Initial Implementation of Super-Resolution Data on the NEXRAD Network

1
Initial Implementation of Super-Resolution Data
on the NEXRAD Network

• Dr. Sebastian TorresCIMMS/NSSL
• A presentation to the Data Quality TeamJune 15,
  2007

2
What is Super Resolution?

• Legacy (existing) Resolution spatial sampling
• Reflectivity 1-km by 1-deg grid
• Doppler 250-m by 1-deg grid
• Super Resolution spatial sampling
• All moments 250-m by 0.5-deg grid
• Finer spatial sampling and smaller resolution
  volume lead to about 50 improvement in range of
  detection for mesocyclone and tornado signatures
  (Brown et al. 2002)

Tornado outbreak in Oklahoma City9 May 2003
(Curtis et al. 2003)
3
Super Resolution for NEXRAD
Signal Proc.
Algorithms
Z
ORDA
ORPG
ORDA
v
Products
w

• Super-resolution data scheduled for operational
  use on NEXRAD
• Short-term goals - Phase I (ORDA Build 10)
• Data used for visualization only
• Produce legacy- and super-resolution data streams
• ORPG algorithms ingest legacy-resolution data
• Super-resolution data produced on lower-elevation
  scans
• Higher likelihood of finding tornado signatures
• Long-term goals - Phase II (ORDA Build gt12?)
• Data used by the algorithms
• Produce super-resolution data stream only?
• Not every algorithm may benefit from
  super-resolution data

4
Characteristics of Super-Resolution Data

• Smearing due to antenna rotation is reduced using
  a data window
• Effective beamwidth is 25 narrower than with
  legacy resolution (Torres and Curtis 2006)
• Better spatial resolution due to smaller
  resolution volume
• Reduced data quality due to data windowing
• Errors of estimates are 30 larger and NEXRAD
  Technical Requirements are not met

5
Super Resolution Goals and Constraints

• Produce stronger signatures of mesocyclones and
  tornadoes
• Assure compatibility with current and future
  (planned) signal processing techniques
• Assure compatibility with current Volume Coverage
  Patterns (ra, va, update times)
• Adhere to CPU load and bandwidth limitations
• Meet NEXRAD requirements for errors of estimates
• Provide acceptable base data to algorithms

6
Super Resolution in the ORDA

• ORDA must produce base data with finer spatial
  sampling and resolution
• Process overlapping 1-deg radials with data
  windowing sampled every 0.5 deg and no range
  averaging
• For each range gate, M time-series data samples
  are weighted
• with von Hann window if clutter filtering is
  not needed
• with Blackman window if clutter filtering is
  needed

7
Super Resolution in the ORDA

• ORDA must produce unfolded reflectivity in the
  Doppler half of split cuts and must include the
  noise power in the metadata for the ORPG
• This additional data are required in the ORPG to
  produce legacy-like data from super-resolution
  data
• ORDA must double the throughput
• Processing pipeline is minimally affected, but
  there are twice the number of radials in a
  super-resolution data volume scan
• ORDA-ORPG transmission bandwidth must accommodate
  faster radial rates and additional data

8
Super Resolution in the ORPG

• ORPG algorithms expect data with legacy
  resolution and quality
• Super-resolution data does not have the required
  resolution or quality for the algorithms
• Must have both legacy and super-resolution data
  streams
• Two instances of the velocity dealiasing
  algorithm
• Radial recombination
• Two super-resolution radials are recombined to
  form one legacy-resolution radial

Zr
1 km
vr
250 m
1 deg
Super Resolution
Legacy Resolution
9
Super Resolution in the ORPG

• The radial recombination algorithm assumes a
  bimodal Gaussian spectrum model
• Recombined data has legacy resolution and
  acceptable data quality
• Recombined super-resolution data will not be
  exactly the same as the true legacy-resolution
  data

10
Super Resolution in the ORPG

• Radial recombination algorithm must deal with
  missing data and several special boundary
  conditions
• Base data at the ORPG is censored and quantized!
• Censoring can occur due to SNR thresholding or
  overlaid echoes

ZBG is derived from a best-guess power (PBG) 0
PBG lt N SNRth
Z11
Z21
Zr
?
Z12
Z22
1 km
Z13
Z23
Z14
Z24
1 deg
11
Super-Resolution Reflectivity
KCRI March 19, 2006
12
Recombined Legacy-Resolution Reflectivity
KCRI March 19, 2006
13
True Legacy-Resolution Reflectivity
KCRI March 19, 2006
14
Summary

• Super Resolution data produces enhanced
  mesocyclone and tornado signatures
• Potential to detect weaker and/or more distant
  tornadoes
• To meet long-term technical goals, NWS will
  implement and use Super Resolution with range
  oversampling techniques
• Staged approach
• Build 10
• Recommended super resolution with azimuthal
  radial recombination
• Beyond Build 12
• Range oversampling techniques
• Algorithms modified to ingest super-resolution
  data