Establish a DSD parameter estimation method by combining both CloudSat and TRMM/PR data with statistical method. Demonstrate a visualization method of global microphysical data Climate map of DSD parameter as well as other meteorological parameters. To see the climatological characteristics in term of cloud and precipitation interaction 10 x 10 deg. box Algorithm validation for GPM/DPR, especially for DSD estimation. Utilization of W-band radar data (CloudSat and future EarthCARE) Current target is validation of DSD estimation by TRMM/PR . Purpose of this study Introduction (Motivation of this study: goal) Major specifications of CloudSat/CPR and TRMM/PR CloudSat/CPR TRMM/PR Orbit Sun-synchronous orbit Altitude about 700 km (local time 1:30) Non sun-synchronous orbit (Inclination = 35 degrees) Altitude about 400 km Swath width (footprint size) Nadir (1.3 km) about 240 km (5 km) Vertical resolution 500 m (240 m sampling) 250 m (partially 125 m sampling) Frequency 94 GHz 13.8 GHz Minimum detectable echo -30 dBZ 17 dBZ Product used in this study 2BGROPROF 2A25 Ver. 6 Data available 2006.8-2011.1 2011.12-current 1997.12-current 36 th Conference on Radar Meteorology (16-20 September, 2013) An analysis method to estimate the climatological drop size distribution (DSD) parameter using spaceborne radars (CloudSat’s w‐band radar and TRMM’s Ku‐band radar) is proposed in this study and a global maps of estimated DSD parameters (e.g. median diameter, D0) are demonstrated. The basic idea of the analysis method is to compare the histogram of radar reflectivity factor (Z) at near‐surface range bin in the over‐lapping Z range (weak to moderate rain echo). Because of the Mie effect, the Z value of rainfall is different between w‐ and Ku‐band and it reflects the difference in the histograms of w‐ and Ku‐band. A general characteristic of global maps of D0 through the year and local time is apparent land‐ocean contrast; larger D0 appears over land and smaller D0 appears over ocean except for relatively small D0 over southeastern Asia to China. Also, relatively larger D0 appears in tropical area and mid latitude summer. Diurnal change of D0 can be seen by comparing the day/night time D0; D0 is larger in the night time over ocean while day time D0 is larger over land. Tropical Ocean shows smaller seasonal change, while larger change are seen over mid‐latitude area. Difference in histogram of Z between CloudSat and TRMM/PR comes from the Mie scattering effect and attenuation. Mie scattering correction closely relating to the DSD (model). Attenuation correction can be done using surface echo (sigma zero) if we consider only surface rain. Reference sigma zero data are prepared using no-cloud data (1 x 1 deg.) Note: this study is targeting the weak to moderate rain where the both radar can observe with certain quality. Key idea σ0 database −10 0 10 20 −10 0 10 20 sigma zero (TRMM) (dB) sigma zero (CloudSat) (dB) ANN 2007 (source: 1 x 1 deg. grid average data) Ocean Land Histogram matching Data Result (Example of estimated D0) 0 1 2 0 0.5 1 DFR (W/Ku) D0 (mm) DFR Rain, Gamma DSD ( =0) DFR =ZW/ZKu 0 0.1 10 20 30 40 0 0.1 dBZ cumulative probability DJF 2006−2007, ascending, lon: 110−120 lat: 20N to 30N (land) lat: 20N to 30N (ocean) original histogram Mie correction with various DSD model CloudSat : convert W‐band histogram to Ku‐band histogram with various DSD models (Mie correction) TRMM: no change Threshold Z (Ku‐band equivalent) Threshold Z Probability Probability k Z k Q t D ND dD N 0 Q t D D e 3.67 /D 0 dD Z D ND dD N 0 D D e 3.67 / D 0 dD N 0 Q t D D e 3.67 /D 0 dD N 0 D D e 3.67 /D 0 dD N 0 1 D D e 3.67 /D 0 dD Q t D D e 3.67 /D0 dD k‐Z relationship （α, β: constant） Qt: extinction cross section, N(D): number concentration, D: diameter sigma: backscattering cross section k‐Z adjustment equation in TRMM 2A25 (ε‐k‐Z relationship) k Z ‐‐‐ε adjusts the α in k‐Z relationship original histogram Comparison determine the most appropriate DSD model Threshold Z Probability DSD models DSD model defined by k‐Z (and ε） relationship gamma DSD with various μ factors (single D0 in fixed gamma DSD model) Fitting result and coverage of DSD model, sing le D0 model is the most appropriate simply offset the histogram of CloudSat. Compare the Z‐histogram of Ku‐band. Equations In order to estimate the path integrated attenuation, no‐cloud (no‐ rain) sigma zero databases are created from CloudSat (TRMM) data. Global dataset: 3-month average, 10 x 10 degree grid (40S to 40N), Land/Ocean, daytime/nighttime December 2006 to November 2007 20 dBZ← →40 dBZ CloudSat 2B GEOPROF TRMM 2A25 Dataset Z, sigma0 Dataset Z, sigma0 histogram (various DSD models) histogram histogram matching (DSD estimation) Sigma‐zero Database Dataset attenuation corrected Z (the lowest clutter‐free bin) (Top) Afternoon (0-3 P.M.= ascending node of CloudSat) (Bottom) Descending - Ascending cloud –free cases clear land/ocean contrast larger diurnal change (day/night contrast) over maritime continent and Amazon smaller diurnal change except for several area (e.g Indian ocean in SON) P207 ‐‐‐ Start point of DSD model ‐‐‐ Gamma DSD model is used ‐‐‐ equation to determine N0 (function of D0 or mu) Relationship among D0, R, Z and ε ) ( ) ( band Ku Z band w Z DFR
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Establish a DSD parameter estimation method by combining both CloudSatand TRMM/PR data with statistical method.
Demonstrate a visualization method of global microphysical data Climate map of DSD parameter as well as other meteorological parameters.
To see the climatological characteristics in term of cloud and precipitation interaction 10 x 10 deg. box
Algorithm validation for GPM/DPR, especially for DSD estimation.
Utilization of W-band radar data (CloudSat and future EarthCARE)
Current target is validation of DSD estimation by TRMM/PR .
Purpose of this studyIntroduction (Motivation of this study: goal)
Major specifications of CloudSat/CPR and TRMM/PR
CloudSat/CPR TRMM/PR
Orbit Sun-synchronous orbitAltitude about 700 km(local time 1:30)
Non sun-synchronous orbit(Inclination = 35 degrees)Altitude about 400 km
Swath width(footprint size)
Nadir (1.3 km)
about 240 km(5 km)
Vertical resolution 500 m(240 m sampling)
250 m(partially 125 m sampling)
Frequency 94 GHz 13.8 GHz
Minimum detectable echo
-30 dBZ 17 dBZ
Product used in this study
2BGROPROF 2A25 Ver. 6
Data available 2006.8-2011.12011.12-current
1997.12-current
36th Conference on Radar Meteorology (16-20 September, 2013)
An analysis method to estimate the climatological drop size distribution (DSD) parameter using spaceborne radars (CloudSat’s w‐band radar and TRMM’s Ku‐band radar) is proposed in this study and a global maps of estimated DSD parameters (e.g. median diameter, D0) are demonstrated. The basic idea of the analysis method is to compare the histogram of radar reflectivity factor (Z) at near‐surface range bin in the over‐lapping Z range (weak to moderate rain echo). Because of the Mie effect, the Z value of rainfall is different between w‐ and Ku‐band and it reflects the difference in the histograms of w‐ and Ku‐band.A general characteristic of global maps of D0 through the year and local time is apparent land‐ocean contrast; larger D0 appears over land and smaller D0 appears over ocean except for relatively small D0 over southeastern Asia to China. Also, relatively larger D0 appears in tropical area and mid latitude summer. Diurnal change of D0 can be seen by comparing the day/night time D0; D0 is larger in the night time over ocean while day time D0 is larger over land. Tropical Ocean shows smaller seasonal change, while larger change are seen over mid‐latitude area.
Difference in histogram of Z between CloudSat and TRMM/PR comes from the Mie scattering effect and attenuation. Mie scattering correction closely relating to the DSD (model).
Attenuation correction can be done using surface echo (sigma zero) if we consider only surface rain. Reference sigma zero data are prepared using no-cloud data (1 x 1 deg.)
Note: this study is targeting the weak to moderate rain where the both radar can observe with certain quality.
Key idea
σ0 database
−10 0 10 20−10
0
10
20
sigma zero (TRMM) (dB)
sig
ma
zero
(C
lou
dS
at)
(dB
)
ANN 2007(source: 1 x 1 deg. grid average data)
Ocean
Land
Histogram matching
Data
Result (Example of estimated D0)
0 1 20
0.5
1
DFR (W/Ku)
D0 (mm)
DF
R
Rain, Gamma DSD (=0)
DFR =ZW/ZKu
0
0.1
10 20 30 400
0.1
dBZ
cum
ula
tive
pro
bab
ilit
y
DJF 2006−2007, ascending, lon: 110−120
lat: 20N to 30N (land)
lat: 20N to 30N (ocean)
original histogram
Mie correction with various DSD model
CloudSat : convert W‐band histogram to Ku‐band histogram with various DSD models (Mie correction)
TRMM: no change
Threshold Z (Ku‐band equivalent)
Threshold Z
Probab
ility
Probab
ility
k Z
k Q tD N D dD N0 Qt D D
e 3.67 /D0 dD
Z D N D dD N0 D D e 3.67 /D0dD
N0 Qt D De 3.67 /D0 dD N0 D D e 3.67 /D0 dD
N01
D D e 3.67 /D0 dD
Qt D De 3.67 /D0 dD
k‐Z relationship （α, β: constant）
Qt: extinction cross section, N(D): number concentration, D: diametersigma: backscattering cross section
k‐Z adjustment equation in TRMM 2A25 (ε‐k‐Z relationship) k Z ‐‐‐ε adjusts the α in k‐Z relationship
original histogram
Comparison
determine the most appropriate DSD model
Threshold Z
Probab
ility
DSD modelsDSD model defined by k‐Z (and ε） relationshipgamma DSD with various μ factors(single D0 in fixed gamma DSD model)
Fitting result and coverage of DSD model, sing le D0 model is the most appropriate simply offset the histogram of CloudSat.
Compare the Z‐histogram of Ku‐band.
Equations
In order to estimate the path integrated attenuation, no‐cloud (no‐rain) sigma zero databases are created from CloudSat (TRMM) data.
Global dataset:3-month average, 10 x 10 degree grid (40S to 40N), Land/Ocean, daytime/nighttime
December 2006 to November 2007
20 dBZ← →40 dBZ
CloudSat2B GEOPROF
TRMM2A25
DatasetZ, sigma0
DatasetZ, sigma0
histogram(various DSD models)
histogram
histogram matching(DSD estimation)
Sigma‐zeroDatabase
Datasetattenuation corrected Z(the lowest
clutter‐free bin)
(Top) Afternoon (0-3 P.M.= ascending node of CloudSat)(Bottom) Descending - Ascending
cloud –free cases
clear land/ocean contrast
larger diurnal change (day/night contrast) over maritime continent and Amazon
smaller diurnal change except for several area (e.g Indian ocean in SON)
P207
‐‐‐ Start point of DSD model
‐‐‐ Gamma DSD model is used
‐‐‐ equation to determine N0(function of D0 or mu)
Relationship among D0, R, Z and ε )(
)(
bandKuZ
bandwZDFR
Evaluation （Comparison of D0 with single frequency precut = TRM/PR only)
36th Conference on Radar Meteorology (16-20 September, 2013) P207 Drop Size Distribution Model Estimation by Comparing CloudSat/CPR and TRMM/PR（N. Takahashi）
• Over land/Daytime
• Seasonal change (DJF, MAM, JJA, SON) of climatological D0 and rain rate （December 2006 to November 2007）
• Over land/Nighttime
• Over water/Nighttime
• Over water/Daytime
– D0 (CloudSat & TRMM) : black circles
– D0 (TRMM only) : colored circles
– bars: rain rate (mm/h)
• Diurnal change – TRMM only
–Convective vs. Stratiform
–over land
–2007
SST (black solid line)Data source: NOAA
Wind speed (open bar)Data source: J‐OFURO2
R (solid bar)
Data source: TRMM/PR
ε (color solid line)Data source: TRMM/PR
January to December
each box represents the 10 x 10 deg. area
36th Conference on Radar Meteorology (16-20 September, 2013) P207 Drop Size Distribution Model Estimation by Comparing CloudSat/CPR and TRMM/PR（N. Takahashi）
• Diurnal change – TRMM only
–Convective vs. Stratiform
–over water
–2007
• Monthly change – TRMM only
–Convective vs. Stratiform
–over water
–2007
• Monthly change – TRMM only
–Convective vs. Stratiform
–over land
–2007
• Monthly change – TRMM (rain)
–NOAA (SST)
– J‐OFURO (Sea surface wind)
–over water
–2007
Related Documents
