations, land interactions, and wind shear associated with the jet stream. Even in this simplified environment, the development of circulating winds at the surface, and the well-organized convective bands which drive the storm, can depend critically on subtle processes that occur over just a few hours. The evolution of the low-level wind and rain fields in a simulation with 2 km grid spacing, depicting the rapid increase in wind speed and organization of the convection over a 30 hour period, is depicted in Figure 1.
Unfortunately, the increasing accuracy of these simulations comes at the cost of greatly increased computational demands, in terms of both processing power and disk space. For simulations such as the one shown below, continued access to a super-computing center with parallel processing capabilities and substantial storage will be necessary for further advances.
Forecasting rapid intensity change of a hurricane is one of the most challenging problems in weather research today. A key to improve hurricane prediction is to develop computer models to resolve the detailed hurricane structures (eye, eyewall, and rainbands) and full coupling to the ocean. A state-of-the-art high resolution, fully coupled atmosphere-wave-ocean model for hurricane prediction has been developed by a research group led by Professor Shuyi Chen at the Rosenstiel School of Marine and Atmospheric Science (RSMAS) at the University of Miami. The University of Miami Coupled Atmosphere-Wave-Ocean Model (UMCM) includes three components, the atmospheric model(s), surface wave model(s), and ocean circulation model(s). Using UMCM we can predict not only the detailed heavy rain and extreme winds at 1-2 km resolution, but also storm driven ocean currents, temperature, and surface waves (Figure 2). The continued development and testing of UMCM will be benefit from the increasing computational power at CCS.
Figure 2: UMCM forecasts of rainrate (mm/h, top-left), significant wave height (m, midright), and surface sea temperature (oC) and current (m/s, bottom) in Hurricane Katrina from 0000 UTC 27 Aug 0000 UTC 30 Aug 2005.
MASTHEAD IMAGE SOURCE: “A snapshot of global clouds as detected from outgoing longwave radiation within a 3.5-km global simulation with GEOS-5. As resolutions in GEOS-5 approach cloud scales, numerous cloud structures are resolved, including a major hurricane in the east Pacific, shallow cumulus clouds in the marine layer over the oceans, convection from daytime heating over land, and deep cumulus and stratus clouds along frontal boundaries and large-scale weather systems” by William Putman, NASA/Goddard.