JAPANs GV Strategy and Plans for GPM

1
JAPANs GV Strategy and Plans for GPM

• K. Nakamura (HyARC/Nagoya Univ.)
• and S. Shimizu (JAXA)

2
Objectives of Japanese GPM Cal/Val

• To confirm the reliability of the GPM standard
  products,
• To quantify the error of the products and confirm
  the characteristics,
• To clarify the origin of the error of the
  products and feed it back to modify the
  algorithms and
• To validate the algorithms using the physical
  parameters observed or estimated from the ground
  validation activities.

3
PR algorithm concept
Stratiform
Height
Snow
PR
Melting Layer
Rain
Radar reflectivity
Rain attenuation ? Surface Reference Method Drop
Size Distribution ? External Parameter (In the
algorithm)
4
DPR algorithm concept
Detectable range of KaPR (35 GHz)
Detectable range of KuPR (14 GHz)
Height
Stratiform
Sensitive observation by the KaPR
Discrimination of snow and rain using
differential attenuation method
Snow
KuPR
KaPR
Melting Layer
Rain
Accurate rainfall estimation using differential
attenuation method (DSD parameter estimation)
Radar reflectivity
Ice/Snow Region insufficient for three
parameters (N0, D0, r)
5
GPM/DPR vs TRMM/PR on algorithm
Attenuation TRMM/PR (Ku-band) (Rain) DSD
uncertainty GPM/DPR KuPR
(Ku-band) (Rain) KaPR (Ka-band) (Rain)
(Cloud) (Water Vapor) (gases) Rain
attenuation correction will be improved. New
uncertain terms attenuation by cloud, water
vapor, and gases
Other difficulties Beam filling same as
TRMM/PR Beam matching new problem
6
GPM/DPR Calibration and Validation
Calibration (by ARC)
Transmit power,Received power,Antenna beam
direction
Assumption(Initial values)
Precip. type classification (Conv./Strat.),Partic
le type (Rain/Snow/Graupel), (DSD (Drop Size
Distribution)),Temp. humidity profile,Melting
layer model,Gaseous attenuation,
Precip. rate/accumulation,Precip. type
classification (Conv./Strat.),Particle type
(Rain/Snow/Graupel), DSD (Drop Size
Distribution) ,
Validation
7
From TRMM experiences

• Simple comparison is never enough.
• Ground-based radar data (especially radar
  reflectivity value) are depended on the radars.
• TRMM is too good to be validated by
  regression-based traditional validation.
• Temporal/spatial mismatching is still problem.
• Precise and comprehensive precipitation system
  measurement is required.
• Physical validation may be more important for
  radar rain retrieval as well as microwave rain
  retrieval.
• Very few occasions of simultaneous observations
  between GV instruments and satellite, especially
  PR.

8
Japanese GV activities

• Japanese calibration and validation will focus on
  DPR in GPM.
• More accurate and sensitive cal/val analyses will
  be required.
• Validation for snow rate will be required for
  DPR.
• Post-launch beam matching measurement between two
  radars (new task of external cal. for GPM/DPR)
  using multiple ARCs
• Algorithm specific validation for each rain
  retrieval algorithm of DPR will be required.
• For this purpose, we need to develop new paradigm
  of algorithm validation and collect many kinds of
  physical parameters for Special validation sites
  are required for the physical validation.
• ? We need to establish Super sites for DPR GV
• (Okinawa, Wakkanai)
• Statistical comparison with long-term
  precipitation data using operational data.
• For this purpose, we need to collect operational
  raingauge data (e.g. AMeDAS data) and other
  operational data.

9
GV New Paradigm Example with PR/DPR
True values in Nature
Reflectivity (Ze), Rain Rate (R)
Compare
Hydrometeor (Rain, Snow, Graupel, etc.)
Remote Sensing
GV algorithm
Rain Rate (R(h))
GV data
Vertical velocity (v(D))
DSD(h), v(D), Particle type, Zm, PWC, etc
Rain (snow) water content (PWC(h)) Density (?
(h)) Drop Size Disribution, etc
In-situ measurement
?
Compare
GV algorithm
Synthesized Nature
Retrival Numerical models
Reproduce physical parameters for forward
calculation from ground-based observation using
GV algorithms
DSD(h)
Assumption
v(D)
Particle types DSD, v(D) Non-Uniformity, etc.
Particle types
Compare
Water vapor Cloud water content (Liquid,
Solid) Oxygen Aerosol Sea Surface
Temperature Noise, etc
forward calculation
Zm14 Zm35
Rain rate (R(h))
Retrieval Algorithm
(Iguchi, 2004)
10
Key issues for success of GV activities

• How do we synthesize physical parameters from GV
  data?
• We need to collect appropriate observation data.
• We need to investigate and collect existing
  observation data. Whether are existing datasets
  enough for reproducing physical parameters for
  forward calculation or not?
• New observation for GV will be need before launch
  of GPM-Core satellite.
• We need to establish GV algorithms for
  reproducing physical parameters.
• We need to validate the physical parameters
  retrieved by GV observations.
• We need to make Zm data by forward calculation.

11
Candidates for GPM GV Supersite

• International Arctic Environmental Research
  Project Group
• Upper air observation by VHF radar

Wakkanai (45.5N, 142E)
Okinawa Subtropical Environment Remote Sensing
Center - C-band multiparameter radar, wind
profiler, etc.
Okinawa (26N, 128E)
12
Issues

• Validation for solid precipitation
• Algorithms and validation methods for retrieval
  of solid precipitation have not established.
  (Physical parameters for DPR algorithm
  development have not been clear.)
• Density, N0, D0 ? Snow rate
• N0 and D0 can be derived by dual frequency radar
  for rain rate. But we have three parameters for
  snow. Statistics of snow density is required.
• We will try to get upper layer data above melting
  level at Okinawa.
• Conventional method using polarization radar for
  the classification of solid particles.
• Spectrum differences in C, Ku, Ka and W for
  detection of terminal velocity of snow.
• We need to collect snow rate and other physical
  parameters in NiCT Wakkanai during winter season
  using wind profilers, Ku/W-band radars,
  multi-parameter radar, etc before launch of
  GPM-core satellite.
• Continuous validation analyses using statistical
  methods will be needed after the launch.

13
Summary

• DPR is steadily being developed by JAXA and NiCT
  for the launch of GPM-Core satellite in winter on
  2010.
• Japanese calibration and validation will focus on
  DPR in GPM.
• New GV paradigm for DPR is proposed. We are now
  designing Japanese GV plan based on the new
  paradigm.
• Construction of adequate physical parameter
  database for forward calculation is the most
  important and concerning problem.