ARRCs Polarimetric Xband Radar

1
Progress Report March 2009 Advances in Phased
Array Weather Radar Research at the University of
Oklahoma R. Palmer, T. Yu, G. Zhang, M. Yeary,
P. Chilson, Y. Zhang, J. Crain
2
Research Topics

• Clutter Mitigation Using Auxiliary Elements for
  the NWRT Phased Array Radar OU POC Bob Palmer
• Real-time Rapid Refractivity Retrieval Using the
  National Weather Radar Testbed Phased Array Radar
  OU POC Bob Palmer
• Experimental Studies and Knowledge Base
  Development for Mixed-phase Hydrometeor OU POC
  Rockee Zhang
• Knowledge-Based Adaptive Sensing Scheduling
  Multi-Task Operations on a Phased Array Radar
  OU POC Tian-You Yu
• Convective Storm Cell Clustering and Tracking for
  Short-Time Forecasting OU POC Mark Yeary and
  Tian-You Yu
• Multi-Beam/Pattern Experiments for the Phased
  Array Radar Calibration and Side-lobe Reduction
  OU POC Guifu Zhang
• Adaptive scanning and PAR OU POC Phil Chilson

3
Clutter Mitigation Using Auxiliary Elements for
the NWRT Phased Array Radar

• Khoi Le1,2, Robert Palmer2,3, Boon Leng Cheong2,
  Tian-You Yu1,2, G. Zhang2,3, S. M. Torres4,5
• 1School of Electrical Computer Engineering,
  University of Oklahoma, USA
• 2Atmospheric Radar Research Center (ARRC),
  University of Oklahoma, USA
• 3School of Meteorology, University of Oklahoma,
  USA
• 4Cooperative Institute of Mesoscale
  Meteorological Studies (CIMMS), USA
• 5NOAA/OAR National Severe Storms Laboratory, USA

4
Advantageous of Phased Array Radars for Weather
Observations

• Multi-mission Phased array radar (MPAR) could
  provide simultaneous air traffic control and
  weather observations
• Rapid Scan Phased array radars are capable of
  electronically steering. With an adaptive
  scanning strategy, such as beam multiplexing
  (BMX), these systems can provide update scans as
  fast as 1min.
• BMX uses very short dwell times (2 pulses), which
  are problematic for clutter filtering based on
  temporal filters

5
NWRT Phased Array Model

• The configuration of the NWRT with its six
  auxiliary elements is modeled to study the ground
  clutter mitigation capability of phased array
  radars in weather observation

6
Example of Phase Steering and Sidelobe Cancelling
7
Comparison of Contaminated and Retrieved Fields
to Original Power and Radial Velocity Fields
Reflectivity
Radial Velocity
8
Simulation Results Sample Size Effects
9
Research Topics

• Clutter Mitigation Using Auxiliary Elements for
  the NWRT Phased Array Radar OU POC Bob Palmer
• Real-time Rapid Refractivity Retrieval Using the
  National Weather Radar Testbed Phased Array Radar
  OU POC Bob Palmer
• Experimental Studies and Knowledge Base
  Development for Mixed-phase Hydrometeor OU POC
  Rockee Zhang
• Knowledge-Based Adaptive Sensing Scheduling
  Multi-Task Operations on a Phased Array Radar
  OU POC Tian-You Yu
• Convective Storm Cell Clustering and Tracking for
  Short-Time Forecasting OU POC Mark Yeary and
  Tian-You Yu
• Multi-Beam/Pattern Experiments for the Phased
  Array Radar Calibration and Side-lobe Reduction
  OU POC Guifu Zhang
• Adaptive scanning and PAR OU POC Phil Chilson

10
Real-time Rapid Refractivity Retrieval Using the
National Weather Radar Testbed Phased Array Radar

• Boon Leng Cheong1, Robert Palmer1,2, Christopher
  Curtis3,4, Tian-You Yu1,5, Dusan Zrnic4, Douglas
  Forsyth4
• 1Atmospheric Radar Research Center (ARRC),
  University of Oklahoma, USA
• 2School of Meteorology, University of Oklahoma,
  USA
• 3Cooperative Institute of Mesoscale
  Meteorological Studies (CIMMS), USA
• 4NOAA/OAR National Severe Storms Laboratory, USA
• 5School of Electrical Computer Engineering,
  University of Oklahoma, USA

11
Refractive Index and Radar Phase

• Radar phase as a function of refractive index
• Change of phase and change of refractive index

Fabry, Meteorological Value of Ground Target
Measurements by Radar, JTECH, 21, 2004
12
Experimental Setup

• Continuous 90 sector scans over 45, EL 0.5
• PRT of 1 ms, 64 consecutive I/Q samples for each
  radial
• Aliasing velocity 23.4 ms-1
• Temporal resolution 5.76 s
• Refractivity reference is set at the start of the
  experiment
• A 1-hour data set was collected. During this
  period, there was a strong low-level northerly
  wind (8 ms-1) causing a light dust storm.
• Subset of the data were extracted to simulate
  shorter dwell periods of 2, 4, 8, 16 and 32
  samples.

13
Comparison Between KOUN and PAR
PAR Refractivity
KOUN Refractivity
Different Radars and Different Algorithms
14
Validation with the Oklahoma Mesonet

• 5-minute temporal sampling

15
Quantitative Analysis of Shorter Dwell Times

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

16
Statistical Comparison to 64-point Dwell

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

17
Research Topics

• Clutter Mitigation Using Auxiliary Elements for
  the NWRT Phased Array Radar OU POC Bob Palmer
• Real-time Rapid Refractivity Retrieval Using the
  National Weather Radar Testbed Phased Array Radar
  OU POC Bob Palmer
• Experimental Studies and Knowledge Base
  Development for Mixed-phase Hydrometeor OU POC
  Rockee Zhang
• Knowledge-Based Adaptive Sensing Scheduling
  Multi-Task Operations on a Phased Array Radar
  OU POC Tian-You Yu
• Convective Storm Cell Clustering and Tracking for
  Short-Time Forecasting OU POC Mark Yeary and
  Tian-You Yu
• Multi-Beam/Pattern Experiments for the Phased
  Array Radar Calibration and Side-lobe Reduction
  OU POC Guifu Zhang
• Adaptive scanning and PAR OU POC Phil Chilson

18
Experimental Studies and Knowledge
Base Development for Mixed-phase Hydrometeor
Rockee Yan Zhang1,3, Guifu Zhang2,3 Students
supported Zhengzheng Li and Andrew Huston 1
School of Electrical and Computer Engineering 2
School of Meteorology 3 Atmospheric Radar
Research Center
A progress report to NOAA-NSSL, period April 08
April 09
19
Key Milestones Achieved
Improved EML chamber facility at 1PP 10 dB
reduced environment clutter Improvement on
highly-sensitive scatterometer system for
dual-polarized radar signature studies.
Compared melting hydrometeor lab scattering
measurement with different melting models
Mie-coated spheres and fractional melting
model Established new T-matrix based knowledge
derivation technology for future Knowledge-aided
algorithms Toward experiment setup for
distributed volume scattering and polarimetric
Phased array study
20
Publication and Discloser
Journal Yan Zhang, Andrew Huston, Michael
Mallo, Zhengzheng Li and Guifu Zhang, A
Scatterometer System for Laboratory Study of
Polarimetric Electromagnetic Signatures of Icy
Hydrometeors, IEEE Transactions on
Instrumentation and Measurement, in
press. Conference paper Andrew Huston, Yan
Zhang, Guifu Zhang, Mark Yeary and Robert T.
Neece, A Laboratory Study of Dual-Polarization
Scattering Characterizations for Meteorological
Objects, I²MTC 2008 IEEE International
Instrumentation and Measurement Technology
Conference, Victoria, Vancouver Island, Canada,
May 1215, 2008 Yan Zhang, Robert Palmer, Guifu
Zhang, Tian-You Yu, Keith Brewster, Mark Yeary,
Ming Xue, Phillip Chilson, "Multi-functional
Airborne External Hazard Monitoring Radar with
Antenna Diversity", SPIE Remote Sensing
Applications for Aviation Weather Hazard
Detection and Decision Support conference, San
Diego, California, Aug 10-14, 2008 Others Relate
d research was also presented on 2008 Aviation
safety conference and several other Project
meetings/seminars
21
Highlight 1 scatterometer system and measurement
Re-furbished and Re-calibrated chamber
environment shows significant Improvement on
clutter suppression and RCS Measurement
accuracy. Dual-polarized RCS measurement for
natural Hailstone samples is performed and
compared With man-made hydrometeors. Strong
difference In ZDR and LDR can be used for
classifications
22
Highlight 2 Theoretical models and lab
measurement
LEFT RCS prediction from coated
sphere-model (layered Mie-model), with different
thickness of Melting layer Bottom comparing
RCS measurement of 3 Wet ice sphere at X-band
with fractional-volume Model prediction.
Conclusion Fractional volume model With
appropriate Mixing procedure Is the best tool
23
Highlight 3 Knowledge development
Knowledge obtained from Laboratory studies has
been applied To predict the performance And
radar signatures of a future airborne
polarimetric array radar. A new scattering
knowledge model is being studied aiming at the
particular radar system application.
New approach of measuring other Dual-pol
variables for volume scattering case is being
formulated. T-matrix and fractional volume
model is used For mixed phased hydrometeor at
both S and X bands. Based on lab-measurement-vali
dated scattering model, we are able to establish
a realistic scattering model relating the storm
microphysics to the dual-pol scattering
parameters leading to a better way for
polarimetric radar signal modeling.
24
Research Topics

• Clutter Mitigation Using Auxiliary Elements for
  the NWRT Phased Array Radar OU POC Bob Palmer
• Real-time Rapid Refractivity Retrieval Using the
  National Weather Radar Testbed Phased Array Radar
  OU POC Bob Palmer
• Experimental Studies and Knowledge Base
  Development for Mixed-phase Hydrometeor OU POC
  Rockee Zhang
• Knowledge-Based Adaptive Sensing Scheduling
  Multi-Task Operations on a Phased Array Radar
  OU POC Tian-You Yu
• Convective Storm Cell Clustering and Tracking for
  Short-Time Forecasting OU POC Mark Yeary and
  Tian-You Yu
• Multi-Beam/Pattern Experiments for the Phased
  Array Radar Calibration and Side-lobe Reduction
  OU POC Guifu Zhang
• Adaptive scanning and PAR OU POC Phil Chilson

25
Knowledge-Based Adaptive Sensing Scheduling
Multi-Task Operations on a Phased Array Radar
Time Balance

• Tian-You Yu

26
Scheduling Multi-task Time Balance (TB)
Inputs parameter for track surveillance
Is it a new cell?
yes
no
Adjust track parameter
Find tracks TB gt 0
If empty
yes
no
Schedule surveillance
Choose track with maximum TB / max occupancy
Schedule this cell
Decrement its TB by Li
Increment all TB by dwell time of scheduled
function
27
Demo of TB to Schedule Multi-task
Tracking two cells and surveillance
Tasks requested
Tasks scheduled
28
Quality Measure I Improvement Factor
29
Quality Measure II Average Frame Time
Frame Time The minimum time period for each task
is executed at least once
30
Summary

• PAR is capable of performing surveillance and
  tacking of multiple storm cells independently and
  adaptively. The concept of Time Balance was
  introduced to schedule these tasking that are
  competing for radar resources.
• Two quality measures were introduced Improvement
  factor and Frame time (can be optimized
  independently based on users need). The
  trade-off between these two measurements were
  demonstrated by both theory and simulations.
• Scheduling multiple tasks for adaptive sensing
  was demonstrated using interpolated WSR-88D data.
• Results suggest that the improvement factor and
  frame time can be improved using a small number
  of samples for surveillance and implementing beam
  multiplexing (BMX) for tracking.

31
Research Topics

• Clutter Mitigation Using Auxiliary Elements for
  the NWRT Phased Array Radar OU POC Bob Palmer
• Real-time Rapid Refractivity Retrieval Using the
  National Weather Radar Testbed Phased Array Radar
  OU POC Bob Palmer
• Experimental Studies and Knowledge Base
  Development for Mixed-phase Hydrometeor OU POC
  Rockee Zhang
• Knowledge-Based Adaptive Sensing Scheduling
  Multi-Task Operations on a Phased Array Radar
  OU POC Tian-You Yu
• Convective Storm Cell Clustering and Tracking for
  Short-Time Forecasting OU POC Mark Yeary and
  Tian-You Yu
• Multi-Beam/Pattern Experiments for the Phased
  Array Radar Calibration and Side-lobe Reduction
  OU POC Guifu Zhang
• Adaptive scanning and PAR OU POC Phil Chilson

32
Convective Storm Cell Clustering and Tracking for
Short-Time Forecasting

• Mark Yeary and Tian-You Yu

33
Convective Storm Cell Clustering and Tracking for
Short-Time Forecasting

• The goal of this research was to develop
  convective storm cell clustering and tracking
  algorithms that will provide short-time forecasts
  at the NWRT. Developments will also continue in
  future studies. The technologies here are
  similar to the Storm Cell Identification
  Tracking (SCIT) algorithms that were developed by
  the NSSL one decade in the past for the WSR-88D,
  but the new techniques here are greatly improved
  and are being developed specifically for the
  phased array radar at the NWRT.
• The SCIT is the algorithm that is used
  operationally for the WSR-88D radars to identify
  and track storm cells. The algorithm identifies
  storm cells by finding contiguous data-points
  that meet empirical reflectivity thresholds.
• The focus of the current research was to develop
  a replacement for the SCIT storm-cell
  identification algorithm. The teams algorithm
  utilizes a new clustering method, known as the
  teams Strong Point Analysis (SPA) algorithm, to
  find reflectivity features. SPA finds features by
  first identifying statistically relevant
  data-points (the strong points).

34
Properties of Strong Point Analysis

• Image Clustering Algorithm
• Consistent
• Identifies similar features in similar images
• Necessary for tracking features in time-sequence
• Stable
• Results vary smoothly with input parameters
• Results tunable to desired level of detail

35
Strong Point Analysis Demonstration
Identified Clusters
36
Input Image Varying Input Parameters
Constant

• June 5, 2008
• MPAR
• 2342Z to0000Z

37
Input Image Varying Input Parameters
Constant

• February 10,2009
• MPAR
• 2000Z to2050Z
• Same parametersas previous

38
Input Image Constant Input Parameters
Varying
Subcluster Level

• July 8,2004
• KDDC(DodgeCity)?
• 0028Z

Upper Sensitivity
39
Research Topics

• Clutter Mitigation Using Auxiliary Elements for
  the NWRT Phased Array Radar OU POC Bob Palmer
• Real-time Rapid Refractivity Retrieval Using the
  National Weather Radar Testbed Phased Array Radar
  OU POC Bob Palmer
• Experimental Studies and Knowledge Base
  Development for Mixed-phase Hydrometeor OU POC
  Rockee Zhang
• Knowledge-Based Adaptive Sensing Scheduling
  Multi-Task Operations on a Phased Array Radar
  OU POC Tian-You Yu
• Convective Storm Cell Clustering and Tracking for
  Short-Time Forecasting OU POC Mark Yeary and
  Tian-You Yu
• Multi-Beam/Pattern Experiments for the Phased
  Array Radar Calibration and Side-lobe Reduction
  OU POC Guifu Zhang
• Adaptive scanning and PAR OU POC Phil Chilson

40
Multi-Beam/Pattern Experiments for the Phased
Array Radar Calibration and Side-lobe Reduction
Guifu Zhang1, Yinguang Li1, Richard J. Doviak2,
John Carter2 and Dave Priegnitz2 1 University
of Oklahoma, Norman, OK 73072 2 National Severe
Storms Laboratory, NOAA, Norman, OK 73072
41
Outline

• National Weather Radar Testbed/Phased Array Radar
  (NWRT/PAR)
• Unique capability of dual-scan (mechanical and
  electronic)
• Multi-beam/pattern experiments
• Clutter reduction
• Power calibration
• Summary

Courtesy of A. Zahrai
42
Multi-pattern Idea originated from antenna
pattern measurements

• Beam width different
• Side-lobe location different
• Further info can be obtained

43
Extended radar equation

• Beam at boresight and off boresight
• Radar equation
• Calibration

44
Multiple patterns Test with simulations

• Maximize the difference in the
• patterns
• Cost function
• Optimal beams 0, 25, 34, 41, 45
• 5 dB reduction in side-lobe

45
A Isolated Convection Case Reflectivity field
obtained with multi-pattern
0 (deg)
34 (deg)
25 (deg)
41 (deg)
45 (deg)
46
Notch Filter and Multi-Pattern Processing
Relative std 0.2
47
Conclusions and Discussions

• PAR measurements with multi-beam/patterns (MPs)
  can reduce side-lobe effects
• MPs can remove both stationary and moving clutter
• MPs preserves weather echo better than a notch
  filter
• Side-lobe reduction is enhanced with optimized
  non-uniform beam separation (0, 25, 34, 41, 45)
• Power calibration with
• Antenna area projection
• Reasonable

48
Research Topics

• Clutter Mitigation Using Auxiliary Elements for
  the NWRT Phased Array Radar OU POC Bob Palmer
• Real-time Rapid Refractivity Retrieval Using the
  National Weather Radar Testbed Phased Array Radar
  OU POC Bob Palmer
• Experimental Studies and Knowledge Base
  Development for Mixed-phase Hydrometeor OU POC
  Rockee Zhang
• Knowledge-Based Adaptive Sensing Scheduling
  Multi-Task Operations on a Phased Array Radar
  OU POC Tian-You Yu
• Convective Storm Cell Clustering and Tracking for
  Short-Time Forecasting OU POC Mark Yeary and
  Tian-You Yu
• Multi-Beam/Pattern Experiments for the Phased
  Array Radar Calibration and Side-lobe Reduction
  OU POC Guifu Zhang
• Adaptive scanning and PAR OU POC Phil Chilson

49
Adaptive Scanning and PARPhil Chilson

• PAR has the capability to perform digital beam
  steering, which allows for new techniques like
  adaptive scanning.
• Algorithms will be designed to scan targets based
  on priority or interest
• Higher priority targets will be scanned with
  greater temporal or spatial resolution
• These algorithms should help improve detection
  and analysis of tornadoes, severe thunderstorms
  and other phenomena

An example adaptive scanning method for a
tornadic supercell. The high priority targets
(red lines) would be scanned at a much higher
temporal and spatial resolution than the low
priority targets (green lines).
50
Optimization parameters for developing adaptive
scans

• Temporal resolution
• What update rate is necessary to fully track
  storm evolution?
• Azimuthal resolution
• What resolution do we need in order to better
  detect small-scale features like TVSs and
  mesocyclone signatures?
• Elevation angles
• Which elevations are the most critical during a
  particular event? What is the ideal elevation
  spacing?
• How do we determine the scanning rate for each
  elevation?
• Data accuracy
• What pulse rate is necessary to ensure that PAR
  matches the accuracy of the WSR-88Ds?
• How do we balance data accuracy with a rapid
  update rate?

51
Current Focus

• Test possible adaptive scanning techniques during
  Winter and Spring 2009
• Example A winter precipitation event sampled by
  PAR on 26-27 Jan 2009
• Use a surveillance scan to detect a target, then
  select the intensive scan based on target range.
• Repeat at regular intervals to adapt to target
  movement.

r lt 50 km
Intensive scans
50 lt r lt 100 km
Surveillance scan
52
New Developments

• A radar simulator is being used to test scanning
  strategy requirements in a controlled environment
• Manipulate data to determine the effects of beam
  oversampling, modifying the number of pulses, or
  adjusting temporal scan rates.
• What are the minimum settings that will allow us
  to obtain reliable data?
• Find an ideal combination of settings that
  satisfies accuracy requirements while allowing
  for rapid update times.

Reference Cheong, B. L., R. D. Palmer and M.
Xue, 2008 A time series weather radar simulator
based on high-resolution atmospheric models. J.
Atmos. Ocean. Tech., 25, 230-243.
53
Next Steps

• Examine other data sources in order to find
  possible improvements to current scanning methods
• Use the new OU-PRIME to provide a comparison with
  PAR.
• What information can OU-PRIME provide that PAR
  cannot currently obtain?
• How can we modify our PAR scanning strategies to
  acquire the missing information?
• Can surface observations or model fields help us
  determine target priority?
• Overlays of radar returns, RUC model results and
  surface observations could help us determine the
  most important targets.
• Can we develop a target priority index that can
  be applied in an adaptive strategy?