WMOD00561 - Laserfiche WebLink

Home Browse Search

Several researchers have discussed possible future advances in radar precipitation estimation with multi parameter and dual polarimetric methods. Matrosov (1998) presents evidence that use of dual- wavelength radars should provide improved snowfall estimates. However, the purpose of the SAA development has been to develop a practical means of Sand SD estimation with the existing WSR-88D radar. Future technological improvements in the WSR-88D may permit the SAA to be improved. However, the philosophy behind the SAA development has been, and will continue to be, to provide a practical means of converting measurements supplied by existing WSR-88D capabilities into useful S and SD estimates. It is planned to add the SAA to the standard suite of WSR-88D algorithms within the next 3 years (O'Bannon, 1998), after existing computers are replaced with more powerful units capable of running additional algorithms. The SAA has some major limitations, caused by the physics of the situation, which will be discussed. Many of these are shared with rain algorithms, especially when shallow stratiform rain exists. · The SAA is designed to be limited to estimation of dry snowfall only. The complication of dealing with mixed rain and snow and/or melting snow with associated "bright band" effects was beyond the scope of the requested work and available resources as stated in Reclamation's October 1994 proposal made part of the MOU. The SAA is limited by the vertical departure of the radar beam from the surface, caused by earth's curvature. The current SAA version cannot distinguish between snow reaching the surface or virga at far range. Both are often detected only in the lowest beam tilt. The SAA does not calculate the vertical profile of snowfall which could be used for range correction. However, the option exists to apply a correction as a function of range. Additional study will be required to provide appropriate range correction equation constants for various regions or individual radars. · The SAA is limited by the ever changing relationship, in space and time, between what a radar measures ("reflectivity") and falling snow particles. A radar does not measure precipitation but reflectivity which can be related to precipitation. Since the relationship between reflectivity and snowfall is ever changing in time and space as snow particle characteristics (Ohtake and Henmi, 1970) and other factors change, a practical average relation must be sought that provides adequate S estimates most of the time. Moreover, the radar is not directly measuring what is happening at ground level beneath each radar bin. Rather, because of earth's curvature, the radar beam scans at ever increasing altitudes as the range (distance) from the radar increases. Because snow particles typically continue to grow as they fall to the surface, scanning well above the surface measures weak reflectivities compared with near-surface values. The result is the relationship between Ze' as measured by the radar, and S, observed at the surface, also changes with range. Consequently, a radar will tend to underestimate S at moderate to far ranges unless a range correction scheme is employed. It should be understood that a radar can provide only estimates of snowfall because of these and other complexities to be discussed. However, it will be shown that useful estimates can be provided. Moreover, a radar offers the important advantage of frequent sampling of large spatially-continuous regions, impractical to achieve with surface instrumentation. The present version of the SAA provides useful products, but the SAA can and should be improved upon. The primary improvement would be to deal with known underestimation at increasing range. Reclamation submitted a proposal to the OSF at their request during January 1997 to pursue a solution to this serious problem. However, the proposed work was not funded. Apparently, the need for a range 2