Radar Wind Profiler Data Quality Issues

1
Radar Wind Profiler Data Quality Issues

• AT741 Presented by Eric James 6 May 2008

2
Overview

• BACKGROUND
• 404 MHz Wind Profilers
• NOAA Profiler Network (NPN)
• DATA QUALITY ISSUES
• Effects of Precipitation
• Contamination by Ground Clutter
• Contamination by Gravity Waves
• Contamination by Migrating Birds
• CONCLUSIONS

3
404 MHz Wind Profilers

• Vertically-pointing radars operating at 404 MHz
  frequency (74.3 cm wavelength)
• 3 beams pointing upwards 1 vertically, 1 north
  at 73.7 degrees elevation, and 1 east at 73.7
  degrees elevation
• 40 by 40 square foot, fixed, phased-array antenna
• 250 m vertical gate spacing, with measurements
  from 0.5 km to 16.25 km above ground level (AGL)
• Hourly wind profiles, consisting of averages of
  sub-hourly (six-minute) observations
• Horizontal winds are derived by
• Determining the vertical velocity from the
  vertically-pointing beam
• Removing the component of this along the diagonal
  beams
• Dividing the remaining radial velocities by the
  cosine of the elevation angle (73.7 degrees)

4
The NOAA Profiler Network (NPN)

• Early profiling tests in a research environment
  involved systems at 10 MHz to 10 GHz frequency
• Operational network tests conducted in Colorado
  in 1980s
• NOAA Profiler Network (NPN) of 404 MHz profilers
  established in Midwestern United States in 1992
• Key requirements for NPN
• Capability for remote, unattended operation
• Frequent (hourly) observations to supplement
  radiosondes
• High vertical resolution (i.e., small range gate
  spacing)
• Dense network of sites
• NPN data valuable in operational forecasting
  environment, and for numerical weather prediction
  (NWP) data assimilation

5
The NOAA Profiler Network (NPN)

• 31 densely spaced sites in Midwestern U.S.A.,
  with core of network centred on Lamont, Oklahoma
  (ARM)
• 3 sites in Alaska, 1 in NY
• Ideal for studying synoptic and mesoscale
  phenomena in Great Plains
• Critical data for operational severe weather,
  aviation, and winter weather forecasting

6
Effects of Precipitation

• Fundamental assumption of the profiler
  observation strategy horizontal homogeneity
• Horizontally homogeneous precipitation causes no
  problems vertical velocity can be directly
  measured (by vertical beam) and removed
• Spatially varying precipitation (implying varying
  fall speeds) over the distance between beams
  (about 2.8 km at 10 km AGL) violates horizontal
  homogeneity assumption
• If taking more than one independent estimate of
  horizontal winds (such as by using 5 beams
  instead of 3), it is possible to flag data as
  erroneous if estimates do not agree
• Temporally varying precipitation requires removal
  of the correct vertical velocity at each time
  (for each sub-hourly observation)

7
Effects of Precipitation

• Secondary issue velocity aliasing
• Occurs when profiler maximum unambiguous velocity
  is too low
• Large hydrometeor fall speeds in presence of
  large horizontal winds can lead to folded radial
  velocities
• Horizontal winds derived from folded radial
  velocities will obviously be wrong
• Example shows velocity folding in derived
  vertical velocity near 2 km

Wuertz et al. 1988
8
Effects of Precipitation

• Yet another issue sidelobe contamination
• Figure shows ratio of precipitation reflectivity
  and clear air turbulence (Bragg scattering)
  reflectivity
• Ratio becomes huge at high frequencies (short
  wavelengths)
• Precipitation within sidelobes could contaminate
  main lobe observations of clear air

Balsley et al. 1982
9
Contamination by Ground Clutter

• Ground clutter is an issue for all radar systems
• Clutter contamination is reduced by the fact that
  the profiler beams point at very high elevation
  angles
• Clutter is increased by the fact that the most
  interesting data are often located in the first
  several range gates, where clutter can be
  appreciable
• Clutter fading (due to movement of the offending
  object) can cause ground clutter to have a
  nonzero spectral width in profiler data,
  corrupting the real Doppler spectrum

10
Contamination by Gravity Waves

• Gravity waves violate the fundamental horizontal
  homogeneity assumption
• Radar reflectivity directly proportional to
  static stability (due to concentration or
  dilution of radar scatterers)
• Gravity waves produce bias in vertical motion
  measurement due to inherent structure inversely
  related vertical motion and static stability
  (reflectivity)
• In this example, vertical velocity is biased
  negative in observations

Nastrom and Vanzandt 1996
11
Contamination by Migrating Birds

• Songbirds migrate between seasonal ranges in
  spring and autumn, usually at night, and at low
  levels (up to 4 km AGL)
• Since they are large, travel in flocks, and are
  full of water, they are readily detectable by
  radar
• Single bird in pulse volume produces trimodal
  velocity spectrum peak velocity due to bird
  speed, secondary peaks due to body motion during
  wing flapping
• Large number of birds produces high returned
  power and broad Doppler velocity spectrum
• Sometimes hard to detect (e.g., in LLJ)

Wilczak et al. 1995
12
Conclusions

• Profiler data useful, but significant
  uncertainties and errors exist, arising from
  combination of profiler measurement strategy and
  certain phenomena in the atmosphere
• Precipitation can violate horizontal homogeneity
  assumption, produce temporally-varying vertical
  velocities, cause velocity aliasing, or
  contaminate clear air data by its presence in
  sidelobes
• Ground clutter can contaminate low-level profiler
  data, particularly when clutter fading occurs
• Gravity waves can violate horizontal homogeneity
  assumption, and produce biases in vertical
  velocity measurements due to their inherent
  structure
• Migrating birds can produce unrealistic wind
  speeds and directions in low level profiler
  observations in certain locations, in certain
  seasons, and at certain times of day