JP3.49
NEXRAD in Space: A Solution to the Hurricane Intensity Prediction Problem
Gregory J. Tripoli, University of Wisconsin, Madison, WI; and V. Chandrasekar, S. S. Chen, G. Holland, E. Im, R. Kakar, W. E. Lewis, F. D. Marks, A. Mugnai, E. A. Smith, and S. Tanelli
It is well known that numerical model forecasts of hurricane intensity have not improved over the past two decades at the rate that track forecasts have. This is attributable to the fact that track forecasts are most strongly tied to the ability to forecast the large scale environment which steers the storm. The large scale environment tends to be quasi two-dimensional (balanced), has space scales on the order of hundreds to thousands of kilometers and time scales of days. It is observable by conventional and current remote observation systems and features long time scales of predictability. On the other hand, hurricane intensity is largely controlled by the internal energy containing dynamics of the storm which have space scales ranging from single kilometers to hundreds of kilometers and time scales of minutes to hours. Moreover, the energy driving the storm is derived from fully three-dimensional motions on the meso-alpha to meso-beta scales almost entirely hidden beneath a thick layer of precipitation laden air. These structures include Vorticial Hot Towers (VHTs), Vortex Rossby waves, Spiral bands and the eye wall itself. Recent papers have shown the VHTs are critical to simulating genesis and even fluctuations in intensity of mature storms. These are essentially strong supercell-like individual cumulonimbi with time scales of hours and space scales of ~10-50 km.
Recent research and real time modeling investigations have clearly demonstrated the ability of numerical models to simulate these structures, sometimes anticipating their existence by virtue of the resolved flow systems, albeit with imperfect geometries and timing. Interestingly, the best successes have been 24-48 hours past the time that their impending signal was captured from the observation system, since it takes that long to contract these effects to the fine scale not resolved by observations. This is well past their predictability limit, however some success have been possible in a probabilistic sense since their existence and progression can be strongly controlled by more predictable features of the storm.
Real time observations on scales closer to the time and space scales of these mesoscale structures would greatly improve the accuracy of forecasts, but would require the assimilation of nearly continuous three dimensional observations taken uniformly within the cloudy air. Moreover, the important weather features tend to be near or below the Rossby radius of Deformation. This means that the controlling feature of the observation is the wind rather than the mass (temperature) field. Recent research into VHTs has confirmed that the forcing of the storm is via the wind field creating a vortex which becomes absorbed into the parent flow. At these scales, the tropical storm is microcosm of the equatorial tropics overall: It's the 3D WIND field that matters; we draw streamlines, not height fields!
Aircraft or UAVs could be a useful observation tool if only they could deploy Doppler radar monitor the complex wind fields of brewing and mature storms continuously on timescales of hours and space scales of tens of kilometers across the tropical latitudes where storm develop and move, but this is unlikely. A more reasonable solution is being developed in the form of a Doppler radar placed in geostationary orbit which is being called “NEXRAD in Space” (NIS). Just one of these radars will continuously take 3D observations of the wind and reflectivity fields over the east Pacific genesis zone, the Caribbean, the Gulf of Mexico and the western tropical Atlantic. The oral presentation will present the argument which demands that this technology, now nearly ready for implementation, be pursued to solve the intensity prediction problem.
Joint Poster Session 3, Tropical Cyclones and Climate Change Poster Session
Monday, 21 January 2008, 2:30 PM-4:00 PM, Exhibit Hall B
Previous paper Next paper