Nobuhiro Takahashi1, Kazumasa Aonashi2, Toshio Iguchi1, Koyuru Iwanami3, Tomoo Ushio4 and Kenichi Ok

1
EGU06 Symposium NH1.06 Progress in Hazardous
Storm Research using Evolving Global
Precipitation Measurement (GPM) Global Water
Cycle (GWC) Remotely Sense Datasets
The Global Satellite Mapping of Precipitation
(GSMaP) Project

• Nobuhiro Takahashi(1), Kazumasa Aonashi(2),
  Toshio Iguchi(1), Koyuru Iwanami(3), Tomoo
  Ushio(4) and Kenichi Okamoto(5)

• National Institute of Information and
  Communications Technology (NICT)
• Meteorological Research Institute (MRI)
• National Institute for Earth Science and Disaster
  Prevention (NIED)
• Osaka University
• Osaka Prefecture University

2
The GSMaP Project

• GSMaP Project
• Started 2002 under JST/CREST (- Nov. 2007)
• 25 members, lead by Prof. Kenichi Okamoto (Osaka
  Prefecture Univ.)
• Goals of the Research
• Production of high precision and high resolution
  global precipitation map by using satellite-borne
  microwave radiometers.
• e.g. spatial resolution0.1º0.1º, temporal
  resolution1 day
• Microwave radiometers (TMI, AMSR-E, SSM/I3)
• Precipitation radar, GEOs visible and IR
  radiometers
• Development of reliable microwave radiometer
  algorithm
• Based on the common physical precipitation model
  with precipitation radar.
• Precipitation map production technique for the
  coming GPM satellites around 2010.

Core Research for Environmental Science and
Technology sponsored by Japan Science and
Technology Agency
3
GSMaP Team Members
Ground Radar Observation Group K. Iwanami (GL),
K. Nakagawa, H. Hanado, Y. Kitamura, Y. Shusse

• 4D data base of cloud physical parameter
• Validation of algorithms

• Precipitation profile model
• Melting layer model
• DSD model

Physical Precipitation Model Group N. Takahashi
(GL), S. Sato, J. Awaka, T. Kozu, Y. Takayabu,
M. Hirose
P. I. K. Okamoto

• Scattering algorithm
• Rain/No rain classification
• Improvement of CRM

Precipitation Retrieval Algorithm Group T. Iguchi
(GL), K. Okamoto, S. Seto, S. Shimizu, K.
Aonashi, H. Eito, T. Inoue

• Global rain mapping
• Evaluation of products
• MWRIR combined algorithm

Global Precipitation Map Group K. Okamoto (GL),
T. Ushio, S. Shige, R. Oki, H. Hashizume, T.
Kubota, M. Kachi, Y. Iida
GL Group Leader
4
Structure of GSMaP Project
Ground Radar Observation Gr.
Algorithm Gr.
Look-up Table
MWR Obs. Data
Physical Precip. Model Gr.
Routine Obs. Campaign Obs. Database
Precipitation Retrieval
feed back
validation
Precipitation Map Products
Global Precipitation Map
TRMM/PR
comparison
High Temporal Resolution Map
Obs. Data
Geo-IR data (MTSAT, GOES, MSG)
MWRIR algorithm
Precip. Database
Radar Algorithm
Global Precip. Map Gr.
5
GSMaP algorithmsImprovements and development

• MWR algorithm --- upgrade the algorithm developed
  by Aonashi and Liu (2000).
• Scattering algorithm --- dual frequency method
• Rain/No-rain classification algorithm over land
• Precipitation profile model
• Melting layer model
• Rain drop size distribution model
• Snow particle model
• MWRIR combined algorithm for high temporal
  resolution map ---use IR-based motion vector to
  advect the rain area of MWR product.
• Motion vector from Geo-IR data
• Motion vector rain estimation with Kalman filter

6
Major updates of GSMaP algorithm(for TMI rain
retrieval )
7
Improvement of Land AlgorithmDevelopment of
Rain/No Rain Classification Algorithm Using
Database
V3.2
Rain/No-rain Determination Rgt0
Rain
Observation Data TB(85V)
Database of TB under No-Rain condition by using
PR and TMI - 1 x 1 deg. / Monthly - Gives
regression parameter (a, b ) and stdev. of the
data base (se)
Retrieval
Rain/No-rain Determination Rain, if
TBe(85V)-TB(85V)gt0
Threshold of No-Rain TB TBe(85V) abTB(22V) -k0
se
Rain rate R0
No-rain
Algorithm
Rain estimation error caused by rain/no-rain
classification (mm/month)
Evaluation
GPROF
GSMaP
1
1. Underestimation in the mid-latitude
1
2
month
month
2. Underestimated in the semi-arid area in summer
3
3. Overestimation in the semi-arid area in winter
8
Development of Precipitation Profile Model
V4.3
Classification of Precipitation Type (Takayabu
Katayama)
Precipitation type database (PR)
(Land) 1 afternoon shower, 2 shallow rain,
3 Midlat front. systems, 4 organized
rain systems, 8 Tibet (Ocean) 5 shallow rain,
6 Midlat front. Systems, 7
organized rain systems
2.5º grid Every 3 months 8 types (ocean3 land5)
Type-1 Afternoon shower type
Precipitation Profile (Hirose)
Height measured from1ºC height (km)
Precipitation profile database (PR)
Integrated in each rain type Prepared for
various surface rainfall rate
Rainfall rate mm/h
Rainfall rate profiles classified by rainfall
rate (Type-1) 0.5, 1, 2, 3, 4, 6, 8, 10, 15, 20,
30, 40, 60, 80, 120, 160, 200 mm/h
9
Dual frequency algorithm over land (Aonashi)
V4.6
Motivation
MCS over West Africa 1998.07.03 orbit 3429
PR NSR
PCT37
Echo Top
PCT85
PCT85 (polarization corrected temperature at 85
GHz) is rather similar to the echo top height
pattern in PR data, while PCT 37 is similar to
PRs NearSurfaceRain (NSR). This result indicates
that the PCT85 represents the temperature near
the echo top for deep convective system.
Algorithm --- Combination of PCT85 and PCT37
instead of using only PCT85 to avoid the
saturation in PCT 85
W85weight W371-weight
rainpct3785 W85rain85W37rainpct37
10
Melting layer model(Takahashi Awaka)
V4.6BB
Melting layer model
Motivation

• To establish the algorithm which has common
  physical model with PRs algorithm.

• Nishitsuji model which is used for the melting
  layer model of TRMM/PR is used to create the
  lookup table.
• Number of layers for radiative transfer
  calculation increased from 14 to 33, 19 layers
  are allocated for the melting layer with 50m
  intervals.

Relationship between PRs rain rate (2A25)
averaged over 10GHz foot print and observed 10
GHz TB (1B11)

• Melting layer model shows better representation
  than rain only model.
• The model calculation is close to the average.
• Difference is clearer for low FH cases.

11
Comparison of Zonal Rain Rate Average(Ocean,
January 1998)
PR 2.2 3.2 4.3 4.6
12
Comparison of Zonal Rain Rate Average(Land,
January 1998)
PR 2.2 3.2 4.3 4.6
13
Utilization of Melting Layer Model
PR V6 GSMaP V4.6 TMI GSMaP V4.6 TMI with melting
layer model

• Overestimates of the GSMaP over winter
  mid-latitude oceans are related to too low
  freezing level heights.
• In experimental results using the algorithm
  integrated with the melting layer model, the
  overestimates decreases largely.

14
MWRIR combined algorithmGSMaP_MV algorithm
Infrared (IR) Data
T0
T0 1hr
Cloud Motion Vector (1hr intervals)
Advection of rain system by Cloud Motion Vector
IRMWR combined
Global Rain Map _at_ TT01hr
Global Rain Map _at_ TT0
GSMaP_MV
Update the rain map by the latest MWR data
Past 1 hour MWR rain retrievals
Microwave Radiometer (MWR)
15
Global precipitation map of GSMaP MV (0.1degs., 1
hour September 2003)
16
Status of satellite-borne microwave radiometer
data processed by GSMaP algorithm
Table1 Status of each satellite data
Table2 Status of each products
Various (6-hourly, 1 day, 1 month)
17
Global precipitation map observed by TMI (monthly
rain rate average Jan 1998Dec 2004)
18
Research Activities in GSMaP

• Introduction of DSD model.
• Development of SSM/I algorithm.
• Join the international PEHRPP (Pilot Evaluation
  of High Resolution Precipitation Product) group
  and evaluate various precipitation maps around
  Japan using Radar-AMeDAS data.
• Develop a rain retrieval algorithm by combining
  TRMM/PR and microwave radiometer algorithms.
• Estimation of DSD profile by using 400MHz wind
  profiler, vertical pointing 35 GHz Doppler radar,
  and micro rain radar.
• Characteristics in the parameters of
  polarimetric radar in convective and stratiform
  rainfall
• Evaluation of retrieved rainfall by using a
  radiative transfer model.
• Improvement of cloud resolving model by comparing
  the brightness temperature between observation
  and simulation.
• Development of 3D-radiative transfer model.
• IR-base algorithm using split window data.