Extended Abstract (2.1M)1.4
The use of the Hydro-Nowcaster for Mesoscale Convective Systems and the Tropical Rainfall Nowcaster (TRaN) for landfalling tropical systems
Roderick A. Scofield, NOAA/NESDIS/ORA, Camp Springs, MD; and R. J. Kuligowski and J. C. Davenport
The Satellite Analysis Branch (SAB) of NESDIS provides satellite derived GOES (Geostationary Operational Environmental Satellite) Quantitative Precipitation Estimates (QPEs) for heavy rain or snow (including Lake Effect) over the contiguous U.S. and Puerto Rico. SAB's efforts concentrate primarily on locations where there is a potential for or occurrence of flash floods or heavy precipitation. The estimates are sent via the Advanced Weather Interactive Processing System (AWIPS) as part of the satellite-derived precipitation message (SPENES). The SPENES message also contains manually produced guidance on satellite analysis, trends, and short range forecasts (Nowcasts). SAB's Home Page that contains graphics products of these estimates is located at: " http://www.ssd.noaa.gov ". Heavy rainfall estimates from the Automatic Hydro- Estimator (H-E) have replaced many of the estimates computed interactively on the Interactive Flash Flood Analyzer (IFFA). An advantage of the H-E is its ability to vastly improve the spatial and temporal coverage of satellite precipitation estimates while improving timeliness. Hourly estimates from the H-E are available on AWIPS. The H-E has evolved into an automatic Hydro- Nowcaster (H-N) to complement and possibly eventually replace the above mentioned Nowcasts produced manually by SAB. Successful application of the H-N has the potential of increasing the lead time for some types of flash floods. However, this is not an easy task since one of the greatest challenges of an operational meteorologist is the short term prediction of the direction and speed of movement of Mesoscale Convective Systems (MCS). Propagation is the controlling influence on the movement of MCSs. At this time, the mechanisms of storm propagation are minimally understood. Any 3 hour Nowcast algorithm must take into account propagation characteristics of MCSs. Movement of MCSs involve the 0 - 3 hour movement/propagation of existing Meso-Beta Cores (MBC) of heavy rainfall. MBC are defined as the coldest cloud top clusters (as determined from the GOES 10.7 mm digital data) embedded within the MCS; MBC have spatial scales of 10 - 100 km and time scales of 1 - 10 hours. As a note, satellite-derived MBC may be relatively conservative features, especially as compared to the fast-changing rainfall cores detected in the radar. Thus, satellite-derived MBC may be more reliable to extrapolate in the short time frame of 0 - 3 hours. The H-N algorithm is a Storm - Mesoscale (Inside-Out) methodology where the speed and direction of the MBC is computed from two consecutive satellite images; preferably, the time between images should be 30 minutes or less. In the above calculation, a pattern correlation is used to determine cloud motions, from which a motion vector is assigned to every pixel in the image. The pattern correlation involves isolating the coldest 25 % pixels within a 25 pixel radius; in most cases such a per cent should capture automatically the MBC. Each pixel that initially had rainfall (as computed by the H-E) is then driven through the motion vector field for three hours and accumulated rain at each location tracked. The trends of the MBC are also incorporated to enhance the predicted rainfall during growth and diminish it during dissipation. A natural extension of the H-N algorithm is to landfalling hurricanes and tropical systems. Therefore, the H-N is being modified into a Tropical Rainfall Nowcaster (TRaN). TRaN will be tested for those tropical systems that are 6 hours or less from making landfall and after the system is over land. Home pages that contain automatic Flash flood estimates (around the world) and Nowcasts from H-N and TraN are on the NESDIS Flash Flood Home Page: "http://orbit35i.nesdis.noaa.gov/arad/ht/ff".
Session 1, Session 1: Continued
Monday, 12 January 2004, 10:45 AM-1:30 PM
Next paper
Recorded presentation