Quality Control Quality Control Problems For VAD Winds Problems For VAD Winds and NEXRAD Level-II and NEXRAD Level-II Winds In the Presence Winds In the Presence of Migrating Birds of Migrating Birds Li Bi 1 , Alan Shapiro 1,2 , Pengfei Zhang 3 and Qin Xu 3 1 School of Meteorology (SOM), University of Oklahoma,Norman,OK 2 Center for Analysis and Prediction of Storms (CAPS), Norman, OK 3 National Severe Storms Laboratory (NSSL), Norman, OK
20
Embed
Quality Control Problems For VAD Winds and NEXRAD Level-II Winds In the Presence of Migrating Birds Li Bi 1, Alan Shapiro 1,2, Pengfei Zhang 3 and Qin.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Quality Control Problems For Quality Control Problems For VAD Winds and NEXRAD VAD Winds and NEXRAD
Level-II Winds In the Level-II Winds In the Presence of Migrating BirdsPresence of Migrating Birds
Li Bi1, Alan Shapiro1,2, Pengfei Zhang3 and Qin Xu3
1School of Meteorology (SOM), University of Oklahoma,Norman,OK2Center for Analysis and Prediction of Storms (CAPS), Norman, OK
3National Severe Storms Laboratory (NSSL), Norman, OK
MotivationMotivationVAD began to be assimilated in Eta in 1997
and in Global model in 1998. VAD use stopped in Jan, 1999. After developing qc for VAD, VAD use was resumed on Mar 2000.
VAD winds contain several sources of errors. Birds flag is one significant source of qc error flag.
Want to minimize false alarms in the qc techniques.
Some previous work (radar Some previous work (radar ornithology)ornithology)
Radars can detect birds out to 200km.Peak bird traffic density occur before
midnight.Migratory bird usually fly with a tailwind.Peak reflectivity can reach 15-30dBz.Echo : Solid disk or doughnut.
Data for this projectData for this project
VAD product with qc flag from NCEP (v-wind exceeds forecast v-wind by 8 m/s.
6-sec sounding data from OUN (Norman).Level II WSR-88D radial velocity data from
nearby KTLX. (11.1km east, 14.8 km north of OUN)
Time period analyzedTime period analyzed
4 months of sounding data and radial wind data from 1 Feb 2001 to 31 May 2001.
16 cases of bird flag ( received from NCEP bird flag marked “5” within 15 mins of the twice-daily sounding time)
16 cases unflagged data within two days of a bird flag case.
MethodMethod
Project the sounding velocity along the balloon trajectory into the direction of the radar beam.
Calculate trajectory:
Trapezoidal rule for the integrals:
dtvyydtuxxt
t balloontt
t
t balloontt 2
112
2
112
,
2
1
2
1
2/)]()([)(,2/)]()([)( 2121
t
t
t
ttvtvdttvtutudttu
Method (continued)Method (continued)
Linearly interpolate the sounding data at the 2 nearest data points in the vertical.
Linearly interpolate the radial velocity data at the nearest 4 radar points to this intersection point.
A sample balloon trajectoryA sample balloon trajectory
An example of comparison of An example of comparison of Vr for bird flag casesVr for bird flag cases
Vr, 28 May 2001, 0010 UTC
0
500
1000
1500
2000
2500
3000
3500
4000
-30 -20 -10 0 10 20 30
Radial Velocity (m/s)
He
igh
t(m
)
radar
sounding
An example of comparison of An example of comparison of Vr for no flag casesVr for no flag cases
Vr, 10 May 2001, 0016 UTC
0
200
400
600
800
1000
1200
1400
1600
1800
-30 -20 -10 0 10 20 30
Radial Velocity (m/s)
Hei
gh
t (m
)
radar
sounding
Statistics Results Statistics Results
Rms error (difference between radial component of the sounding velocity and the radial wind)
Rms error over all 16 bird flag cases is 2.69m/s compare with 2.00m/s over all 16 unflagged cases.
Maximum error in the radial wind. 4.37m/s over all 16 bird flag cases compare with 3.75 over all unflagged cases.
RMS error in VrRMS error in Vr
Maximum error Maximum error
DiscussionDiscussion
Radar data in bird flag cases did had higher errors than in the unflagged cases (over 20%).
True bird contamination? We inspected the KTLX reflectivity data for
circumstantial evidence of birds (disks or annular rings of high reflectivity).
We checked the second elevation angle (1.450) to minimize ground clutter contamination.
Reflectivity field check for Reflectivity field check for bird-flag cases bird-flag cases
Reflectivity field check for Reflectivity field check for bird-flag cases (continued)bird-flag cases (continued)
Reflectivity field check for no Reflectivity field check for no flag cases (cross-wind)flag cases (cross-wind)
Reflectivity field check for no Reflectivity field check for no flag cases (two-disk)flag cases (two-disk)
ConclusionConclusion
For the 16 bird-flag cases, 10 cases didn’t have bird reflectivity signatures.
For the 16 unflagged cases, 10 cases didn’t have bird reflectivity signatures.
Results suggest current bird flag may be providing false birds alarms in as many as 50% of the cases.
Of the 16 unflagged cases, only 1 case of bird contamination eluded the qc check.There may have been 3 additional cases where bird were present but flying largely with the wind (not contaminating the data).