Proceedings with Extended Abstracts (single PDF file) - Radio ...

The vertical component from the VTB hashigher time resolution (2.5 min.) than thatof the MVD (5 min.). The strong upwardairflow seen at 0100 JST in the tower measurementis caused by the passage of thegravity current front. The VTB methodcould, however, not observe this updraft,because clutter raised noise level around 0m s -1 in the Dopper velocity spectrum. TheMVD method, however, could capture thisupdraft.The spectrum in high mode shows thatthere were small droplets above 3 km after0140 JST. These droplets could be from acloud which produces rain from about 0300JST. The second trip echoes from the dropletswere observed in the spectrum of thelow mode, which masked the echo fromclear air as well as the clutter in the verticalbeam observation. The bias of the VTB seenafter 0140 JST in Fig. 2 could be attributableto the droplets. The MVD has, however,no such bias, because although therewere two peaks in the spectrum of obliquebeams, one from the atmosphere and theother from droplets, the two peaks were quitedifferent in the Doppler velocity. After 0300JST when rain was observed at the surface,however, the echo from the rain masked thatfrom clear air even in the oblique beams andmade it impossible to use the MVD method.Figure 3 indicates the vertical airflow measuredwith the tower versus that from theVTB and the MVD. The data used for theVTB are 0001 – 0137 JST, just before thesecond trip echo began to contaminate thereceived signal. The MVD tends to overestimatevertical airflow and the VTB tends tounderestimate. The correlation coefficientfor the MVD is greater than that of VTB,and bias of the MVD is less than that of theVTB. We concluded that the MVD has abetter accuracy than that of the VTB at lowaltitude. To confirm the reliability of theMVD at high altitude, we analyze the gravitycurrent.Vertical velocity [m s -1 ]Vertical velocity derivedfrom the profiler [m s -1 ]0.60.40.20-0.2-0.4-0.60000 0030 0100 0130 0200 0230 0300Time [JST]Fig. 2. Time series of vertical air motion derivedfrom the tower and the profiler from0000-0300 JST on 30 Dec. 1997.0.60.40.20.0-0.2-0.4VTB (210 m)MVD (203 m)Y v= 0.75X - 0.075, σ v= 0.15Y m= 1.33X - 0.021, σ m= 0.10Corr. σ NVTB 0.49 0.15 36MVD 0.77 0.10 36-0.6-0.6 -0.4 -0.2 0 0.2 0.4 0.6Vertical velocity measured with the tower at 200 m[m s -1 ]Fig. 3. Scatter diagram of vertical velocity measuredwith the tower vs that estimated from thewind profiler. The data for the VTB are collected0001 - 0137 JST and for the MVD are 0000 -0251 JST.600Tower (200m)VTB (210 m)MVD (203 m)3. Analysis of the gravity currentWind profiler with RASS is one of the besttools to visualize a gravity current, becauseit can observe not only wind but also the temperatureprofile. Figure 4 shows the timeheightcross section of horizontal wind andreal temperature derived from the tower (below250 m) and the wind profiler with RASS(above 250 m). Time is right to left so thatthe gravity current is moving from right toleft as it would on a map. The wind directionsat low altitude change from NE to NWwith temperature decrease at 0100 JST, whenthe front of the gravity current passed overHeight [m AGL]400200015 12 09Time [JST]Figure 4: Time-height cross section of wind vectorsand temperature derived from the profiler(above 250 m ) and the tower. Full barb is 5 m s -1 .060300299