1. Introduction to the Dual-Polarized WSR-88D
Don Burgess
OU CIMMS/NSSL (Ret.)
Storm-Scale Data Assimilation Workshop
October 2011

2. Explaining Dual-Polarization
Dual-polarization radars emit EM waves with horizontal and vertical polarizations.
- Alternating H & V Transmission requires an expensive fast switch and longer acquisition times

3. WSR-88D Dual-Polarization Upgrade
Simultaneous Transmission And Reception (STAR); Slant 45
Transmit at 45o, receive at both horizontal and vertical
There is a PROBLEM!

4. WSR-88D Dual-Polarization Upgrade
Simultaneous Transmission And Reception (STAR); Slant 45
Transmit at 45o, receive at both horizontal and vertical
There is a PROBLEM! Split the power; 3-dB sensitivity loss

5. List of New WSR-88D Dual-Pol Outputs
3 New Variables (like moments)
Differential Reflectivity (ZDR; Zdr)
Correlation Coefficient (CC; Rhv)
Specific Differential Phase (KDP; Kdp)
3 New Algorithms
Melting Layer
Hydrometeor Classification
QPE
9 NEW Precipitation Estimation Display Products

6. New Product #1: Differential Reflectivity (ZDR)
Definition
Possible Range of Values
Units
Abbreviated Name
Measure of the log of the ratio of the horizontal to vertical power returns
-4 to 10
Decibels (dB)
ZDR
The differential reflectivity is a measure of the log of the ratio of the horizontal to vertical power returns in a pulse volume. Its values range from -4 to 8 in units of decibels (dB). In AWIPS and the literature, differential reflectivity is abbreviated as ZDR.
Horizontal Reflectivity
Vertical Reflectivity

7. ZDR Physical Interpretation
Spherical
(drizzle, small hail, etc.)
Horizontally Oriented
(rain, melting hail, etc.)
Vertically Oriented
(i.e. vertically oriented ice crystals)
ZDR ~ 0 dB
ZDR > 0 dB
ZDR < 0 dB Pv Pv Pv Ph Ph Ph
Zh ~ Zv
Zh > Zv
Zh < Zv
This chart here summarizes the general physical interpretation of differential reflectivity. For spherical hydrometeors such as drizzle drops, the power returned from both the horizontal and vertical pulses are approximately equal. This leads to a ratio of horizontal to vertical powers of approximately 1. The log of this ratio is therefore zero, which means ZDR is approximately 0 dB. The same logic can be applied for the horizontally and vertically oriented columns. Horizontally oriented hydrometeors such as rain drops will have positive ZDRs, while vertically oriented hydrometeors such as vertically oriented ice crystals will have negative ZDRs.

8. Typical ZDR Values for Various Targets

9. New Product #2 Correlation Coefficient (CC)
Definition
Possible Range of Values
Units
Abbreviated Name
Measure of similarly of the horizontally and vertically polarized pulse behavior within a pulse volume
0 to 1
None
CC (AWIPS)
ρHV (Literature)
The correlation coefficient is a measure of how similarly the horizontally and vertically polarized waves are behaving within a pulse volume. Its values can range from 0 to 1 and have no units. In AWIPS, this variable is referred to as CC, but in the literature it will be referred to as rhoHV.

10. CC Physical Interpretation
Non-Meteorological
(birds, insects, etc.)
Metr (Uniform)
(rain, snow, etc.)
Metr (Non-Uniform)
(hail, melting snow, etc.)
Shapes are complex and highly variable. Horizontal and vertical pulses will behave very differently with these objects
Shapes are fairly simple and do not vary much. Horizontal and vertical pulses behave very similarly with these objects
Shapes can be complex and are mixed phase. Horizontal and vertical pulses behave somewhat differently with these objects
Low CC (< 0.85)
High CC (> 0.97)
Moderate CC (0.85 to 0.95)

11. Correlation Coefficient (CC) Typical Values
Non-Precip
Precip

12. New Product #3: Differential Phase Specific Differential Phase Shift (KDP)
Definition: gradient of the difference between phase shift in the horizontal and vertical directions
Units: degrees per kilometer (o/km)
Differential phase shift

13. What ΦDP Means t6
ΦDP

14. Gradients Most Important = KDP!!!
KDP Has Big Advantages
Immune to partial beam blockage, attenuation, radar calibration, presence of hail
Used primarily for rainfall estimation and locating heavy rain
Gradients Most Important = KDP!!!

15. Typical Values :: KDP
Generally KDP is highly correlated with the amount of liquid water and almost liberally to rainfall rate.
The main difference between ZH and KDP is that ZH gets contributions from all hydrometeors, including those comprised of ice, whereas KDP is not sensitive to ice particles.
Almost linearly related to rainfall rate
KDP is great at indicating high amounts of liquid precipitation

16. Melting Layer Detection Algorithm
Run in the RPG
Mixed phase hydrometeors: Easy detection for dual-pol!
Z typically increases (bright band)
ZDR and KDP definitely increase
Coexistence of ice and water will reduce the correlation coefficient (CC ~ )
Algorithm overlay product for top and bottom of melting layer
User Selectable MLDA, RUC, Sounding

17. ML Product in AWIPS

18. Hydrometeor Classification Algorithm (HCA)
Run in RPG
Algorithm makes best guess of dominant radar echo type for each gate
Display Product for each radar elevation angle
Based on Fuzzy Logic
Tornado debris category to be added
Lgt/mod
rain
Heavy
Hail
“Big
drops”
Graupel
Ice
crystals
Dry
snow
Wet
Unknown
AP or
Clutter
Biological
Current Classification Options

19. 20000 ft MSL
SOO-DOH Images\kcri_0.5_HC_ _0638.png

20. Hydrometeor Classification Algorithm Challenges
Run in the RPG
Verification Limitations: We need the A10 aircraft
“Fuzzy” Logic; assumes Zdr Accuracy
Typical Radar sampling limitations (snow at 2000 ft AGL may not be snow at the surface)

21. Dual-Pol QPE Algorithm
Run in the RPG
Uses Z, Zdr, Kdp
9 new products
Match Legacy PPS
Instantaneous Rate
User Selectable (Up to 10 durations) for the NWS
Difference products
Legacy Products still available

23. The WSR-88D Dual-Pol Upgrade

24. WSR-88D Dual-Pol Calibration
Dual-Pol calibration is more complex than Legacy calibration
Zdr calibration
Initial System PHIdp calibration
Components/outputs are temperature sensitive
Initial system PHIdp calibration working well
Zdr calibration still under investigation
Full system calibration
Vertical pointing = NO
Cross-polar calibration = Not Yet
We think Zdr not calibrated to < 0.1 dB

25. “Investigate System ZDR” Basics
Part of “ZDR” comes from the system
Different losses in H & V transmit and receive paths
ZDRtrue = ZDRmeasured - ZDRsys
ZDRsys initially measured during off-line calibration, then adjusted for drift over time
ZDRsys = Initial ZDRsys + drift compensation
Updated each
volume scan
Offline calibration
Retrace + 8 hr check

26. 0.99<Rhv ; Range: 20-60 km; Elevation: 2.4 deg; Height < 2.5 km
0.65 dB observed value
Zdr sys error = .42 dB
0.23 dB expected value

27. 0.99<Rhv ; Range: 20-60 km; Elevation: 2.4 deg; Height < 2.5 km
0.45 dB observed value
Zdr sys error = .22 dB
0.23 dB expected value

28. The WSR-88D Upgrade Deployment
All WSR-88Ds upgraded = NO
10-14 days radar downtime during upgrade
System Test: Apr 10 – Sep 10
KOUN: April 2010
Ops Test: Sep 10 – May 11
Vance: Feb 2011
Beta Test: Jun 11 – Aug 11
Wichita: June 2011
Phoenix: June 2011
Pittsburgh: July 2011
Morehead City: July 2011
Full Deployment: Sep 2011 to Apr 2013

29. Sep 26, 2011 Legend Deployment Complete Deployment In Progress
Deployment Scheduled
Legend
Radar coverage shown is at 10,000 ft AGL or below

30. Jan 2, 2012 Legend Deployment Complete Deployment In Progress
Deployment Scheduled
Legend
Radar coverage shown is at 10,000 ft AGL or below

31. Jun 18, 2012 Legend Deployment Complete Deployment In Progress
Deployment Scheduled
Legend
Radar coverage shown is at 10,000 ft AGL or below

32. Dec 01, 2012 Legend Deployment Complete Deployment In Progress
Deployment Scheduled
Legend
Radar coverage shown is at 10,000 ft AGL or below

33. Apr 22, 2013 Legend Deployment Complete Deployment In Progress
Deployment Scheduled
Legend
Radar coverage shown is at 10,000 ft AGL or below

34. Life Gets More Complicated With 5-cm and 3-cm Radar34

35. The Result for 5-CM & 3-CM Scattering response different for
shorter wavelengths (above)
ZDR and KDP are different at the shorter wavelengths (right)

36. Summary WSR-88D Dual-Pol deployment underway; lots of data in 2012
More work needed on dp calibration and dp algorithms
Some model verification work with dp data can be done soon
Lots of model verification work with dp data can be done with time
How will dual-pol information be assimilated into models?
Assimilate dual-pol variables?
Assimilate dp algorithm output?
Drop size and ice distributions?
Something else?

37. Questions?

38. Impacts of radar wavelength
Fields of measured Z, ZDR, and ΦDP at C and S bands for the storm on 03/10/2009 at 0309 UTC.
El(C) = 0.41°, El(S) = 0.48°. C-band radar is at X = 0, Y = 0.
The areas of visible negative bias of Z caused by attenuation at C band are marked as A and B (left top panel).
Taken from: Gu, et al. (conditionally accepted to JAMC)

39. Those Darn Laws of Physics Again
The Full Radar Equation is Ugly
Radar Scattering Cross-Section Equation is Ugly
We Simplify Things at 10-cm Wavelength
10-cm: Rayleigh Approximation
Scattering by particles whose radii are ~1/10 of the radar wavelength or smaller
5-cm & 3-cm we must use the full Mie
Scattering Equation…the Ugly Equation

40. Important Information
The fundamentals of this presentation and other Dual-Polarization training materials for outreach (NWS, media, others) are at: 40