GPS

1
Validation of GPS Radio Occultation Results in
Japan and India T. Tsuda(1), D.N. Rao(2), M.V.
Ratnam(1,2), J. Furumoto(1), T. Nakmaura(1), H.
Hayashi(1), Y. Aoyama(3), S. Satoh(4), and Y.
Murayama(4) (1) Research Institute for
Sustainable Humanosphere (RISH), Kyoto
University (2) National Atmospheric Research
Laboratory (NARL), Indian Space Research
Organization (ISRO), (3) National Institute for
Polar Research (NIPR) (4) National Institute of
Information and Communications Technology (NICT)
FORMOSAT3/COSMIC Workshop October 16-18, 2006
2
Global Earth Observation Systems System
(GEOSS) Theme (2) Water cycle and climate
changes in the Asia-monsoon region
22-3 Application of GPS radio occultation (RO)
data for studying temperature and humidity
variations in the tropical troposphere
(1) Development of retrieval algorithm for GPS
RO data Y. Murayama (NICT), Y. Aoyama
(NIPR) (2) Assimilation of GPS RO data into a
meso-scale numerical weather prediction model
Y. Shoji, H. Seko, T. Kawabata, K. Aonashi, M.
Kunii (MRI Meteorological Research
Institute) (3) Validation and scientific
application of GPS RO data T. Tsuda, M.
SHiotani, J. Furumoto, H. Hayashi, T. Nakamura,
T. Horinouchi, N. Nishi (RISH, Kyoto University)
3

• Objective of the study
• We obtain accurate profiles of temperature and
  humidity observed with the GPS RO technique for
  extensive oceanic and land areas, whose accuracy
  and vertical resolution are comparable to those
  with radiosonde measurements.
• We validate the GPS RO profiles through a
  comparison with ground-based observation
  radiosondes (routine and campaign), Raman lidar,
  wind profiler, other satellite observation
  e.g., AQUA-AIRS and objective analysis.
• GPS RO data is assimilated into a meso-scale
  numerical weather prediction model to improve the
  forecast accuracy.
• Using high-quality GPS RO data, we study the
  temporal and spatial characteristics of
  temperature and humidity, particularly in the
  tropical troposphere in the Asian monsoon region,
  where routine ground-based observations are
  relatively sparse.

This paper A preliminary report on validation of
the FORMOSAT-3/COSMIC GPS RO profiles with
simultaneous radiosonde and Raman lidar results.
We are particularly interested in the height
resolution of the humidity profiles.
4
High resolution radiosonde data by routine and
campaign soundings
Kyoto-U, Shigaraki MU obs.
Ground-based validation sites
India, Gadanki MST radar obs.
Japan Meteorol. Agency (JMA) 18 stations, 2/day
We obtain original records of routine radiosonde
soundings at meteorological agencies (twice daily
at 0 and 12 GMT) with sampling interval of 2-5
seconds (10-30 meter height resolution) Japan
(JMA) 18 sites Malaysia 7 sites
Singapore 1 site Vietnam 3 sites
NICT, Okinawa obs.
Kyoto-U/LAPAN Equatorial Atmosphere Radar (EAR),
Indonesia
Malaysia (7 sites), Vietnam (3 sites),
Singapore Met. Offices
5
Major Ground-Based Validation Sites of the
Project in the Asia Monsoon Region
(1) The MU Radar Obs., Shigaraki, Japan (34.8N,
136.1E) RISH
Ground-based validation sites
(4) MST Radar Observatory Gadanki, India (13.5N,
9.2E) NARL-ISRO
(2) Okinawa Obs., Ogimi, Japan (26.7N, 128.1E9
NICT
(3) Equatorial Atmosphere Radar (EAR), Koto
Tabang, Indonesia (0.2S, 100.3E), RISH and LAPAN
6
(1)
The Middle and Upper (MU) Atmosphere Radar, RISH,
Kyoto Univ (34.8N, 136.1E)
LIDAR (Raman, Rayleigh)
46.5 MHz, 1MW, 103mf 475 Yagi antenna array
7
(2)
Raman lidar (April 2006 - ) (26.7N, 128.1E)
Intensive Radiosonde Campsigns
C-band (5.3GHz) Polarimetirc radar
8
Equatorial Atmosphere Radar (EAR), West Sumatra,
Indonesia (0.2S, 100.3E)
(3)
VHF (47MHz) wind profiler Antenna array (110 m
in diameter) 560 sets of 3-elements Yagi
antennas. Peak transmitting power 100 kW.
?
Radiosonde Lidars X-band radar Micro rain
radar Optical rain gauge Ceilometer Radiometer Dis
drometer GPS receiver
9
(4)
Gadanki MST radar facility (13.5N,
79.2E) National Atmosphere Research Laboratory
(NARL), Indian Space Research Organization
(ISRO)
VHF wind profiler 53 MHz, 1MW, 100mx100 m,
Yagi antenna array Raman and Rayleigh lidars
Intensive radiosode soundings VAISALA RS80,
RS92 daily at 12 GMT from Apr 10 - Sep 25
176 profiles
10
Comparison of GPS RO and radiosonde results at
Gadanki, India
NARL continued daily radiosnde launches at 12 UT
(VAISALA RS80 and RS92) from Apr. 10 to Sep.
25. Among a total of 176 profiles, 27 GPS RO
events occurred within 3o in latitude and
longitude from Gadanki irrespective of the time
difference (19 coincidences occurred after day
number 194).
Radiosonde launch site
A GPS RO event occurred very close to Gadanki on
July 24, 2006. Rasiosonde 13.48N, 79.18E,
0013 UT GPS-RO 1348N, 79.23E, 0345 UT The
horizontal distance at 5 km altitude was about 35
km.
11
Comparison of temperature (DRY) between GPS-RO
and radosonde results at Gadanki on July 24, 2006
12
Temperature (WET)
Comparison of temperature and humidity profiles
between 1D-Var estimates for GPS-RO and radosonde
results at Gadanki on July 24, 2006
?
Water Vapor Pressure
Relative Humidity
13
Comparison between the COSMIC GPS RO measurements
at 0049 UT on September 5, 2006 and a radiosonde
result at Okinawa
Temperature
Refractive Index
Pressure
BlueRadiosonde Black COSMIC (1D-Var) Red
COSMIC (dry)
NICT, Ogimi (26.7N, 128.1E) Launch at 2349UT,
Oct 4
Humidity
Tangent Point
0049 UT, Oct 5, 26.84E, 131.49E
14
Comparison of humidity profiles between GPS RO
1D-Var (0049 UT on September 5, 2006) and
radiosonde launched at 2349UT on Oct 4, from
Okinawa, Japan
15
Vertical Gradient of Refractive Index Mdn/dz M
is mainly determined by dq/dz in a moist
atmosphere
a b c
Relative contribution of the terms (a), (b) and
(c) to M (refractive index gradient)
16
Time-height variation of the water vapor in
comparison with 1st and 2nd height derivative of
refractive Index by using radiosonde data
Radiosonde profiles were observed along the
equator in the eastern Pacific during a research
voyage of a SOWER campaign, Shiotani et al.,
2002, J. Meteor. Soc. Japan, 80, 897- 909
17
Comparison of refractivity and its 1st and 2nd
gradient between GPS RO 1D-Var (0049 UT on
September 5, 2006) and radiosonde launched at
2349UT on Oct 4, from Okinawa, Japan
18
Comparison of Height Derivative of Refractive
Index
19
Cloud distribution looks horizontally
homogeneous. But ..
Cloud distributions (water vapor as well) in the
tropics are characterized by vigorous temporal
and spatial variations, which should be
considered in validating the GPS-RO results.
20
Time - Height Variations of U, T and q with
Atmospheric Radars
Contour Time-height section of convective clouds
observed with the X-band radar Line Temperature
perturbations observed with EAR-RASS
Contour Water vapor variations with
EAR-RASS Arrow wind velocity vector (zonal
vertical) with EAR
21
Time Height Variations of specific humidity
(g/kg) at night (top) and day/night (bottom)
observed with a Raman Lidar in Shigaraki, Japan
January 6-11, 2006 (integration 15 minutes)
???
4km
June 24-28, 2005 (integration 30 minutes)
Height resolution 0- 0.5km 60 m
0.5 - 1 km 120 m
1 2 km 240 m
2 4 km 480 m
22
Water Vapor Profiles Observed with a Raman lidar
at Shigaraki (72m x 1h)
March 20, 2006
March 21, 2006
March 23, 2006
March 24, 2006
23
Observation periods of a Raman lidar at Shigaraki
and Okinawa in 2006
24
GPS RO events of FORMAT-3/COSMIC (with the lowest
height below 10km) around Japan between July 29
and August 7
Aug 2 0803 UT
Aug 1 0813 UT
Shigaraki MU radar Observatory (34.8N, 136.1E)
25
Comparison of a humidity profile between the
Raman lidar (solid line) at the MU radar
observatory and the GPS RO results (dotted line)
of FORMOSAT-3/COSMIC observed at 0803 UT on
August 2, 2006 near the MU radar observatory.
The lidar profiles (72 m height resolution) are
averaged for one hour. In the left panel the
hourly lidar profile at 1116 UT is plotted,
while all available profiles are superimposed in
the right panel.
26
Comparison of a humidity profile between the
Raman lidar (solid line) at the MU radar
observatory and the GPS RO results (dotted line)
of FORMOSAT-3/COSMIC observed at 0813 UT on
August 1, 2006 near the MU radar observatory.
The lidar profiles (72 me height resolution)are
averaged for one hour. In left panel the hourly
lidar profile at 1334 UT is plotted, while all
available profiles are superimposed in the right
panel.
27
Expected Distribution of GPS RO data in the Asia
Monsoon Region
Intensive ground validation sites
Gadanki
MU obs.
Okinawa
EAR
28

• SUMMARY
• We operate four validation sites for GPS RO in
  Japan (Shigaraki, Okinawa), Indonesia and India,
  where wind profilers, lidars and other
  complementary facilities are installed.
• At these site, intensive radiosonde soundings
  (Vaisala, Meisei) were conducted. Some
  coincidence with GPS RO events occurred. We hope
  the frequency will significantly increase if
  accurate prediction of LEO satellite overpass is
  provided well in advance.
• Case studies in India and Okinawa show a very
  good consistency in the humidity profile (q)
  between GPS RO and radiosonde results with a
  height resolution of several hundred meters.
• Height derivative of refractive index (dn/dz) is
  mostly determined by the humidity gradient
  (dq/dz) in a moist atmosphere. So, the behavior
  of thin water vapor layers can be studied by
  using dn/dz, without retrieving q by 1D-var.
• In the tropics, temporal and spatial variations
  of q are very large, Continuous monitoring of q
  and T with lidar/radar is important.
• GPS RO data is especially important over ocean in
  the Asian monsoon region, where routine balloon
  observations are sparse.

29
Thank you
30
Comparison of T and q profiles between GPS-RO,
radiosonde and AQUA-AIRS
31
Comparison between COSMIC and Singapore
radiosonde data
32
(No Transcript)
33
COSMIC GPS RO events within 500 km at night
Occultation event within 500 km Onna
127.850E 26.4990N 15-Jul-2006 123700
129.560E 22.6191N 15-Jul-2006 123700
129.574E 22.7052N 15-Jul-2006 123200
129.866E 25.3847N 16-Jul-2006 122100
128.226E 27.2834N 16-Jul-2006 122100
128.230E 27.3768N 22-Jul-2006 105100
130.293E 25.8163N 22-Jul-2006 105100
130.299E 25.9023N 22-Jul-2006 104500
130.404E 28.5873N 23-Jul-2006 103500
128.591E 29.9030N 19-Aug-2006 183400
127.637E 28.4792N 25-Aug-2006 180900
129.653E 27.1944N 28-Aug-2006 153700
130.501E 27.2085N 30-Aug-2006 152000
129.163E 23.2813N 31-Aug-2006 151000
128.141E 22.1868N
Ogimi 128.159E 26.6781N 06-Sep-2006
143600 131.756E 23.9913N 06-Sep-2006
143500 131.775E 24.0220N 06-Sep-2006
142800 132.327E 24.9348N 07-Sep-2006
142100 131.091E 25.6144N 07-Sep-2006
142100 131.155E 25.7232N 07-Sep-2006
141300 131.767E 26.8435N 08-Sep-2006
135900 131.350E 29.0928N 09-Sep-2006
135200 130.311E 30.0698N 09-Sep-2006
144900 126.045E 30.6546N 11-Sep-2006
135500 130.282E 29.2870N 12-Sep-2006
134100 131.124E 27.2821N 13-Sep-2006
140100 124.092E 25.9779N
34
Vapour Pressure, First and Second Derivative
Comparison between Gadanki Radiosonde and COSMIC
Gadanki 13.48N 79.18E 0013UT COSMIC 13.48N7
9.23E 0345UT
35
Cloud distribution looks horizontally
homogeneous, But ..
36
Cloud distributions (water vapor as well) in the
tropics are characterized by vigorous temporal
and spatial variations, which should be
considered in validating the GPS-RO results.
37
Okinawa Subtropical Environmental Remote-sensing
Center, Ogimi facility (26.7ZN, 128.1E) UHF
(443MHz) wind profiler L-band (1.3 GHz) wind
profiler
C-band (5.3GHz) Polarimetirc radar
Radiosonde Ceilometer Disdrometer GPS receiver
38
Time Height Variations of q (top) and T (bottom)
observed with a Raman Lidar from 0003-0033LT
(left), and 0150-0221 LT (right) on November
15, 2005
Water vapor mixing ratio (27m/90sec)
Temperature (45m/120sec)
39
6????????
11?????????????????????????????????