Introduction to the COMS Meteorological Data Processing System Myoung-Hwan Ahn/RSRL 2005. 06. 02

CMDPS2005. June 2, APSDEU-6, Seoul

Introduction to the COMS Meteorological Data Processing

System

Myoung-Hwan Ahn/RSRL

2005. 06. 02


CONTENTS

I. Introduction

II. Baseline Products

III. Examples

IV. Calibrations

V. Future Plans


I. Introduction


COMS

1995

2002

2003

2004

2005

2006

2007

1996

1997

1998

1999

2000

2001

2008

2009

1995.8 STEPI “Meteorological Sensors for Satellite

1997.11 “Research on the Korean Meteorological Satellite”

2000.3 “Project for Centennial Meteorological Events”2000.8 Meteorological Satellite Forum2001.4 Feasibility study for COMS

2003.9 COMS Project Kick-off

System Design ReviewPreliminary Design ReviewCritical Design ReviewMission Readiness Review

Pre Ship ReviewLaunchData Service

2000.12 National Plan for Long-term Space Development

2002.4 Preliminary work for COMS


Requirement of Ka-band Payload(TBC)

Ka-band Service Coverage

58dBW

Refer to right figureCoverage

0.6degree/each beamBeamwidth

400MHz(100MHz/channel)Bandwidth

13dB/KG/T

Minimum EIRP edge of coverage

Uplink : 29.6 ~ 30.0GHz

Downlink : 19.6 ~ 20.3GHz

Frequency(Ka-band)

COMS Mission-1


2,500km X 2,500km

- Total : 8 times∙ 10:00 ~ 17:00 : 6 times,∙ 22:00, 02:00 : 2 times

Number of observation

- Calibration type : Solar Calibration - Accuracy of Radiometric Calibration : 3%

Sensor Calibration

11bitDynamic Range

0.3 at Nyquist frequencyMTF7500.01623.412.0408658600.02033.017.7407658700.03146.227.120670

10100.03258.332.020625

10700.05685.255.320555

11700.067115.572.220490

10900.085145.892.520443

10000.100150.010020412

SNRNEdLMax. Rad.[Wm-2

um1sr-1]

Nom. Rad

[Wm-2

um-1sr-1]

Band Width[nm]

Band Center[nm]

Band Center &

Band Width &Nominal Rad.&

Max Rad. &NEdL &

SNR (Sensitivityof sensor)

8 fixed bandNo. of BandCoverage

500m X 500mSpatial Resolution

RequirementsItem • Ocean Sensor Coverage

COMS Mission-2Requirement of Ocean Sensor(TBC)


COMS Mission-3 Meteorology

Continuous monitoring of weather events with multi-channel ImagerEarly detection of sever weather such as tropical cyclones, heavy

rainfall, dust out break, etc.Long term data acquisition for climate study

Channels

Channel

Spectral band(㎛ )

Application

VIS1 0.63Daytime cloud imagery

Detection of special event (yellow dust, fire, haze, etc.),Atmospheric motion vector

SWIR 3.75 Nighttime fog/stratus, Fire detection, Surface temperature

WV1 6.2 Upper atmospheric water vapor, Upper atmospheric motion vector

WIN1 10.8 Standard IR split window channel(cloud, Sea surface temperature, Yellow sand detection)

WIN2 12.0 Standard IR split window channel(cloud, Sea surface temperature, Yellow sand detection)


COMS System

Command,Telemetry

Raw Data,HRIT/LRIT

HRIT/LRIT

Satellite Operation Control Center/ Back-up Data Processing Center

Meteo/Ocean DataApplication Center(Primary Data Processing Center)

HRIT/LRITForeignMeteorological DataReceiving Station

CommunicationSystem MonitoringFacility

Specialized Organization/ Domestic User

Foreign Meteorological Organization/ Foreign User

Exclusive Line

Meteorological Information Supply

Various SiteInternet

Internet

Ka-band RF Signals

Raw Data


CMDPS

PIAI

Unprocessed

Cloud free land

Cloud free sea

Cloud contaminated

Cloud filled

Snow/Ice contaminated

Unclassified

Cloudy Clear

AMV CType SC/IceCTP

OLR

SSTFog

CLD

Pre-processing dataAux. Data, response function, etc.

Level 1.5 data

Real TimeValidation

DistributionArchiveApplications

Validation Data

TPW

CMDPS


CMDPS COMS Meteorological Data Processing System (CMDPS)

System to produce geophysical parameters from the satellite measured raw radiance data.

The system includes; algorithms for each baseline products various auxiliary data such as the surface emissivity, etc. radiative transfer model calibration monitoring scheme and algorithm interfaces for between algorithms, and between CMDPS and O

S validation procedures.


II. Baseline Products


Baseline productsGroup Product Spatial

ResolutionTemporal

ResolutionAccuracy(Thres.)

Remarks

Cloud Detection

CLD 1x1 pixel =obs METRI

CSR ~ 100km =obs METRI

AMV AMV ~ 50km 3 hr. 9 m/s METRI

Surface Information

SST 1x1 pixel hourly 1.2 K SNU

LST 1x1 pixel hourly 2 K Kongju NU

Sea Ice/SC 1x1 pixel daily METRI

Insolation 1x1 pixel hourly 10 % Pukyoung NU

Water vapor Information

UTH ~ 50km =obs 20 % Kyoungpook NU

TPW ~ 50km =obs 20 % Kyoungpook NU

Cloud Information

CLA 3x3 pixel =obs SNU

CTT/CTH 1x1 pixel or3x3 pixel

=obs 3 K1 km

SNU

Fog 1x1 pixel =obs EWU

PI 1x1 pixel =obs 20 % bias Kangnung NU

OLR ~ 10km hourly 10 % SNU

Aerosol Information

AI 1x1 pixel =obs Pusan NU

AOD 10x10 pixel =obs 35 % YSU


Name New? Issues RemarksCLD/CLR No Accuracy Improvement Probability Info., Use of NWP products, SFC1

LST Yes New with enough acc. SFC is mandatory for enough accuracy,

SST No Accuracy Improvement Different algorithms for spatial and temporal scale

Insolation Yes New regression coeffs. Better reference ground data and CLD information

Snow/Ice Yes Important in LSM Combining with SSM/I, AVHRR, and MODIS

AMV No Accuracy Improvement Height assignment using IR1/WV, IR1/IR2

TPW Yes NWP application Regression with better database

UTH Yes NWP application Regression and Physical approach

CLD Info. No Accuracy Improvement Use of NWP products, better understanding of Chs.

CLD Cls. No Accuracy Improvement Use of NWP products with better RTM

CLD Phs. Yes Important for nowcasting Better understanding of Chs.

Fog No Accuracy Improvement Better threshold values for different conditions

OLR Yes Expect better accuracy Better calibration with improved aux. data

PI No Accuracy Improvement Use Microwave data and precise topography

AI No Accuracy Improvement Better threshold and SFC information

AOD Yes Became important info. Accurate LUT and SFC information1 SFC includes the surface information acquired by both off-line and on-line

Baseline products


Sea Ice/Snowdetection

OLR

TPW UTH

Cloud Amount

Cloud Phase

AMV

SST

Insolation

LSTAOD

Aerosol Detection

PI

Cloud TypeCOD

CTT&CTH

Fog

Cloudy Clear

COMS Level 1B Image data Auxiliary Data

Cloud Detection

OLR: Outgoing LW RadiationAOD: Aerosol optical depthCOD: Cloud optical depthCTT: Cloud top temp.CTH: Cloud top height

TPW: Total precipitable waterUTH: Upper Tropospheric HumidityAMV: Atmospheric motion vectorPI: Precipitation index

Baseline products


III. Examples


Cloud Phase-I Key issues

Is WV channel useful for cloud phase How accurate for multi-layer cloud ?

Approaches Simulation

8.7 vs. 6.3 Application to MODIS data

comparison

Baum et al.(2000, JGR)


Cloud Phase-II


Cloud Phase-III


AODa)

Look Up Table0 = 0-70o ,5o

0 = 0-180o ,10o

s= 0-70o ,5o

a = 0-3 ,0.1s = 0-1, 0.1

RTM (6S)a= F(0 , 0s,p,s)

COMS Vis. ChannelTOA Reflectance (’p)

Cloud Removal

LOOK UP TABLE METHOD’s= F(0 , 0s,p)b=0.0 … in 6S

COMS Vis. ChannelTOA Reflectance (p)

GOES-9 Vis. ChannelSurface Reflectance (’s)

s (1 month 2nd min. reflectance)

Aerosol Optical Depth-I


Aerosol Optical Depth-II


Aerosol Optical Depth-IV


Comparison of Aeronet with GMS5 AOD(2002/3/30-4/30)

Aeronet AOD

0.0 0.5 1.0 1.5 2.0 2.5 3.0

GM

S5

AO

D

0.0

0.5

1.0

1.5

2.0

2.5

3.0Shirahama(lat=33.69, lon=135.36)Osaka (lat=34.65, lon135.59)Che-Ju(lat=33.28, lon=126.17)Anmyon(lat=36.322, lon=126.195)

Y=0.9381454401X+0.0389830969r2=0.6327175486

Aerosol Optical Depth-V


Source

Atmospheric aerosol optical properties (e.g., sphericity,wo, n(r), refractive index)

location of the aerosol layerRayleigh optical depthgaseous absorption

Surface reflectance uncertaintybidirectional reflectance contamination

Instrument calibrationnoise

Radiative transfer model plane-parallel approximationmultiple scattering

Sources of uncertainty(Knapp et al., 2002)

Aerosol Optical Depth-VI


IV. Calibration


Visible Channel Calibration-I There is onboard calibrator for the infrared channels There is no onboard calibrator for visible channel, although accurate

calibration is required for the derivation quantitative values such as the aerosol optical depth and for climatic applications

For the real time monitoring/update of the visible channel calibration coefficients, we are going to take the vicarious calibration approach

The main concept for the approach originates from the current EUMETSAT visible channel calibration method, which compares the measured radiance and theoretically derived radiance

Thus, for the beginning, we builds up the theoretical model for the calculation of the theoretical radiance

Also, great efforts are given to find out bright and stable surface target for the determination of the slope of the calibration curve.


16 days of Data

TargetIdentification

COMSobservation

RTM(6S)

Quality ControlCalibration

COMSL1.5/2.0

(other Satellite)

ECMWF

CLIMATE

Satellite levelradiance

Pixelextraction

CalibrationCoefficient

Mean digitalcount

Visible Channel Calibration-II


6S(RTM) Input Data Geometry

Solar Zenith and Azimuth Angles Sensor Zenith and Azimuth Angles Date

Atmosphere Total Precipitable Water Total Ozone Aerosol Characteristics (Type, AOT)

Target BRDF(Bidirectional Reflectance Distribution Function) Height

Sensor Band, Response function

Visible Channel Calibration-III


Visible Channel Calibration-IV

R @ band 1 NDVI

STDV of R@ Band 1

NormalizedSTDV ofR @ Band 1

0.005 0.01 0.015 0.02 0.03 0.04 0.05 0.1


12Apr20040040

Visible Channel Calibration-V


Case 1 (25.05S, 137.05E)

100 105 110 115 120 125

Rad

ianc

e

6080

100120140160180200

6SMODIS

Julian day in 2004100 105 110 115 120 125

Err

or

-0.4

-0.2

0.0

0.2

0.4

Case 2 (28.45S, 138.55E)

100 105 110 115 120 125

Rad

ianc

e

6080

100120140160180200

Julian day in 2004100 105 110 115 120 125

Err

or

-0.4

-0.2

0.0

0.2

0.4

6SMODIS

Visible Channel Calibration-VI3 X 3 grids average (0.1o X 0.1o) for TERRA MODIS Band01 [620~670nm]Error = R6S-RMODIS / RMODIS


Case 3 (26.85S,138.05E)

100 105 110 115 120 125

Rad

ianc

e

6080

100120140160180200

6SMODIS

Julian day in 2004100 105 110 115 120 125

Err

or

-0.4

-0.2

0.0

0.2

0.4

Case 4 (24.05S, 138.05E)

100 105 110 115 120 125

Rad

ianc

e

6080

100120140160180200

6SMODIS

Julian day in 2004100 105 110 115 120 125

Err

or

-0.4

-0.2

0.0

0.2

0.4

Visible Channel Calibration-VII


Visible Channel Calibration-VIII Other approaches

Using the bright cloud targets Using the cloud albedo characteristics Clear ocean targets with in-situ observation

Expected accuracy Preliminary results show that absolute accuracy of 10% is relativel

y easily achieved With well known surface characteristics of the target area, 5% is a

chievable Further works are required for the characterization of uncertainty s

ources and reducing the uncertainty


V. Future plan


Future Plans2003

2004

2005

2006

2007

2008

Conceptual Design

S/W Development

S/W Standardization

CMDPS

Integration to OS

Validation

Real time operation

Interface design

Interface develop.

Simulation Data

Evaluation Data

Cal. concept

Cal. Development(I)

Cal. Development(II)

S/W Prototype

Other major TaskDevelopment of validation strategy for the raw and derived productsPreparation for the user service

Documents

Introduction to the COMS Meteorological Data Processing System Myoung-Hwan Ahn/RSRL 2005. 06. 02