Assimilation and Impact of GPS Radio Occultation data

1
CDAAC Update
2
Issues
3
Open Loop Status at JPL

• UCAR and JPL have agreed on a minumiuCurrently
  tracking in OL mode each day UT 10 - 16
• Setting OL occultations working well
• Rising OL occultations presently 15 per 6 hours
  that look good
• Bug fix of scheduler should bring this to 50 /
  six hours
• Transition to L1 phased-lock-loop (PLL) at -10 km
  line-of-sight altitude (LSA)
• Transition to L2 PLL 3-4 seconds later (15-20 km
  tangent point height)
• UCAR is expected to receive data by next week
• JPL still needs to write the data translator
• UCAR and JPL have agreed on a minimum
  requirement document for first flight software
  version (expected in a few weeks)
• Things look positive - need to see the data

4
CDAAC Open Loop Data Processing Status

• CDAAC processed some initial OL data from JPL
  (Aug. 04)
• Format of OL data and content will change again
  in next release
• Significant changes to the CDAAC software are
  expected including
• Level 0 - Level 1 translator
• Additional external data are needed (50 bps data
  stream)
• Format of excess phase file may have to be
  changed to include OL model output
• Interface to inversion software will have to be
  changed
• Changes will be significant but it is not clear
  yet what is the best way to accomplish them
• Work on this will hopefully start March 14 05

5
Receiver generated doppler (model) agrees
reasonably well with UCARs model. several Hz
6
Reasonable SNR can be observed for open loop data
(right of red line)
7
(No Transcript)
8
Refractivity profile from one of the early OL
profiles from SAC-C
9
OL starts at 3000, Data bit modulation cannot
be removed after 4000 samples
10
Bit Grab Receivers

• COSMIC has to develop GPS receivers to collect
  and send to CDAAC all GPS data bits (50 bps data
  stream)
• Contract with CU has been negotiated - FFP for
  25 k to design and build 10 such receivers
• UCAR/COSMIC is purchasing all the hardware
  computers, special GPS receiver boards, etc
• Receivers must be installed at global sites
  (Boulder, Kyoto, Potsdam, Sao Paulo, Wellington,
  Johannesburg, .) host sites must still be found
  and negotiated
• Prototype is currently collecting data bits at CU
  
• Establishing this network is challenging - but on
  track
• Without this network OL data profiles degrade

11
Fiducial Network for COSMIC

• Three COSMIC fiducial sites installed in TWP
• Two COSMIC fiducial sites delivered to Brazil
  (just passed customs)
• Eleven fiducial sites are regularily received in
  real-time from NRCan (IGS)
• Six more sites will become available by March 11
  (this week)
• 30 sites are received with 15 minutes latency
  from CDDIS
• Presently we receive sufficient amount of
  fiducial data for the COSMIC mission goals (still
  we plan to add 20 more real-time sites for
  greater redundancy)

12
Revising CDAAC QC

• Existing CDAAC QC has been reviewed based on
  feed-back from first assimilations at NCEP and
  comparison with ECMWF reanalysis
• Generally the existing QC was found to work well
  by rejecting most bad occultations
• However, some bad occultation still passes the QC
  with errors up to 40
• By closely inspecting those profiles it was found
  that these outliers can be divided into five
  categories
• Too late truncation of L1 when this signal is
  lost
• Too late truncation of L2 when this signal is
  lost
• Too low truncation of the radio holographic
  bending angle profile
• Bad signals (tracking errors and 1 second spike
  (CHAMP))
• Poor adjustment of climatology for statistical
  optimization

13
Improved CDAAC QC

• Threshold for controlling the truncation of L1
  and L2 have been tightened to avoid big errors in
  the troposphere related to L1 and L2 truncation
  time
• A new more conservative QC check has been
  introduced to control the truncation height of
  the radio holographic bending angle profile
• A New QC check has been introduced to catch bad
  profiles based on relative deviations from the
  climatology bending angle and deviations between
  RO scale height and the scale height of the
  climatology
• A more robust technique for adjusting the
  climatology for statistical optimization has been
  implemented
• New QC criteria greatly reduces the number and
  magnitude of outliers

14
Result Improvements at UCAR
Old
New
15
BUFR and data distribution UCAR
16
Space Weather data analysis at CDAAC

• Defining near-real-time ionospheric products and
  formats - draft to be sent out to the space
  weather community with a request for feedback
• Absolute TEC to all GPS satellites in view at all
  times - automatic editing algorithm to delete
  outliers and correct for cycle slips -
  solving for differential code biases (one day
  averages) of COSMIC GPS receivers (POD antennae)
  - near-real-time calibration of TEC
• TIP processing - conversion from raw TIP data
  to radiances (getting radiance code from Scott
  Budzien, NRL) - calculate pointing direction from
  orbit and attitude of spacecraft
• Electron density profiles - derive orbit electron
  density from TEC observations to be used for
  initialization (upper boundary) of profile
  retrievals - use horizontal gradients from
  Global Ionospheric Maps (GIMs) for improved
  ionospheric profiling (getting GIMs either from
  JPL or from CODE)

17
Near Real-Time CHAMP Orbit Overlap Results (vs.
JPL) with Bernese v5.0
Arcs for every CHAMP data dump
Position
Velocity
18
COSMIC GPS Antenna Test at Ball Aerospace
19
Antenna Requirements
20
COSMIC GPS Limb Antenna GainSAD0deg
L2
L1
21
COSMIC GPS Limb AntennaL3 Phase (FOV7 X
40)SAD(0)
L3 Phase (mm)
Gradient Magnitude(mm/deg)
22
COSMIC GPS POD AntennaL3 Phase (FOV60)SAD(0)
L3 Phase (mm)
Gradient Magnitude(mm/deg)
23
POD Excess Phase Status

• POD - To Do
• - Analyze orbit quality for all missions,
  optimize reduced-dynamic tracking
• - Investigate simultaneous use of data from 2
  POD antennas
• - Additional real-time
• Excess Phase
• - Test single-difference and zero-difference
  (GRACE-USO) processing
• - correct cycle slips in non-occulting links
• - modify software to process data with short
  gaps in non-occulting links
• - include antenna phase center corrections in
  processing
• - Complete documentation and algorithm
  description

24
CDAAC Status Summary

• A lot of work and exciting challenges remain
• There appear to be no show-stoppers or
  insurmountable problems - just a lot of tough
  problems
• CDAAC team is working hard and making progress
• CDAAC will be completed for launch assuming that
• No unexpected problems in OL data
• Postpone or receive additional help to deal with
  RoadRunner
• TBB network will be installed by NRL and NRL
  provides TIP software modules
• No staffing changes (hired J. Johnson for Bit
  Grabber project, fiducial tasks and software
  support)
• Minimize distractions of CDAAC staff between now
  launch