GV for Taiwans precipitation retrieval research

1
GV for Taiwans precipitation retrieval research

• Wann-Jin Chen
• Chung Cheng Institute Technology of National
  Defense University, Taiwan

2
Outline

• Introduction
• Estimates of Rainfall using TMI data over the
  Ocean during Mei-Yu and Typhoon season
• Estimates of Rainfall using TMI data over the
  Land during Mei-Yu and Typhoon season
• Ground Validation
• Future work

3
Two Heavy rainfall Season

• Mei-Yu (Meso-scale convective systems), occurred
  during May and June.
• Typhoon occurred during August through October.
• Many natural disasters in Taiwan areas caused by
  the two weather systems, so the following studies
  were focused on them.

4
Estimates of Rainfall using TMI data over the
Oceanduring Mei-Yu season
5
Methodology Match-up with TMI_Tb and Rain Gauge
6
The regression of Multichannel during Mei-Yu
season
7
Typhoon SINLAKU
TMI-2002-09-05_1407
GMS-2002-09-05_1400
8
Quantification Validation during Mei-Yu season
19982001 18 points R2 0.812
9
Estimates of Rainfall using TMI data over the
Oceanduring Typhoon season
10
The regression over ocean
11
Rainfall Rate (mm/hr) by Regression. 37769
(this study) Typhoon MINDULLE
Rainfall Rate (mm/hr) by 2A12.
Delta Rainfall Rate (mm/hr).
12
Validation
R0.74, RMS3.75 mm/hr, Bias 0.68 (1.24) ,
Points66 (CO25, ST41)
CO ? Convective ST Stratiform
Average Rain Rate of Rain Gauge 4.9
mm/hr. Average Rain Rate of Estimated 5.6
mm/hr.
13
Validation (2A12)
R 0.45, RMS5.76 mm/hr, Bias2.52 (-0.29),
Points 66
14
Estimates of Rainfall using TMI data over the
Landduring Typhoon season
15
Data Collection
The scan patterns and resolutions of TRMM
sensors, including TMI, PR, and VIRS.
The illustration is a scatter of ground rain
gauge in Taiwan
16
Scattering Index over Land( SIL ) Equation

• Developing Scattering Index over Land( SIL )
  equation for Taiwan land area.
• Using TRMM/TMI 19.35V, 21.3V and 85.5V GHz
  channel.
• Under non-scattering atmosphere conditions, that
  identified weather as cloud-free on Taiwan region
  with GOES-9 IR images.

17

• TMI data set used in the study.

18

• The Rainfall Retrievals over Land
• The scatter diagrams of the SIL and Gauge Rain
  rate

Gauge Rain rate (mm/hr)
Sample74
SIL(K)
RR(rain rate) 0.126 SIL 1.239
19
LRCT(Land Rainfall Retrieval by Chen and Tsai)
20
Validating the Accuracy of Estimated Rainfall
over land _Typhoon AERE (sample23)
21
Validating the Accuracy of Estimated Rainfall
over land _Typhoon MINDULLE (sample74)
22
Error analysis time difference

• Satellite rainfall retrievals have a good result
  when comparing with rain gauge with 10 min. later
  after satellite overpass for the Mei-Yu cases.

23
Over the land _ Mei-Yu
Validating the Accuracy of Estimated Rainfall
with QPESUMS Data ( -/ before/after satellite
overpass)
24
RMS
June
May
( -/ before/after satellite overpass 0
same time )
25
Correlation Coefficient
June
May
( -/ before/after satellite overpass 0
same time )
26
(No Transcript)
27
The scatter diagrams of the estimated rainfall
and the Gauge Rain rate
Correlation Coefficient 0.82
Correlation Coefficient 0.81
Rainfall Retrievals derived from LRCT(2005)
(left), Ferraro (1994) (right) ( Black solid line
is the regression line between Est.-RR and rain
gauge RR, and the red solid line is for the
equivalent values between them. )
28
Error analysis terrain effect

• Satellite rainfall retrieval underestimated on
  upsloping side of the mountain and overestimated
  on the downsloping side of the mountain.

29
(No Transcript)
30
(No Transcript)
31
Future works

• Improve rain gauge QC.
• Develope more algorithms for other weather
  systems in Taiwan area.
• Use more passive microwave radiometer sensors,
  like AMSR, AMSU in our algorithm.
• Compare satellite rainfall retrievals with those
  derived from ground-based radar.

32
The End !Thanks for your attention!