CDAAC status

1
CDAAC status
2
CDAAC 1.0

• Delivered on time to NSPO
• Two CDs including over 120 source code packages
  and 800 Megabytes of test data for 3 satellite
  missions

3
CDAAC 1.0 features

• Completely re-written for Bernese 5.0
• Improved LEO orbit computation
• Post-processing now can be spread to all nodes of
  a cluster
• Improved statistical optimization and error
  characterization in atmospheric inversions
• Ionospheric processing added to CHAMP and SAC-C
  missions

4
CDAAC 1.0 features (cont)

• Automatic Ionosonde comparisons
• New 1D variational assimilation for moisture
  computation
• Web-based operator interface and configuration
  editor
• Extensive documentation improvement
• Bug fixes and code cleanup

5
Missing from CDAAC 1.0
6
POD and Excess Phase StatusBill SchreinerDoug
Hunt
7
(No Transcript)
8
POD and Excess Phase (cont)
9
Post-Processed CHAMP Orbit Overlap vs. JPL
Daily 24 hour arcs
Position
Velocity
10
Near Real Time CHAMP Orbit Overlap
Arcs for every CHAMP data dump
Position
Velocity
11
Radio Occultation Sensor Evaluation (ROSE) -
GFZ/JPL/UCAR Inter-comparison(extracted from
Jens Wickert presentation at NCAR Summer
Colloquium - 2004)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
1D VAR moisture assimilationTae-Kwon Wee
17
1D-Var Retrieval Scheme

• 1D-Var method now used replacing previous
  statistical retrieval code (simret)
• Initial tests show high stability of algorithm
  and realistic statistics
• Further work is necessary to settle on optimal
  configuration parameters and covariance matrices
• Documentation available online

18
Ionospheric Data ProductsStig SyndergaardBill
Schreiner
19
Ionospheric data products

• CDAAC plans to provide the following ionospheric
  data products
• GPS receiver
• Absolute Total Electron Content (TEC) to all GPS
  satellites in view at all times
• Occultation TEC and derived electron density
  profiles
• Tiny Ionospheric Photometer (TIP)
• Nadir intensity on the night-side from radiative
  recombination emission at 1356 Å
• Derived F-layer peak density and critical
  frequency
• Location and intensity of ionospheric anomalous
  structures
• Tri-Band beacon
• Phase and amplitude of radio signals at 150, 400,
  and 1067 MHz
• TEC between the COSMIC satellites and ground
  receivers

20
Ionospheric occultation data processing CDAAC
1.0

• CHAMP electron density profiles
• Compared with in situ electron density from CHAMP
  Langmuir Probe ()

• Currently looking into bias to Langmuir Probe
  data
• Work to be done on combining RO data with other
  data types

21
Combination with TIP and Global Ionospheric Maps
(GIMs)

• TIP provides valuable information about
  horizontal gradients along the subsatellite track

• GIMs (derived from ground-based GPS) provide
  global information about horizontal gradients

22
Open Loop data statusSergey SokolovskiyBill
SchreinerDoug Hunt
23
Preparing for open-loop (OL) tracking of
L1 tropospheric RO signals in COSMIC
Advantage free of tracking errors. Advantage
acquisition of rising occultations. Disadvantage
generally, lower SNR. Due to limited bandwidth
in LEO, OL tracking needs Doppler model. RO
signal is acquired and transmitted to the ground
with un-removed data modulation (must be removed
in post-processing). JPL started acquisition of
the OL setting occultations on SAC-C. The data
are forwarded to CDAAC for analysis and
feedback. CDAAC is exploring different options
for removal of data modulation.
24
Validation of the receiver-running Doppler
model against more accurate computer-running
model based on N-climatology, in general, shows
good agreement. Some things (2-3Hz offset
at PLL-OL transition and gt5Hz deviation at
HSLlt-100km) are suggested for the future
improvement.
25
Strong atmospheric modulation and noise do not
allow to clearly distinguish the data bits in
acquired signal (this, in part, must be the
reason for PLL tracking errors).
the samples likely Affected by data bits
?
the samples likely not affected by data bits
26
An example of retrieval of the SAC-C OL data
(lat23o) with ad hoc removal of data modulation
based on magnitude of the phase lapse
27
Currently, statistics of the SAC-C OL retrievals
are not as good as PLL. This can be related to
incomplete removal of data modulation in the
post-processing (currently using ad hoc method
based solely on the phase lapse). CDAAC is
looking into possibility of independently
obtaining data bit stream from all GPS satellites
by ground receivers for complete data
demodulation in the post-processing (Open Source
GPS Software). Concern rising occultations on
SAC-C are not acquired yet.
28
CDAAC operational concernsDoug HuntKarl
HudnutMaggie Sleziak-Sallee
29
CDAAC preparation for launch

• For launch in late 2005, we will need to sort
  out
• Complete reprocessing with CDAAC 1.0 code
• Dual cluster software issues
• Purchasing of production clusters and I/O
  machines
• I/O machine replacement issues
• Show and tell display
• Fiducial network development
• Porting CDAAC to 64 bit machines
• Final archival policies
• Processing other missions, such as GRACE
• Other items?

30
Dual string processing
Scp forwarding
TACC
Data sources USN, CDDIS, JPL, etc
FTP
Web or FTP
Scp forwarding
Scp forwarding
NFS mount, Postgres client
Previous day rsync
UCAR security perimeter
31
Computer Room Requirements

• Need 2 new Beowulf Clusters for dual string CDAAC
• Beowulf Clusters and racks require
• 2.5 x 4 x 6.2 (width x length x height)
• Beowulf Clusters and racks consume approx. 3 kW
  and dissipate heat at approx. 9000 BTU/Hr
• Total power con-sumption 21.5 kW.
• Total heat dissipation 64,500 BTU/Hr.

32

• A large monitor will be hung in the entryway of
  the FL4 COSMIC offices
• This will show a continuous display of COSMIC
  data recently processed
• An application which displays COSMIC data in a
  cool way needs to be developed
• Initial ideas
• A very high resolution flat-panel monitor
  attached to a fast PC running a web browser
• A special URL on the I/O machines web server
  which would display several animated Earth
  images, one for each COSMIC satellite
• (see www.cosmic.ucar.edu/champ.html)
• Other ideas?

33
Fiducial Network

• Current plans for COSMIC real time fiducial data
• Get 15 minute, 1 second igsHrf data from CDDIS
• Supplement this with streaming data from NR CAN
  (canHrf)
• Finally, add several stations set up by COSMIC in
  the South Pacific, Brazil and perhaps other
  places (cosHrf)
• Software in place to convert these data types to
  level1a (fidGps and f30Gps)

34
Lots to do