WMOD00595 - Laserfiche WebLink

Home Browse Search

6. APPLICATION OF A RADAR WIND PROFILER The radar wind profileI' was installed and operated near Blairsden tBL)J), which is located about 6 kIn northeast of Johnsville (fig 2.6) during the 1993-94 wimer, in the hope of obtaining near continuous vertical velocities oyer the valley between the dispensers and the target ridge. This continuous data collection would help document the wave cloud actiyity and how the frequency and magnitude of this ,"-ave affect crystal transport. At the time of installation, no "clean" site existed an:p,vhere in the valley to install the radar. Random noise can m;erwhelm the incoming signal to the profileI' and essentially contaminate its wind finding capability. Xoise is generated by wind blowing through the trees nearby or by moving telephone or power lines. The vertical velocities measured by the profileI' are much more sensitive to this noise than are the horizontal velocities. Unfortunately, almost all the vertical yelocity data measured by the profileI' were contaminated during non-precipitating eyents. During rain or snow, the signal backscattered from the precipitation to the radar is generally much larger than the background noise. The radar at this point is simply measuring the fall speed of the ice crystals, which is the combination of the terminal yelocity of the particle and the influence of the free air vertical motions. Because the Doppler spectra were recorded, it was hoped that a bimodal spectra would allow differentiation of the clear air vertical motions yersus the snowflake terminal yelocity. However, even in these situations, the turbulence associated with the air passing over the Sierra broadened the spectral width and eliminated the possibility of resoh-ing vertical motions. Even with the noise problems, much information was gained by looking at horizontal winds at high time resolution during seeding events. Three separate days will be examined. On these days, 50 pet of the randomized seeding cases that were predicted to have affected the target area occurred. A discussion of the randomized cases for 1993-94 is giyen in section 7.1. Figure 6.1 shows the hourly horizontal winds for January 24, 1994, as measured by the profileI'. ...-\11 heights are above m.s.l. (mean sea level) unless otherwise noted. Each hour, 52 wind samples were collected at 100-m vertical intervals. For a given layer to have an hourly wind computed, at least 50 pct of the 52 samples from each horizontal wind component must have a mean wind within 2 m S.l of each other. The figure shows that between 0000 CTC and 0600 eTC, periods of missing data were caused by ground clutter decreasing the signal to noise ratio. From 0600 to 1300, the antenna looking west is contaminated by clutter and only a southerly wind component is observed. Precipitation begins at 1400 GTC lO600 P.s.t.) as seen on figure 6.2. Station BOX is the closest gauge to the profileI'. The precipitation increases the depth over which winds can be calculated and increases the signal to noise ratio, allowing accurate winds to be computed. The westerly wind component is affected by clutter aboye about 5 kIn, and is obvious from the discontinuity in wind direction. Horizontal velocities are corrected lor vertical velocity contamination. This procedure basically subtracts the horizontal component of a particle's terminal yelocity from the U and V components. " i Figure 6.3 compares the hourly profiler winds with the rawinsonde winds for 1700 LTC lO900 P.s.t... It must be remembered in looking at this compalison that the rawinsonde is looking at the instantaneous winds, but the profileI' is averaging oyer 1 h. In general, the comparison in speed and direction is fairly good. The main differences are in the lowest 500 m and in the layer from 3000 to 4500 m~ The differences are most likely caused by the mountain lee waye. The profileI', located on the east side of the yalley, is seeing a return flow in the lowest 32