Refractivity Measurements from Ground Clutter Using the National Weather Radar Testbed Phased Array

1
Refractivity Measurements from Ground Clutter
Using the National Weather Radar Testbed Phased
Array Radar

• B. L. Cheong1, R. D. Palmer1, C. Curtis2,3, T.-Y.
  Yu4, D. Zrnic3 and D. Forsyth3
• 1School of Meteorology, The University of
  Oklahoma, Norman, USA
• 2Cooperative Institute of Mesoscale
  Meteorological Studies (CIMMS), Norman, USA
• 3NOAA/OAR National Severe Storms Laboratory,
  Norman, USA
• 4School of Electrical and Computer Engineering,
  The University of Oklahoma, Norman, USA

Supported by NOAA/NSSL under cooperative
agreement NA17RJ1227
http//arrc.ou.edu/
2
Overview

• Introduction to the National Weather Radar
  Testbed in Norman, Oklahoma
• Theoretical Background of Refractivity Retrieval
  From Clutter and Overview of OUs Refractivity
  Retrieval Algorithm
• Initial Measurements Using the Phased Array Radar
  (PAR) and Study of Multi-Mission Capability

3
National Weather Radar Testbed NWRT

• Centerpiece of NWRT is the Phased Array Radar
  (PAR), which is an S-band, klystron-based radar
  using a SPY-1A phased array antenna
• NOAAs National Severe Storms Laboratory (NSSL)
  operates the PAR on the North Campus of the
  University of Oklahoma
• Major goal is to improve severe storm warning
  lead time using phased array radar
• Provide rapid update data for assimilation into
  numerical models
• The NWRT provides a location for
  testing/validation of advanced weather radar
  processing algorithms
• Ultimately, the goal is to perform multi-mission
  tasks including weather surveillance (including
  refractivity), aircraft tracking, etc.

4
Phased Array Radar (PAR)
USAs first research facility dedicated to phased
array radar meteorology
5
Refractivity and EM Waves

• Refractive index n
• Near the earth surface, n 1.0003
• Refractive index is translated into radar phase
• Radar phase from stationary ground clutter
  (constant r) should be constant if refractivity
  remains unchanged
• Refractivity

6
Refractivity, Moisture, and Temperature

• Refractivity
• p air pressure
• T air temperature
• e vapor pressure

• N changes dominated by e

7
Example Phase Measurements

• Can we measure refractive index by using radar
  phase?
• Use phase difference from multiple ground targets
  along same radial for range resolution.
• For S-band radars, the phase wraps very quickly
  with increasing range (every 5cm!)

8
Overview of OUs Refractivity Retrieval Algorithm
Phase measurement for a map of reference phase
Phase measurement during operational time
A map of Phase Difference ??
Image Processing Clutter Quality, Masking,
Smoothing
Radial gradient ? ?N
9
Refractivity During a Dust StormSept 28, 2005,
1859-1920 UTC
10
Comparison Between KOUN and PAR
PAR Refractivity
KOUN Refractivity
Different Radars and Different Algorithms
11
Oklahoma MesonetRefractivity and Refractivity
Change
12
PAR/Mesonet Comparison
5-minute temporal sampling
13
Qualitative Analysis of Shorter Dwell Times

• Due to high SNR of ground clutter, even a
  2-sample dwell produces reasonable results!

14
Statistical Comparison to 64-Point Dwell

• For a 2-sample dwell, the RMS error from the
  reference is approximately 1 N-unit

15
Conclusions

• Provided introduction to the PAR and the NWRT in
  Norman, Oklahoma, USA
• Reviewed theory of the measurement of surface
  refractivity (moisture) from ground clutter
  signals
• Described processing steps necessary for
  implementation
• Results from PAR provided and compared to
  near-operational KOUN radar and Mesonet surface
  stations
• 2-sample dwell has been shown to be sufficient
  for refractivity retrieval using the PAR,
  allowing the possibility for multi-mission
  operation

16
BACKUP SLIDES
17
Phase Difference to Refractivity

• With ?ref and nref measured from a day with
  homogeneous refractivity

Fabry et al., On the extraction of near-surface
index of refraction using radar phase
measurements from ground targets, JTech, 14,
978-987, 1997
18
Experiment Results Interpretation

• The first of two case studies
• 28 September 2005
• - Forecasts from the previous day indicated that
  a cold front was likely to advance SE through
  central Oklahoma during the afternoon of 28
  September 2005,
• - Surface winds from the NNW accompanied the
  cold front, resulting in a major dust storm for
  the area
• - KOUN was operated from approximately1900 -
  1930 UTC
• - PAR simultaneously collected data, directed to
  the North of NSSL

19
28 September 2005 Experiment
Unisys

• 1200 UTC, 28 September
• 300 mb

20
28 September 2005 Experiment
Unisys

• 1200 UTC, 28 September
• 850 mb

21
28 September 2005 Experiment
Unisys

• 0000 UTC, 29 September
• 300 mb

22
28 September 2005 Experiment
Unisys

• 0000 UTC, 29 September
• 850 mb

23
28 September 2005 Experiment
Unisys

• 0000 UTC, 29 September
• Surface analysis

24
28 September 2005 Experiment

• - Winds in Norman shift to the NNW at approx.
  1820 UTC
• - Td and N gradient located along the wind
  shift, but T gradient lags the wind shift by 30
  km
• - Between 1900 and 1930 UTC, when KOUN is
  operated, N field appears to largely be a result
  of the Td gradient

25
28 September 2005 Experiment

• - ?N from beginning of the experiment
• - By 1935 UTC, we see what appears to be a
  wave-like structure of /- ?N

26
28 September 2005 Experiment

• KOUN comparison with Oklahoma Mesonet
• - General trend in KOUN ?N field (SE to NW) is
  supported by surface station observations
• - Wave-like pattern of ?N behind the front -
  this is not resolved by the Oklahoma Mesonet

KOUN Oklahoma Mesonet
27
PAR image provided by Boon Leng Cheong
28 September 2005 Experiment

• KOUN comparison with PAR
• - At 1935 UTC, features present in ?N field are
  confirmed by PAR observations
• - Is there a meteorological significance to the
  wave-like ?N pattern observed by KOUN?

KOUN PAR
28
28 September 2005 Experiment
Need to develop a method to measure the wave
structure contained within the ?N field Sample
?N along slices, oriented perpendicular to the
position of the frontal boundary and parallel to
the mean wind (from SE to NW)

• Diagram of slices sampled through ?N
• field for 1935 UTC, 28 September 2005

29
28 September 2005 Experiment
After eliminating outliers, use linear regression
to determine slope (green) Remove slope from
5-km running-mean mean (red) to isolate
sinusoidal ?N signal (black, bottom) Fourier
Analysis determines wavelength of sinusoid
(approx. 23 km)
30
28 September 2005 Experiment - Summary

• Experiment Summary
• Refractive index retrieval using KOUN has allowed
  us to observe small-scale gravity waves
  associated with a cold front.
• Features observed in the ?N field on 28 September
  are supported by simultaneous data recorded with
  the PAR, as well as Oklahoma Mesonet surface
  station observations.

31
Overview

• Introduction to the National Weather Radar
  Testbed in Norman, Oklahoma
• Theoretical Background of Refractivity Retrieval
  From Clutter and Overview of OUs Refractivity
  Retrieval Algorithm
• Initial Measurements Using the Phased Array Radar
  (PAR) and Validation Using Scanning Radar and
  Surface Stations

32
Phased Array Radar (PAR) Design

• Passive array of 4,352 elements
• S-band transmitter (3.2 GHz)
• 1.5-2.1 degree beamwidth over 45 degrees

33
Overview

• Introduction to the National Weather Radar
  Testbed in Norman, Oklahoma
• Theoretical Background of Refractivity Retrieval
  From Clutter and Overview of OUs Refractivity
  Retrieval Algorithm
• Initial Measurements Using the Phased Array Radar
  (PAR) and Validation Using Scanning Radar and
  Surface Stations

34
Overview

• Introduction to the National Weather Radar
  Testbed in Norman, Oklahoma
• Theoretical Background of Refractivity Retrieval
  From Clutter and Overview of OUs Refractivity
  Retrieval Algorithm
• Initial Measurements Using the Phased Array Radar
  (PAR) and Validation Using Scanning Radar and
  Surface Stations