Application of GPS Radio Occultation Data to Weather Analysis and Prediction

1
Application of GPS Radio Occultation Data to
Weather Analysis and Prediction

• Bill Kuo
• NCAR/MMM Division
• and
• UCAR/COSMIC Office

2
GPS Radio Occultation

• The 24 GPS satellites are distributed roughly in
  six circular orbital planes at 55o inclination,
  20,200 km altitude and 12 hour periods.
• Each GPS satellite continuously transmits signals
  at two L-band frequencies, L1 at 1.57542 GHz (19
  cm) and L2 at 1.227 GHz (24.4 cm).
• A ray passing through the atmosphere is refracted
  due to the vertical gradient of density.

GPS satellite
Low Earth Orbiting (LEO) satellite
3
GPS Occultation
Basic measurement principle Deduce atmospheric
properties based on precise measurement of phase
delay and amplitude.
4
GPS/MET Experiment
The GPS/MET Team

• UCAR, JPL, and U. of Arizona, collaborated on the
  GPS/MET program in 1993, with a goal to
  demonstrate the radio occultation sounding
  technique for Earths atmosphere.
• The satellite was launched on April 13, 1995, and
  data were collected for two years.

5
Two Recent GPS Occultation Missions

• SAC-C
• Argentine spacecraft carrying a multispectral
  imager and magnetometer
• Launched Nov. 21, 2000
• 98.3o inclination
• 700 km orbit
• Both rising and setting occultation

• CHAMP
• a German mission for magnetometry and gravity
  mapping
• Launched July 15, 2000
• 87o inclination
• 435 km orbit
• 5 year mission life

6
Examples of GPS RO soundings
7
A GPS/MET sounding through an intense upper-level
front
0055 UTC October 22, 1995 over China
8
Location of a CHAMP sounding
9
CHAMP compared to two radiosondes in Typhoon
18UTC 7/29/01 Ishigakijima 00UTC 7/30/01
CHAMP (red/blue lines) 2008 UTC 29 July 2001
10
CHAMP Comparison with Hualien radiosonde
The Hualien station is affected by cloud
and precipitation
11
Detection of gravity waves with GPS
Dry GPS/MET retrieval
From T. Tsuda
12
Characteristics of GPS RO Data

• Limb sounding geometry complementary to ground
  and space nadir viewing instruments
• High accuracy (equivalent to lt 1 deg K from 5-25
  km)
• High vertical resolution (0.1 km surface - 1km
  tropopause)
• All weather-minimally affected by aerosols,
  clouds or precipitation
• Independent height and pressure
• Requires no first guess sounding
• Independent of radiosonde calibration
• No instrument drift
• No satellite-to-satellite bias

13
How accurate is GPS/MET sounding?

• CHAMP and SAC-C Missions provide opportunity for
  Cross-comparison

14
Cross-comparison between CHAMP and SAC-C
Paris of SAC-C and CHAMP soundings within 30 min
of each other
From Hajj et al. (2003)
15
Comparison with Radiosondes
16
Radiosonde Stations and Manufacturers
Vaisala/Australia IM-MK3/India
Shang/China MEISEI/Japan Mars/MRZ VIZ
AIR
Others
From Junhong Wang
17
Quantifying Regional Differences

• Calculate the mean absolute difference in
  refractivity between CHAMP and Radiosondes (NCR)
• Calculate the mean of the absolute value of the
  difference in refractivity between CHAMP and the
  ECMWF (NCE)
• Compute the ratio (D NCR /NCE)
• Perform calculation using radiosonde data from
  different regions of the world from June 2001 to
  present.

From D.Rossiter (UCAR Summer Student)
18
Regional Results
19
India Statistical Plots
Regional Measurements June
2001 to present
ECMWF - CHAMP
Mean Standard
Deviation Count
CHAMP-Sonde
ECMWF-Sonde
20
Comparison of radiosonde performance
IM-MK3 (India)
Vaisala (Australia)
21
Preliminary Conclusions

• The RO technique is accurate enough to
  differentiate differences in performance of
  various types of radiosonde systems.
• IM-MK3 (used by India) appears to have the
  poorest performance.
• Vaisala (used by Australia and others) gives the
  best performance.
• These results are consistent with independent
  assessment performed by Wang et al. (2003).

22
Using GPS soundings to evaluate global analyses
23
ECMWF NCEP/NCAR reanalysis
100 mb Temperature
24
GPS NCEP/NCAR Reanalysis
100 mb Temperature
25
GPS ECMWF analysis
100 mb Temperature
26
Comparison between SAC-C and global analyses
Dec 2001
40 S 90S
27
Assimilation of GPS RO data
28
Assimilation of CHAMP data

• Zou et al. (2003) performed an assimilation of
  CHAMP GPS RO data during the two-week period of
  16-31 July 2002.
• Assimilation of bending angles using NCEP 3DVAR
  system with 6-h update cycle
• Data provided by UCAR COSMIC Data Analysis and
  Archival Center (CDAAC)
• Out of 3030 GPS RO soundings, 1427 penetrated
  below 850 mb
• NO-GPS experiment Include all data currently
  used by NCEP
• GPS experiment Include all data currently used
  by NCEP CHAMP data

29
Assimilation of CHAMP data
500 mb height
500 mb height
850 mb q
850 mb q
Southern Hemisphere
From Zou et al. (2003)
30
Assimilation of CHAMP data
500 mb height
500 mb height
850 mb q
850 mb q
Northern Hemisphere - tropics
From Zou et al. (2003)
31
Assimilation of CHAMP data
500 mb height
Southern Hemisphere
From Zou et al. (2003)
32
A forecast impact trial with CHAMP
radio occultation measurementsSean Healy,
Adrian Jupp and Christian Marquardt

33
Trial Run from UK Met Office
We are assimilating CHAMP refractivity profiles
provided by GFZ, between May 26, 2001 - June 11,
2001. Each profile contains 120 refractivity
values (150 max), with a vertical separation of
200m. We do not assimilate refractivity below
4km because of the well known biases.
Observation errors are based on Kursinskis
1997 estimates, but we have inflated them to 2
at the surface, falling linearly to 0.2 at 10km.
QC based on a 1D-Var calculation. Note, we
only obtain 40 measurements per assimilation
cycle.
34
NWP forecast fit to radiosondes at 50hPa(Trop)
35
NWP forecast fit to radiosonde 250hPa (SH)
36
Summary of UK Met Office Study (Healy et al. 2003)
We have performed the 1st impact study with RO
data at the Met Office. Given the small number
of observations, results are very encouraging.
PMSL in tropics improved, but largest impact
seen in the southern hemisphere. Upper-troposphe
re and lower stratosphere show improvements in
the 250hPa Temp (SH) and 50hPa(all) . The
results would support the case for assimilating
RO measurements operationally. Prospects of
obtaining measurements from a constellation
(e.g., COSMIC, ACE) are very exciting in the
light of these results.
37
Case / model domain / GPS occs
38
Observations

• GPS refractivity
• Conventional observations
• QuickScat surface wind vector
• SSM/I Precipitable water
• SSM/I Rainfall rate
• SSM/I vertically integrated cloud liquid water
• SSM/I surface wind speed

39
Differences between GPS and No-GPS Experiments
in MSLP
40
Differences between GPS and No-GPS Experiments
in MSLP

41
Differences between GPS and No-GPS Experiments
in MSLP
42
MM5 4DVAR assimilation of 47 GPS RO over a
one-day period
Prediction of an Antarctic cyclone
43
(No Transcript)
44
RO observational operator (1)
How well does the Abel-retrieved N(z) represent
the local atmospheric refractivity at the
estimated ray tangent point in the presence of
strong horizontal gradients?
From Sergey Sokolovskiy
45
RO observational operator (2)
1) Modeling of RO observable in straight-line
approximation (Syndergaard et al., 2002).
Since the same approximation is used in forward
and inverse modeling, the errors are
partially cancelled out.
Modeling of excess phase along straight line
Inversion of the straight-line excess phase
2) Curved trajectories (rays) corresponding to
Abel-retrieved refractivity can be used,
instead of the straight lines, to improve
modeling of RO observable (Ahmad, 1998)
46
RO observational operator (3)
Once the trajectories (straight or curved), used
for modeling of RO observable, do not depend on
the model refractivity, the modeled observable is
linearly related to that refractivity
1) Forward linear operator (straight-line excess
phase path)
where D is simply the length of j ray inside k
voxel or contribution from k pixel (interpolation
required). Index k (a 3-D set of indices) counts
all voxels (pixels) contributing to excess phase
along rays.
2) Inverse linear operator (Abel-inverted
refractivity)
Observational linear operator
47
COSMIC (Constellation Observing System for
Meteorology, Ionosphere and Climate)

• 6 Satellites launched in 2005
• Three instruments
• GPS receiver, TIP, Tri-band beacon
• Weather Space Weather data
• Global observations of
• Pressure, Temperature, Humidity
• Refractivity
• Ionospheric Electron Density
• Ionospheric Scintillation
• Demonstrate quasi-operational GPS limb sounding
  with global
• coverage in near-real time
• Climate Monitoring

48
(No Transcript)
49
COSMIC Spacecraft Status
Minotaur launch vehicle
COSMIC spacecraft
GPS receiver
50
COSMIC Soundings in a Day
Red dots are current radiosonde sites, green dots
are 2500 COSMIC soundings in 24-h.
51
GPS radio occultation missions
52
COSMIC EQUARS Soundings in a Day
COSMIC EQUARS Radiosondes
53
Research Opportunities

• Develop cost-effective assimilation techniques
  for GPS RO data assimilation
• Comparison of 3DVAR/4DVAR/EnKF assimilation
  approaches
• Comparison of different observation operator (ray
  tracing, refractivity, etc)
• Optimal coupling between Nadir passive satellite
  measurements with GPS RO data
• Demonstrate the value of GPS RO in global and
  regional scale weather analysis and prediction.
• Use of GPS RO data to study atmospheric
  circulations where traditionally there are few or
  no soundings (e.g., oceans, polar regions,
  Tibetan plateau, etc)

54

• Use of GPS RO data (from on-going and future
  missions) to evaluate the performance of global
  climate models (e.g., CCSM)
• Geopotential height changes, refractivity
  climatology
• Use of GPS RO data to improve the understanding
  of global hydrological cycle (e.g., combining GPS
  RO with other satellite data).
• Study the exchange between upper troposphere and
  lower stratosphere (GPS RO is of highest accuracy
  in this region)
• Use of GPS RO data in field programs (location of
  an occultation can be predicted two weeks in
  advance)
• Many other possible research topics

55
Operational Demonstration

• A major objective for COSMIC is to demonstrate
  the usefulness of GPS RO soundings in operational
  numerical weather prediction.
• Many operational NWP centers have expressed
  strongly interest in using the COSMIC data
• NCEP
• ECMWF
• UK Met Office
• Met. Services of Canada
• Chinese Met. Administration
• Danish Met. Institute
• German Weather Service
• Japan Met. Agency
• Major impact in Op. NWP
• is expected!

Canadian Met. Center 7/14/03
56
Summary

• GPS radio occultation technique is a very
  promising atmospheric observing system for both
  weather and climate
• High vertical resolution
• High density (depending on the number of
  missions)
• High accuracy
• No instrument drift (important for climate
  detection)
• GPS RO technique is here to stay (COSMIC is
  coming, and more missions are being planned)
• GPS RO data will be valuable to
• climate change analysis, validating of climate
  simulations
• operational weather prediction
• ionospheric research and space weather forecasting