METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
Algorithm Theoretical Basis Document
（二行） for Cloud Top Height Product
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
MOURI Kouki*, SUZUE Hiroshi*, YOSHIDA Ryo* and IZUMI Toshiharu**
（二行）
IMAI Takahito* and YOSHIDA Ryo**
（三行）
Abstract
The cloud top height product is an element of the fundamental cloud product, which
incorporates cloud mask, cloud type/phase and
cloud top height. The Meteorological Satellite Center
Abstract
(MSC)This
of Japan
Meteorological
Agency
(JMA)
separated
cloud type/phase
andProduct
cloud top
paper describes the algorithm theoretical basis cloud
for themask,
Himawari-8
Cloud Mask
height,
which
were
contained
in
miscellaneous
level
2
products
in
the
Himawari-7
era,
and
put them
(CMP) developed by the Meteorological Satellite Center. CMP is part of the Fundamental Cloud
together
as awhich
fundamental
cloudcloud
product.
Thetype
cloud
product
the three
elements
Product,
incorporates
phase,
andtop
topheight
altitude
from incorporates
Himawari-8/AHI
and has
been of
cloud
top
height,
temperature
and
pressure,
and
its
algorithm
is
based
on
EUMETSAT’s
NoWCasting
operational since 7 July 2015.
Satellite
Application
Facility
satelliteofdata
from Advanced
Himawari Imager
The
CMP algorithm
is (NWCSAF).
based on the The
cloudinputs
maskare
technique
NWC-SAF
for MSG/SEVIRI
and
(AHI)
observations,
vertical
profile
data
from
a
radiative
transfer
model
and
cloud
type
data from
cloud
NOAA/NESDIS for GOES-R/ABI. Most cloud detection tests involve threshold methods
baseda on
type/phase
of two profiles.
weeks from
20 July to are
2 August
2015
radiativeproduct.
transfer Evaluation
calculation conducted
using NWPover
dataaasperiod
atmospheric
The thresholds
modified
showed
that
cloud
top
height
was
underestimated
in
comparison
to
corresponding
values
from
MODIS
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
(theresults
Moderate
Resolution
Imaging
Spectroradiometer)
andcomparison
Calipso (Cloud-Aerosol
and CMP
Infrared
showed
that the CMP
hit ratio
derived from such
was more than Lidar
85%. The
Pathfinder
Satellite
Observation).
algorithm was applied to SEVIRI data, with results showing hit ratios with the MODIS product
were around 85% for all seasons.
（三行）
1. Introduction
（一行）
Himawari-8 is the new geostationary satellite of the
The cloud top height product provides some of the basic
Japan Meteorological Agency (JMA). It was launched on
information for level-2 products, and some cloud
7 October 2014 and began operation on 7 July 2015. The
products carries
of the Meteorological
Satellite Center
of
satellite
the Advanced Himawari
Imager(MSC)
(AHI),
Japan isMeteorological
Agency
(JMA)
has intheir
which
greatly improved
over past
imagers
termsown
of
individual
top and
height
retrieval procedure.
Since
its
number cloud
of bands
its temporal/special
resolution
(Bessho
et was
al. launched
2016). in
Using
data, operation
JMA’s
Himawari-8
2014 AHI
and began
Meteorological
Satellite
Center
(MSC)
developed
the
on 7 July 2015, MSC has separated cloud mask, cloud
Himawari-8 Cloud Mask Product (CMP).
type/phase and cloud top height from miscellaneous
Cloud mask is used to discriminate cloudy pixels
cloud products and put them together into a single output
from clear ones in satellite data. Numerous satellite
called
therequire
fundamental
cloud information.
product. It isByrational
products
cloud mask
way ofto
create
level-2
using existing(SST)
cloud(Yasuda
mask, cloud
example,
Sea products
Surface Temperature
and
type/phase
and cloud
top Radiance
height information.
The
Shirakawa 1999),
Clear Sky
(CSR) (Uesawa
2009) and Aerosol
al. 2003)
are
fundamental
cloudDetection
product (Okawara
is also et
used
for High
derived
in
clear
pixels.
Conversely,
Cloud
Grid
Resolution Cloud Analysis Information (Suzue et al.
Information
(Tokuno
is processed
cloudy
2016),
which
is a 2002)
level-2
product. in
With
16 pixels.
bands,
There were no MSC cloud mask products before the
Himawari-8 additionally has better observation
start of Himawari-8’s operation; each satellite product
performance
than Himawari-7
As aisresult,
had its own cloud
mask process.(MTSAT-2).
As cloud mask
also
previous
cloud
top
height
estimation
data
from
required for most Himawari-8 products, MSC decided to
Himawari-7
canasnoalonger
be put
to practical
use by MSC.
develop CMP
resource
available
for Himawari-8
products in common.
CMP isdata
a partfrom
of the Fundamental
Multi-band
observation
geostationary
Cloud
Product
(FCP),
which
contains
cloud
phase, type
meteorological satellites (e.g., Cloud Top Temperature
and top height for Himawari-8/AHI pixels (Mouri et al.
2016a, 2016b). FCP has been produced since Himawari-8
began
operation.
and Height
(Meteo France 2012) and the ABI Cloud
This
paper
describes
the algorithm
theoretical
Height Algorithm
(Heidinger
2012))
were basis
used forin
CMP.
previous studies. MSC uses the algorithm adopted for
（一行）
Cloud Top Temperature and Height (Meteo France 2012)
2. Algorithm
because it is well established and relatively easy to
understand.
2.1
Overview
This
document
outlines
theCMP
cloudalgorithm
top heightisproduct
Cloud
detection
in the
based and
on
comparison
data and
clear sky
data
its algorithm.ofTheobservation
product includes
information
on cloud
calculated
from
numerical
weather
prediction
(NWP)
top height, cloud top temperature and cloud top pressure.
data. If observation data differ from clear sky data, the
The document also describes the required input data,
pixel is cloudy. To discriminate between cloudy and clear
practical considerations and related evaluation. The
pixels, it is desirable for observation variables in clear
procedure
the product’s
derivation
is based
on the
and
cloudyforconditions
to differ
and for
the physical
NWCSAF
algorithm
France
reason
behind
the (Meteo
difference
to 2012).
be explicit (e.g.,
temperature, reflectance and emissivity of the top and
transmittance
1.1. Content relating to clouds and the atmosphere).
Rather than encompassing retrieval of these five
variables,
most
of the tests
in the CMP
algorithm
involve
This is the
Algorithm
Theoretical
Basis
Document
for
the use of brightness temperature or reflectance data with
cloud top height, temperature and pressure. It describes
similar tendencies. Cloud mask tests are based on the
scientific
considerations,
and other aspects
cloud
mask
algorithm limitations
of the NoWCasting
(NWC)of
the algorithm
used.
Satellite
Application
Facility (SAF) (Meteo-France,
2012) and the National Oceanic and Atmospheric
Administration
(NOAA)/National
Environmental
1.2. Definitions, acronyms
and abbreviations
Satellite, Data, and Information Service (NESDIS)
(Heidinger and Straka III, 2012).
Radiative transfer calculation using NWP data for
* System Engineering Division, Data Processing Department, Meteorological Satellite Center
** Meteorological Satellite and Observation System Research Department, Meteorological Research Institute
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
(Received August 31, 2015, Accepted November 20, 2015)
** System Engineering Division, Data Processing Department, Meteorological Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-- 33
1 --
－ 33 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
ABI: Advanced Baseline Imager (mounted on the
（二行） in the same way.
GOES-R satellite).
AHI:
Advanced
Algorithm
Theoretical Basis for Himawari-8
Cloud Mask Product
Himawari Imager (mounted on
2. Cloud top height product outline
（二行）
Himawari-8 for observation with 16 bands; central
IMAI
* and YOSHIDA
Ryo
**height product provides data on cloud top
wavelengths: 0.47, 0.51, 0.64, 0.86, 1.6,
2.3,Takahito
3.9, 6.2,
The cloud
top
6.9, 7.3, 8.6, 9.6, 10.4, 11.2, 12.4, 13.3 microns) .
（三行）
height, cloud top temperature and cloud top pressure as
Calipso: Cloud-Aerosol Lidar and Infrared Pathfinder
calculated using the algorithm adopted for NWCSAF
Satellite Observation.
cloud top height. The cloud type from the Cloud
Abstract
This paper
describes
the algorithm
theoretical Type/Phase
basis for the product
Himawari-8
Cloud
Product
CT: JMA’s Cloud
Type
product,
which includes
(Mouri
et Mask
al. 2016)
is important
(CMP)
developed
by
the
Meteorological
Satellite
Center.
CMP
is
part
of
the
Fundamental
Cloud
information on cloud type and cloud phase. The cloud
because it is used to determine one of three height
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
top heightoperational
product requires
in order to
retrieval methods (see Subsection 2.2).
since 7 CT
Julyinformation
2015.
switch methods
cloud
top height
calculation.
Thefor
CMP
algorithm
is based
on the cloud mask technique of NWC-SAF for MSG/SEVIRI and
for GOES-R/ABI.
Most cloud detection
testsdata
involve threshold methods based on
emissivity of wavelength.
Cloud emissivity
2.1. Input
ε(𝜆𝜆𝜆𝜆): CloudNOAA/NESDIS
radiative
transfer
calculation
using
NWP
data
as
atmospheric
profiles.
The thresholds are modified
is the ratio of radiation to black body in a temperature.
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
GOES-R: Geostationary Operational Environmental
i. AHI observation data (6.2, 7.3, 11.2 and 13.3 microns)
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
Satellite -R
Series. was applied to SEVIRI data, with resultsRadiance
are required
for the cloud
algorithm
showing data
hit ratios
with the MODIS
producttop height
GSM: Global
Spectral
Model
(JMA’s
NWP model).
product. The spatial resolution is 2 km at the sub-satellite
were
around
85% for
all seasons.
JMA: Japan Meteorological Agency.
（三行）
point, and normalized geostationary projection (Japan
MODIS:
Moderate
Resolution
Imaging
1. Introduction
Spectroradiometer (mounted on NASA’s Aqua and
（一行）
Terra satellites).
Himawari-8 is the new geostationary satellite of the
MSC: Meteorological Satellite Center.
Japan Meteorological Agency (JMA). It was launched on
Numerical
Weather
7NWP:
October
2014 and
began Prediction.
operation on 7 July 2015. The
NASA: National
Aeronautics
Space Administration.
satellite
carries the
Advancedand
Himawari
Imager (AHI),
λ(lambda):
Center
wavelength
of
AHI
bands.
which is greatly improved over past imagers in terms of
its
number
bands and
its temporal/special
resolution
LRC:
Localofradiative
center.
This is the maximum
ε(λ)
(Bessho
et al. via
2016).
data, from
JMA’s
pixel accessed
a largeUsing
adjacentAHI
ε(λ) pixel
the
Meteorological
Satellite
Center
(MSC)
developed
the
target pixel. In this cloud top height product, λ is set as
Himawari-8 Cloud Mask Product (CMP).
11.2 microns for LRC calculation. When a target pixel
Cloud mask is used to discriminate cloudy pixels
is cloudy, the LRC is a thick part in the cloud area.
from clear ones in satellite data. Numerous satellite
This concept
utilized
to judge
the cloudBy
drop
phase
products
requireis cloud
mask
information.
way
of
at
the
edge
of
a
cloud
area,
and
is
incorporated
into
the
example, Sea Surface Temperature (SST) (Yasuda and
Shirakawa
Clear Sky
Radiance
(CSR) (Uesawa
GOES-R 1999),
ABI algorithm
(Heidinger
2012).
measured
radiance,
𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 represents
𝑅𝑅𝑅𝑅{𝑚𝑚𝑚𝑚,𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜,𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
2009)
and Aerosol
(Okawara
et al. 2003)
are
𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐} (𝜆𝜆𝜆𝜆):Detection
opaque
cloud radiance
an
𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜in represents
derived
clear pixels.
Conversely,
Cloud atGrid
arbitrary
level
as
calculated
using
a
radiative
transfer
Information (Tokuno 2002) is processed in cloudy pixels.
model,
and no
𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐MSC
radiance
via
𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 represents
There were
cloud masksurface
products
before the
clear air as calculated using a radiative transfer model.
start of Himawari-8’s operation; each satellite product
RTTOV:
code
developed
by
had its ownRadiative
cloud masktransfer
process. As
cloud
mask is also
EUMETSAT
(Eyre
1991). products, MSC decided to
required
for most
Himawari-8
SST: SeaCMP
surface
develop
astemperature.
a resource available for Himawari-8
products
in common.
CMP is at
a part
of thewavelength
Fundamentalλ.
𝑇𝑇𝑇𝑇(𝜆𝜆𝜆𝜆): Brightness
temperature
the center
Cloud
Product
(FCP),
which
contains
cloud
type
The brightness temperature of other bands phase,
is expressed
and top height for Himawari-8/AHI pixels (Mouri et al.
2016a, 2016b). FCP has been produced since Himawari-8
Meteorological Agency 2013) is used. The output rate is
began
operation.
one product
per hour.
This paper describes the algorithm theoretical basis for
CMP.
ii. Radiative transfer data calculated from GSM output
（一行）
RTTOV
is
utilized
for
radiative transfer calculation. The
2. Algorithm
variables acquired from RTTOV are radiances observed
by Overview
the AHI for individual wavelengths based on the
2.1
Cloud detection
in body
the CMP
algorithm
is based
conditions
of a black
held in
equilibrium
with on
air
comparison
observation
and clear levels.
sky data
temperature of
at the
altitudes data
of mandatory
The
calculated
from
numerical
weather
prediction
(NWP)
wavelengths are 6.2, 7.3, 11.2 and 13.3 microns. The
data. If observation data differ from clear sky data, the
vertical levels are 32 mandatory ones and the surface, but
pixel is cloudy. To discriminate between cloudy and clear
radiances up to the tropopause level are used in most
pixels, it is desirable for observation variables in clear
casescloudy
for height
estimation.
and
conditions
to differ and for the physical
reason behind the difference to be explicit (e.g.,
temperature,
and emissivity of the top and
iii. Cloud typereflectance
data
transmittance
to clouds
andtop
theheight
atmosphere).
CT data are relating
utilized for
the cloud
product.
Rather
than
encompassing
retrieval
of
these
five
Only cloudy pixels and those for which the satellite
variables, most of the tests in the CMP algorithm involve
zenith angle is smaller than 84 degrees are processed in
the use of brightness temperature or reflectance data with
the product. For satellite zenith angle limitation, an angle
similar tendencies. Cloud mask tests are based on the
slightlymask
smaller
than theoflimitation
of 85 degrees
for
cloud
algorithm
the NoWCasting
(NWC)
RTTOV isApplication
set.
Satellite
Facility (SAF) (Meteo-France,
2012) and the National Oceanic and Atmospheric
Administration
(NOAA)/National
Environmental
iv. Altitude and temperature
of mandatory
levels and
Satellite,
Data,
and
Information
Service
(NESDIS)
surface
(Heidinger and Straka III, 2012).
Radiative transfer calculation using NWP data for
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 34 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 34 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
The relevant altitudes and temperatures are acquired
（二行）Under such conditions, application of the intercept
from NWP data.
method results in correction of lower-layer radiation, and
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
is suitable for single-layered semi-transparent cloud like
（二行）
2.2. Process flow of the cloud top height product
cirrus.If the intercept method does not produce suitable
IMAI Takahito* and YOSHIDA
Ryo
**
results, the
radiance
ratioing method is applied; if this
（三行）
Cloud top height estimation involves one of the
following:
also fails to produce suitable results, the interpolation
method is utilized. When the radiance ratioing method or
the interpolation method is applied after the intercept
Abstract
This method
paper describes the algorithm theoretical method,
basis for LRC
the Himawari-8
Cloud
Maskestimation.
Product This is
i. The interpolation
is utilized for
radiance
(CMP)
developed
by
the
Meteorological
Satellite
Center.
CMP
is
part
of
the
Fundamental
ii. The radiance ratioing method
also used for fractional cloud. For opaqueCloud
and fractional
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
iii. The intercept
methodsince 7 July 2015.
cloud, the interpolation method is utilized. Such cloud is
operational
and only for
radiation
from theand
cloud top is
The CMP algorithm is based on the cloud maskoptically
techniquethick,
of NWC-SAF
MSG/SEVIRI
NOAA/NESDIS
Most
tests
threshold methods based on
These methods
are describedforinGOES-R/ABI.
the next chapter
(seecloud
Fig. detection
observed
byinvolve
the AHI.
radiative
transfer
calculation
using
NWP
data
as
atmospheric
profiles.
The
thresholds
are modified
When an inversion layer
is present,
cloud is set at its
1 for the process flow), and their selection depends on the
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
upper side. This procedure is also applied for opaque
CT cloud type. For semi-transparent cloud, the intercept
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
cloud
belowhit
theratios
middle
level
discriminated
by CT.
method is used.
Such cloud
is optically
thin, anddata,
radiation
algorithm
was applied
to SEVIRI
with results
showing
with
theasMODIS
product
from the lower
is also
observed
by the AHI with this
werelayer
around
85%
for all seasons.
type.
（三行）
1. Introduction
began operation.
（一行）
This paper describes the algorithm theoretical basis for
Himawari-8 is the new geostationary satellite of the
CMP.
Japan Meteorological Agency (JMA). It was launched on
（一行）
7 October 2014 and began operation on 7 July 2015. The
2. Algorithm
satellite carries the Advanced Himawari Imager (AHI),
which is greatly improved over past imagers in terms of
2.1 Overview
Cloud detection in the CMP algorithm is based on
its number of bands and its temporal/special resolution
comparison of observation data and clear sky data
(Bessho et al. 2016). Using AHI data, JMA’s
calculated from numerical weather prediction (NWP)
Meteorological Satellite Center (MSC) developed the
data. If observation data differ from clear sky data, the
Himawari-8 Cloud Mask Product (CMP).
pixel is cloudy. To discriminate between cloudy and clear
Cloud mask is used to discriminate cloudy pixels
pixels, it is desirable for observation variables in clear
from clear ones in satellite data. Numerous satellite
and cloudy conditions to differ and for the physical
products require cloud mask information. By way of
reason behind the difference to be explicit (e.g.,
example, Sea Surface Temperature (SST) (Yasuda and
temperature, reflectance and emissivity of the top and
Shirakawa 1999), Clear Sky Radiance (CSR) (Uesawa
transmittance relating to clouds and the atmosphere).
2009) and Aerosol Detection (Okawara et al. 2003) are
Rather than encompassing retrieval of these five
derived in clear pixels. Conversely, Cloud Grid
variables, most of the tests in the CMP algorithm involve
Information (Tokuno 2002) is processed in cloudy pixels.
the use of brightness temperature or reflectance data with
There were no MSC cloud mask products before the
similar tendencies. Cloud mask tests are based on the
start of Himawari-8’s operation; each satellite product
cloud mask algorithm of the NoWCasting (NWC)
had its own cloud mask process. As cloud mask is also
Satellite Application Facility (SAF) (Meteo-France,
required for most Himawari-8 products, MSC decided to
2012) and the National Oceanic and Atmospheric
develop CMP as a resource available for Himawari-8
Administration
(NOAA)/National
Environmental
products in common. CMP is a part of the Fundamental
Satellite, Data, and Information Service (NESDIS)
Cloud Product (FCP), which contains cloud phase, type
(Heidinger and Straka III, 2012).
and top
for Himawari-8/AHI
(Mouri
et al.
Fig.height
1 Process
flow of the cloud pixels
top height
product
opaque
or semi-transparent
cloud is derived Radiative
from the cloud
type calculation
product. using NWP data for
transfer
2016a,Information
2016b). FCPonhas
been produced
since Himawari-8
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 35 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 35 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
2.3. Output data
（二行）
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
i. Cloud top height (m)
ii. Cloud top temperature (K)
iii. Cloud top pressure (hPa)
iv. Quality flags
（二行）
IMAI Takahito* and YOSHIDA Ryo**
（三行）
Chapter 4 provides information on the format of these
Abstract
describes
the algorithm
basis for the Himawari-8 Cloud Mask Product
data, which allThis
havepaper
a spatial
resolution
of 2 km theoretical
at the
(CMP)
developed
by
the
Meteorological
Satellite
Center.
CMP is part of the Fundamental Cloud
sub-satellite point. The output rate is one product per hour.
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
Fig. 2 Outline of the interpolation method
Cloudy pixels
contain valid
fine pixels contain
operational
since data
7 Julyand
2015.
invalid (-9999)The
data.
Pixels
with a issatellite
zenith
anglemask technique of NWC-SAF for MSG/SEVIRI and
CMP
algorithm
based on
the cloud
for GOES-R/ABI. Most cloud detection
tests involve
threshold
methods based on
smaller thanNOAA/NESDIS
84 degrees are contained.
3.2. Radiance
ratioing
method
radiative transfer calculation using NWP data as atmospheric profiles. The thresholds are modified
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
3. Cloud top height product algorithm
This approach is also known as the CO2 absorption
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
method
(Menzel
et al with
1983)the
and
the COproduct
2-slicing method.
algorithm was applied to SEVIRI data, with results
showing
hit ratios
MODIS
If cloud is assumed to be a single layer, radiance as
3.1. Interpolation
method
were around
85% for all seasons.
（三行）
determined by the AHI is formulated as
Radiance observed by the AHI and the vertical profile of
1. Introduction
black
body radiance as calculated using the RTTOV
radiative transfer model（一行）
are compared in this approach,
Himawari-8 is the new geostationary satellite of the
and the two mandatory levels sandwiching the observed
Japan Meteorological Agency (JMA). It was launched on
The ration
7radiance
October are
2014determined.
and began operation
on 7of
Julyinterpolation
2015. The
between carries
the radiances
of theseHimawari
levels is linearly
derived
satellite
the Advanced
Imager (AHI),
based ison
the improved
observed
Forin terms
pressure
which
greatly
overradiance.
past imagers
of
its
number of abands
and its
temporal/special
resolution
interpolation,
pressure
logarithm
is utilized.
Height,
(Bessho
et al.temperature
2016). Using
AHI data,from
JMA’s
pressure and
are determined
the
Meteorological
Satellite
Center
(MSC)
developed
the
interpolation ratio and the two mandatory levels. This
Himawari-8 Cloud Mask Product (CMP).
method is applicable for levels below the tropopause layer.
Cloud mask is used to discriminate cloudy pixels
When an inversion layer is present, cloud is set at its
from clear ones in satellite data. Numerous satellite
upper
siderequire
in the cloud
MSC algorithm.
A wavelength
of 11.2
products
mask information.
By way
of
microns
is
adopted,
as
this
represents
a
window
region
example, Sea Surface Temperature (SST) (Yasuda and
Shirakawa
1999), This
Clear approach
Sky Radiance
(CSR)to(Uesawa
for
water vapor.
is similar
window
2009) andestimation
Aerosol Detection
et al.
are
channel
(Neiman (Okawara
et al 1993)
and2003)
radiance
derived
in
clear
pixels.
Conversely,
Cloud
Grid
fitting (Hamman et al 2014). Figure 2 shows an example
Information (Tokuno 2002) is processed in cloudy pixels.
of pressure determination using this method. The vertical
There were no MSC cloud mask products before the
profile of radiance at 11.2 microns is calculated using
start of Himawari-8’s operation; each satellite product
radiative
transfer
model.
The As
interpolation
of
had its own
cloud mask
process.
cloud maskratio
is also
radiance
between
the two products,
levels sandwiching
required for
most Himawari-8
MSC decided the
to
develop CMP
as a resource
available
for Himawari-8
observed
radiance
reflects the
interpolated
pressure
products the
in common.
CMP
a part ofshows
the Fundamental
between
two levels.
Thisisexample
an observed
Cloud
Product
(FCP),
which
contains
cloud
phase,
type
radiance of 85.2 mW converted to 591.4 hPa.
and top height for Himawari-8/AHI pixels (Mouri et al.
2016a, 2016b). FCP has been produced since Himawari-8
began
= 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 (𝜆𝜆𝜆𝜆) + (1 − 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁)𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆) (1)
𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 (𝜆𝜆𝜆𝜆)operation.
This paper describes the algorithm theoretical basis for
CMP.
where N is the cloud amount. If 𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 is known,
（一行）
height
can
be
estimated
using the approach of the
2. Algorithm
interpolation method. Equation (1) is transformed to
Equation
(2).
2.1
Overview
Cloud detection in the CMP algorithm is based on
comparison
and
clear sky
(𝜆𝜆𝜆𝜆) observation
(𝜆𝜆𝜆𝜆) − 𝑅𝑅𝑅𝑅
𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 (𝜆𝜆𝜆𝜆) − 𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐of
= 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 �𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜data
(2) data
𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆)�
calculated from numerical weather prediction (NWP)
data. If observation data differ from clear sky data, the
Cloud top pressure is estimated from the ratio of radiance
pixel is cloudy. To discriminate between cloudy and clear
between two wavelengths (Chahine 1974, Smith and Platt
pixels, it is desirable for observation variables in clear
1978).
and
cloudy conditions to differ and for the physical
reason behind the difference to be explicit (e.g.,
(𝜆𝜆𝜆𝜆1 )−𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆
𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁1 �𝑅𝑅𝑅𝑅and
(𝜆𝜆𝜆𝜆1 )
𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 (𝜆𝜆𝜆𝜆1 )−𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐reflectance
temperature,
emissivity
of1)�the top and
=
(3)
(𝜆𝜆𝜆𝜆
)−𝑅𝑅𝑅𝑅
(𝜆𝜆𝜆𝜆
)
𝑅𝑅𝑅𝑅
(𝜆𝜆𝜆𝜆
)−𝑅𝑅𝑅𝑅
(𝜆𝜆𝜆𝜆
𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁
�𝑅𝑅𝑅𝑅
𝑚𝑚𝑚𝑚
2
𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
𝑐𝑐𝑐𝑐
𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
2
𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
transmittance relating 2 to𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
clouds2 and
the2)�atmosphere).
Rather than encompassing retrieval of these five
variables,
mostofof the
the tests
in the CMP
involve
The value
left-hand
side algorithm
is derived
once
the
use
of
brightness
temperature
or
reflectance
data
𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 (𝜆𝜆𝜆𝜆1 ) , 𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 (𝜆𝜆𝜆𝜆2 ) , 𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆1 ) and 𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆2 ) with
are
similar tendencies. Cloud mask tests are based on the
determined. The value of 𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 (𝜆𝜆𝜆𝜆1 ) and 𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 (𝜆𝜆𝜆𝜆2 )
cloud mask algorithm of the NoWCasting (NWC)
on the right-hand side depend on the vertical profile. As
Satellite Application Facility (SAF) (Meteo-France,
almost equal:
𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁
1 and
2012)
and𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁2theareNational
Oceanic and Atmospheric
Administration
(NOAA)/National
Environmental
𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁1
≈ 1 Data, and Information Service (NESDIS)
(4)
Satellite,
𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁2
(Heidinger and Straka III, 2012).
Radiative transfer calculation using NWP data for
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 36 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 36 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
(251.6 hPa for the 11.2(B14)/6.2(B08)-micron pair, 257.5
（二行）
Equation (3) can be written as
hPa for the 11.2(B14)/7.3(B10)-micron pair, and 261.8
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 (𝜆𝜆𝜆𝜆1 )−𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆1 )
𝑅𝑅𝑅𝑅𝑚𝑚𝑚𝑚 (𝜆𝜆𝜆𝜆2 )−𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆2 )
=
𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 (𝜆𝜆𝜆𝜆1 )−𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆1 )
𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 (𝜆𝜆𝜆𝜆2 )−𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 (𝜆𝜆𝜆𝜆2 )
hPa for the 11.2(B14)/13.3(B16)-micron pair. The final
（二行）
retrieved cloud top height is the minimum pressure (at the
(5)
IMAI Takahito* and YOSHIDA
Ryo** determined from one of the three band
highest altitude)
（三行）
Values of 𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 (𝜆𝜆𝜆𝜆1 ) and 𝑅𝑅𝑅𝑅𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 (𝜆𝜆𝜆𝜆2 ) satisfying
pairs. This follows the intercept approach of NWCSAF.
Equation (5) are then derived from the vertical profile of
3.3. Intercept method
Abstract
𝑅𝑅𝑅𝑅𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 is essentially surface radiance
This paper describes the algorithm theoretical basis for the Himawari-8 Cloud Mask Product
calculated using
radiativeby transfer
model, butSatellite
its
(CMP)the
developed
the Meteorological
Center.
CMP is method
part of the
Fundamental
Cloudis adopted
The Intercept
(Schmetz
et al. 1993)
accuracy may
be limited
in incorporates
association with
of water
Product,
which
cloudthat
phase,
type and top altitude from Himawari-8/AHI and has been
for semi-transparent clouds. In the NWCSAF procedure,
operational
sinceAccordingly,
7 July 2015. measured clear
vapor estimation
in NWP.
a linear expression is first set up via the least square
CMPifalgorithm
is based
on theincloud
sky radiance is The
adopted
clear pixels
are present
the 3mask technique of NWC-SAF for MSG/SEVIRI and
NOAA/NESDIS for GOES-R/ABI. Most cloud detection
testsininvolve
threshold
basedradiances
on
approach
a scatter
diagrammethods
of observed
for
x 3 pixel array centered on the target pixel. MSC uses
radiative transfer calculation using NWP data as atmospheric
profiles.
The
thresholds
are
modified
two bands. The data contained in the diagram are
three bandusing
pairs
(11.2(B14)/6.2(B08)-micron
pair,
offsets
determined on the basis of comparison
with the MODIS cloud mask product. Initial
radiances of 16 or so pixels around the target pixel, and a
11.2(B14)/7.3(B10)-micron
pair
andfrom such comparison was more than 85%. The CMP
results showed that the CMP
hit ratio derived
33 showing
x 33-pixel
is formed.
There product
are no limitations
algorithm was applied
with results
hittemplate
ratios with
the MODIS
11.2(B14)/13.3(B16)-micron
pair)to SEVIRI
in thedata,related
on surface type when the template is made. Next, the
were around 85% for all seasons.
algorithm. Figure 3 shows an example of pressure
（三行）
intersection of the black body radiance curve for each
retrieval using this method. The horizontal axis represents
level is derived, along with the relevant linear expression.
pressure (hPa), and the vertical axis shows the value of
began
operation.channel’s radiance at the intersection
1. Introduction
The window
the left-hand side (measured ratio) minus the right hand
（一行）
This
paper
algorithm theoretical
basis
indicates
the describes
height ofthesemi-transparent
cloud
in for
the
sideHimawari-8
(simulated ratio)
of
Equation
(5).
is the new geostationary satellite of the
CMP.
troposphere. The cloud top height is then retrieved as in
Japan Meteorological Agency (JMA). It was launched on
（一行）
the
interpolation
method.
There are three band pairs in
7 October 2014 and began operation on 7 July 2015. The
2. Algorithm
this algorithm (11.2/6.2 microns, 11.2/7.3 microns and
satellite carries the Advanced Himawari Imager (AHI),
11.2/13.3
microns). The minimum pressure (at the highest
which is greatly improved over past imagers in terms of
2.1
Overview
Cloudisdetection
in theamong
CMP the
algorithm
is based
on4
its number of bands and its temporal/special resolution
altitude)
selected from
three pairs.
Figure
comparison
of observation
data and
clear sky
(Bessho et al. 2016). Using AHI data, JMA’s
shows an example
of how pressure
is retrieved
usingdata
this
calculated
from
numerical
weather
prediction
(NWP)
Meteorological Satellite Center (MSC) developed the
method. The retrieved radiance of 11.2 microns for the
data. If observation data differ from clear sky data, the
Himawari-8 Cloud Mask Product (CMP).
11.2/6.2-micron pair, for example, is 17 mW (Fig. 4 (a)),
pixel is cloudy. To discriminate between cloudy and clear
Cloud mask is used to discriminate cloudy pixels
that for ittheis 11.2/7.3-micron
pair is 23.9variables
mW (Fig.
(b)),
pixels,
desirable for observation
in 4clear
from clear ones in satellite data. Numerous satellite
and that
for the
11.2/13.3-micron
16.0themW
(Fig. 4
and
cloudy
conditions
to differ pair
andisfor
physical
products require cloud mask information. By way of
(c)). Figure
4 (d)
that to
cloud
pressure
reason
behind
the shows
difference
be top
explicit
(e.g.,is
example, Sea Surface Temperature (SST) (Yasuda and
temperature,
reflectance
and emissivity
of 11.2
the top
and
Shirakawa
1999),
Clear
Sky Radiance
(CSR) (Uesawa
determined from
the retrieved
radiance of
microns.
Fig. 3 Outline
of the
radiance
ratioing method
transmittance
relatingforto11.2
clouds
and isthe
atmosphere).
2009)
and Aerosol
Detection (Okawara
al. 2003) are
B14:B16
is 11.2/13.3-micron
pair,et B14:B10
is
The radiance profile
microns
calculated
using
11.2/7.3-micron
and B14:B08
is 11.2/6.2-micron
Rather
than
encompassing
retrieval
of
these
five
derived
in clearpair,
pixels.
Conversely,
Cloud Grid
RTTOV. The retrieved radiances of the three pairs
pair.
variables, most of the tests in the CMP algorithm involve
Information (Tokuno 2002) is processed in cloudy pixels.
indicate individual pressure levels in the vertical profile.
the use of brightness temperature or reflectance data with
There were no MSC cloud mask products before the
In this case, 17 mW marks 95.16 hPa for the
similar tendencies. Cloud mask tests are based on the
start
operation;
eachindicates
satellite product
The of Himawari-8’s
dashed
line
the
11.2/6.2-micron
pair. ofIn thetheNoWCasting
same way,
the
cloud
mask algorithm
(NWC)
had its own cloud mask process.
As cloud
11.2(B14)/13.3(B16)-micron
pair,
the mask
chainis also
line
11.2/7.3-micron
pair Facility
indicates (SAF)
139.5 (Meteo-France,
hPa and the
Satellite
Application
required forthe
most
Himawari-8 products, MSC
decided
to
represents
11.2(B14)/7.3(B10)-micron
pair,
and the
2012)
and the pair
National
Oceanic
develop CMP as a resource available for Himawari-8
11.2/13.3-micron
indicates
95.11 and
hPa. Atmospheric
The value of
solid line represents the 11.2(B14)/6.2(B08)-micron pair.
Administration
products in common. CMP is a part of the Fundamental
95.11 hPa from the(NOAA)/National
11.2/13.3-micron pairEnvironmental
is adopted as
The arrows show the point of intersection with the zero
Satellite,
Data,
and
Information
Service
Cloud Product (FCP), which contains cloud phase, type
the cloud top pressure, as it is the lowest level. (NESDIS)
line
of
the
measured-simulated
ratio
and
each
pair
line
(Heidinger and Straka III, 2012).
and top height for Himawari-8/AHI pixels (Mouri et al.
Radiative transfer calculation using NWP data for
2016a, 2016b). FCP has been produced since Himawari-8
radiance.
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 37 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 37 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
（二行）
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
（二行）
IMAI Takahito* and YOSHIDA Ryo**
（三行）
Abstract
This paper describes the algorithm theoretical basis for the Himawari-8 Cloud Mask Product
(CMP) developed by the Meteorological Satellite Center. CMP is part of the Fundamental Cloud
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
operational since 7 July 2015.
The CMP algorithm is based on the cloud mask technique of NWC-SAF for MSG/SEVIRI and
NOAA/NESDIS for GOES-R/ABI. Most cloud detection tests involve threshold methods based on
radiative transfer calculation using NWP data as atmospheric profiles. The thresholds are modified
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
algorithm was applied to SEVIRI data, with results showing hit ratios with the MODIS product
were around 85% for all seasons.
（三行）
1. Introduction
began operation.
（一行）
This paper describes the algorithm theoretical basis for
Himawari-8 is the new geostationary satellite of the
CMP.
Japan Meteorological Agency (JMA). It was launched on
（一行）
7 October 2014 and began operation on 7 July 2015. The
2. Algorithm
satellite carries the Advanced Himawari Imager (AHI),
which is greatly improved over past imagers in terms of
2.1 Overview
Cloud detection in the CMP algorithm is based on
its number of bands and its temporal/special resolution
comparison of observation data and clear sky data
(Bessho et al. 2016). Using AHI data, JMA’s
calculated from numerical weather prediction (NWP)
Meteorological Satellite Center (MSC) developed the
data. If observation data differ from clear sky data, the
Himawari-8 Cloud Mask Product (CMP).
pixel is cloudy. To discriminate between cloudy and clear
Cloud mask is used to discriminate cloudy pixels
pixels, it is desirable for observation variables in clear
from clear ones in satellite data. Numerous satellite
and cloudy conditions to differ and for the physical
products require cloud mask information. By way of
reason behind the difference to be explicit (e.g.,
example, Sea Surface Temperature (SST) (Yasuda and
Fig. 4 1999),
OutlineClear
of theSky
intercept
method
temperature, reflectance and emissivity of the top and
Shirakawa
Radiance
(CSR) (Uesawa
The
horizontal
axis
shows
the
observed
radiance
for
11.2
microns, and relating
the vertical
axis represents
observed
to clouds
and thethe
atmosphere).
2009) and Aerosol Detection (Okawara et al. -22003)
are-1 -1 transmittance
-1
sr
(cm
)
.
radiance
of
band
pairs.
The
unit
is
mW
m
Rather than encompassing retrieval of these five
derived in clear pixels. Conversely, Cloud Grid
(a) Scatter diagram of 11.2 and 6.2 micron observed radiances.
variables,
in theisCMP
Information
2002)
is red
processed
cloudy pixels.
The(Tokuno
curve (the
solid
line) ofinradiative
transfer calculation
formost
11.2 of
andthe6.2tests
microns
also algorithm
shown. involve
the
use
of
brightness
temperature
or
reflectance
data with
There
were
no
MSC
cloud
mask
products
before
the
(b) Scatter diagram and curve for 11.2 and 7.3 microns (as per Fig. 4 (a))
start of
each
satellite
product
(c) Himawari-8’s
Scatter diagramoperation;
for 11.2 and
13.3
microns
(as per Fig. similar
4 (a)) tendencies. Cloud mask tests are based on the
(d)own
Simulated
radiance
profileAsforcloud
11.2 mask
microns
as calculated
using
RTTOV
cloud
mask
algorithm of the NoWCasting (NWC)
had its
cloud mask
process.
is also
Satellite Application Facility (SAF) (Meteo-France,
required for most Himawari-8 products, MSC decided to
2012) and the National Oceanic and Atmospheric
develop CMP as a resource available for Himawari-8
Administration
(NOAA)/National
Environmental
products in common. CMP is a part of the Fundamental
Satellite, Data, and Information Service (NESDIS)
Cloud Product (FCP), which contains cloud phase, type
(Heidinger and Straka III, 2012).
and top height for Himawari-8/AHI pixels (Mouri et al.
Radiative transfer calculation using NWP data for
2016a, 2016b). FCP has been produced since Himawari-8
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 38 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 38 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
4. Input/output data
4.1. Input data
i. AHI data
（二行）
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
（二行）
IMAI Takahito* and YOSHIDA Ryo**
（三行）
Brightness temperatures and radiances at 6.2, 7.3, 11.2
and 13.3 microns are observed by the AHI.
Abstract
Fig. 5 File format and pixel order
Thiscalculated
paper describes
the RTTOV
algorithm
theoretical basis for the Himawari-8 Cloud Mask Product
ii. Vertical profiles
using the
radiative
Data are written in flat binary format.
transfer code(CMP) developed by the Meteorological Satellite Center. CMP is part of the Fundamental Cloud
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
The vertical
profiles since
of brightness
temperatures and
5. Data provision for general users
operational
7 July 2015.
radiances are calculated
RTTOV.
The on
wavelengths
The CMP using
algorithm
is based
the cloud mask technique of NWC-SAF for MSG/SEVIRI and
GOES-R/ABI. Most cloud detection
tests involve
threshold
methods
based
on
are the sameNOAA/NESDIS
as those of AHI for
data.
The cloud
top height
product
is used
exclusively
by
radiative transfer calculation using NWP data as atmospheric
profiles.
The
thresholds
are
modified
MSC and is not provided to external users.
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
iii. Cloud type product results
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
Cloud typealgorithm
information
required.to SEVIRI data, with results
6. Cloud
tophitheight
wasis applied
showing
ratiosproduct
with theevaluation
MODIS product
were around 85% for all seasons.
4.2. Output data
（三行）Evaluation was carried out over a period of two weeks
from 20 July to 2 August 2015. The cloud top height
began
productoperation.
was compared to its MODIS counterpart (Baum
This
paper
the algorithm
theoretical
basis
et al. 2012) describes
and Calipso
(Winker et
al. 2006).
In for
the
CMP.
match-up process, pixels within 2 km and within 5
（一行）
minutes
of
the
scanning
time between cloud top height
2. Algorithm
product data and MODIS or Calipso data were picked up.
1. Cloud
Introduction
i.
top pressure, cloud top altitude and cloud top
（一行）
temperature
Himawari-8 is the new geostationary satellite of the
These data are output in flat binary form at 1 byte per
Japan Meteorological Agency (JMA). It was launched on
pixel.
Figure
5 shows
the pixel
order,onand
Table
1 shows
7 October
2014
and began
operation
7 July
2015.
The
the
specifications
cloud top
pressure,
altitude
and
satellite
carries the for
Advanced
Himawari
Imager
(AHI),
temperature.
Theimproved
start countover
is minus
127 in allincases,
Table
3 shows the results of evaluation for the cloud top
which is greatly
past imagers
terms and
of
2.1
Overview
Cloud
detection
in theand
CMP
algorithm
based
on
its number
bands isand
its temporal/special
resolution
the
physicalof
quantity
determined
using
height
product
of MODIS
Calipso
duringisthis
period.
comparison
of observation
datarepresentation
and clear sky
data
(Bessho et al. 2016). Using AHI data, JMA’s
Figure 6 shows
a time-series
of mean
calculated
from
numerical
weather
prediction
(NWP)
Meteorological
Satellite
Center
(MSC)
developed
the
Physical quantity = slope × (𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 + 127)
errors and root mean square errors for cloud top height. In
data. If observation data differ from clear sky data, the
Himawari-8 Cloud Mask Product (CMP).
the cloud top height product, cloud top height was
pixel is cloudy. To discriminate between cloudy and clear
Cloud mask is used to discriminate cloudy pixels
ii. Quality flag data
underestimated
in comparison
to corresponding
pixels,
it is desirable
for observation
variables in values
clear
from clear ones in satellite data. Numerous satellite
These data
are also
output
in information.
flat binary form
1 byte
from cloudy
MODISconditions
and Calipsotothroughout
and
differ andtheforperiod.
the physical
products
require
cloud
mask
By atway
of
per
pixel. Table
2 detailsTemperature
quality information
per pixel.and
reason behind the difference to be explicit (e.g.,
example,
Sea Surface
(SST) (Yasuda
temperature, reflectance and emissivity of the top and
Shirakawa 1999), Clear Sky Radiance (CSR) (Uesawa
transmittance
to clouds and the atmosphere).
2009) and Aerosol
(Okawara et for
al. cloud
2003)top
arepressure,
Table 1Detection
Output specifications
altitude andrelating
temperature
Rather than encompassing retrieval of these five
derived in clear pixels. Conversely, Cloud Grid
variables, most of the tests in the CMP algorithm involve
Information (Tokuno 2002) is processed in cloudy pixels.
the use of brightness temperature or reflectance data with
There were no MSC cloud mask products before the
similar tendencies. Cloud mask tests are based on the
start of Himawari-8’s operation; each satellite product
cloud mask algorithm of the NoWCasting (NWC)
had its own cloud mask process. As cloud mask is also
Satellite Application Facility (SAF) (Meteo-France,
required for most Himawari-8 products, MSC decided to
2012) and the National Oceanic and Atmospheric
develop CMP as a resource available for Himawari-8
Administration
(NOAA)/National
Environmental
products in common. CMP is a part of the Fundamental
Satellite, Data, and Information Service (NESDIS)
Cloud Product (FCP), which contains cloud phase, type
(Heidinger and Straka III, 2012).
and top height for Himawari-8/AHI pixels (Mouri et al.
Radiative transfer calculation using NWP data for
2016a, 2016b). FCP has been produced since Himawari-8
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 39 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 39 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
Table 2 Pixel quality information
（二行）
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
（二行）
IMAI Takahito* and YOSHIDA Ryo**
（三行）
Abstract
This paper describes the algorithm theoretical basis for the Himawari-8 Cloud Mask Product
(CMP) developed by the Meteorological Satellite Center. CMP is part of the Fundamental Cloud
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
operational since 7 July 2015.
The CMP algorithm is based on the cloud mask technique of NWC-SAF for MSG/SEVIRI and
NOAA/NESDIS for GOES-R/ABI. Most cloud detection tests involve threshold methods based on
radiative transfer calculation using NWP data as atmospheric profiles. The thresholds are modified
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
algorithm was applied to SEVIRI data, with results showing hit ratios with the MODIS product
were around 85% for all seasons.
（三行）
1. Introduction
began operation.
（一行）
This paper describes the algorithm theoretical basis for
Himawari-8 is the new geostationary satellite of the
CMP.
Japan Meteorological Agency (JMA). It was launched on
（一行）
7 October 2014 and began operation on 7 July 2015. The
2. Algorithm
satellite Table
carries3 Mean
the Advanced
Himawari
Imager
(AHI),
errors and root mean square errors for the period from 20 July to 2 August
which is greatly improved over past imagers in terms of
2.1 Overview
Cloud detection in the CMP algorithm is based on
its number of bands and its temporal/special resolution
comparison of observation data and clear sky data
(Bessho et al. 2016). Using AHI data, JMA’s
calculated from numerical weather prediction (NWP)
Meteorological Satellite Center (MSC) developed the
data. If observation data differ from clear sky data, the
Himawari-8 Cloud Mask Product (CMP).
pixel is cloudy. To discriminate between cloudy and clear
Cloud mask is used to discriminate cloudy pixels
(a)
pixels, it (b)
is desirable for observation variables in clear
from clear ones in satellite data. Numerous satellite
and cloudy conditions to differ and for the physical
products require cloud mask information. By way of
reason behind the difference to be explicit (e.g.,
example, Sea Surface Temperature (SST) (Yasuda and
temperature, reflectance and emissivity of the top and
Shirakawa 1999), Clear Sky Radiance (CSR) (Uesawa
transmittance relating to clouds and the atmosphere).
2009) and Aerosol Detection (Okawara et al. 2003) are
Rather than encompassing retrieval of these five
derived in clear pixels. Conversely, Cloud Grid
variables, most of the tests in the CMP algorithm involve
Information (Tokuno 2002) is processed in cloudy pixels.
the use of brightness temperature or reflectance data with
There were no MSC cloud mask products before the
similar tendencies. Cloud mask tests are based on the
start of Himawari-8’s operation; each satellite product
cloud mask algorithm of the NoWCasting (NWC)
had its own cloud mask process. As cloud mask is also
Satellite Application Facility (SAF) (Meteo-France,
required for most Himawari-8 products, MSC decided to
2012) and the National Oceanic and Atmospheric
develop CMP as a resource available for Himawari-8
Fig. 6 Time-series representation of mean errors and root mean square errors
(NOAA)/National
productsThe
in common.
a part ofthe
thedate,
Fundamental
horizontalCMP
axis isrepresents
the left verticalAdministration
axis represents mean
error, and the rightEnvironmental
vertical axis
Satellite,
Data,
and
Information
Service (NESDIS)
Cloud Product
(FCP),
which
contains
cloud
phase,
type
represents root mean square error.
(Heidinger
Straka
2012).
and top The
height
Himawari-8/AHI
pixels
(Mourifor
et MODIS
al.
fig for
(a) and
the fig (b) show
evaluation
and
Calipsoand
cloud
top III,
height,
respectively.
Radiative transfer calculation using NWP data for
2016a, 2016b). FCP has been produced since Himawari-8
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 40 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 40 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
7. Assumptions and limitations
（二行） 2014: Remote sensing of cloud top pressure/height
from SEVIRI: analysis of ten current retrieval
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
i. The semi-transparent cloud top height algorithms are
（二行）
algorithms. Atmos.Meas.Tech,2842-2943.
assumed for single-layer cloud rather than multi-layered
Heidinger A. K.,2012: ALGORITHM THEORETICAL
IMAI top
Takahito
* and YOSHIDA
**
cloud. The accuracy of semi-transparent cloud
height
BASIS Ryo
DOCUMENT
ABI Cloud Height.
estimation is lower for multi-layered cloud.
（三行）
Japan Meteorological Agency, 2013: Himawari-8/9
ii. The accuracy of cloud top height estimation is lower
Himawari Standard Data User's Guide, 5.
if the cloud type is not properly classified. Cloud top
Meteo France, 2012: Algorithm Theoretical Basis
Abstract
This paper describes
the algorithm
theoretical
basis
for the Himawari-8
Mask Product
height for semi-transparent
cloud classified
as opaque
in
Document
for “Cloud Cloud
Products”(CMa-PGE01
v3.2,
(CMP)
developed
by
the
Meteorological
Satellite
Center.
CMP
is
part
of
the
Fundamental
Cloud
the cloud type product is underestimated, while that for
CT-PGE02 v2.2 & CTTH-PGE03 v2.2).
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
opaque cloud
correctly
Menzel W. P., W. L. Smith, and T. R. Stewart, 1983:
operational
since 7classified
July 2015.as opaque is
overestimated. The CMP algorithm is based on the cloud mask technique
Improvedof Cloud
Motion
Wind Vector and
and Altitude
NWC-SAF
for MSG/SEVIRI
NOAA/NESDIS
GOES-R/ABI.
cloud detection
tests involve
threshold
on
iii. Accuracy
for fractional for
cloud
is lower if Most
the cloud
Assignment
using
VAS. methods
Journal based
of Climate
and
radiative
transfer
calculation
using
NWP
data
as
atmospheric
profiles.
The
thresholds
are
modified
does not fully cover the target pixel.
Applied meteorology, 22, 377-384.
using offsets determined on the basis of comparison with the MODIS cloud mask product. Initial
Mouri K., T. Izumi, H. Suzue, and R. Yoshida, 2016:
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
Acknowledgements:
Basis
Document
algorithm was applied to SEVIRI data, with results Algorithm
showing hit Theoretical
ratios with the
MODIS
product of cloud
type/phase product. Meteorological Satellite Center
were around 85% for all seasons.
（三行） Technical Note, 61, 19-31.
The authors gratefully acknowledge the two anonymous
reviewers who provided helpful comments on this
1. Introduction
manuscript,
and also thank the NASA Langley Research
（一行）
Center Atmospheric Science
Data Center for providing
Himawari-8 is the new geostationary satellite of the
the MODIS and Calipso data used in the study.
Japan Meteorological Agency (JMA). It was launched on
7 October 2014 and began operation on 7 July 2015. The
References:
satellite
carries the Advanced Himawari Imager (AHI),
which is greatly improved over past imagers in terms of
its
number
of W.
bands
and its R.
temporal/special
resolution
Baum,
B. A.,
P. Menzel,
A. Frey, D. Tobin,
R. E.
(Bessho
2016). A.Using
AHI data,
Holz, S.etA. al.
Ackerman,
K. Heidinger,
and P.JMA’s
Yang,
Meteorological
Satellite
Center
(MSC)
developed
2012: MODIS cloud top property refinements the
for
Himawari-8 Cloud Mask Product (CMP).
Collection 6. J. Appl. Meteor. Clim., 51, 1145-1163.
Cloud mask is used to discriminate cloudy pixels
Chahine, M. T. 1974: Remote sounding of cloudy
from clear ones in satellite data. Numerous satellite
atmospheres,
The mask
singleinformation.
cloud layer.ByJ. way
Atmos.
products
require I.cloud
of
Sci.,31,233-243.
example, Sea Surface Temperature (SST) (Yasuda and
Shirakawa
Skyradiative
Radiance
(CSR) model
(Uesawa
Eyre
J. R.,1999),
1991: Clear
A fast
transfer
for
2009)
and Aerosol
al. 2003)Dept.
are
satellite
soundingDetection
systems. (Okawara
ECMWF etResearch
derived
in
clear
pixels.
Conversely,
Cloud
Grid
Tech. Memo. 176.
Information (Tokuno 2002) is processed in cloudy pixels.
Hamman.U., A. Walther, B. Baum, R. Bennartz, L.
There were no MSC cloud mask products before the
Bugliaro, M. Derrien, P. N. Francis, A. Heidinger, S.
start of Himawari-8’s operation; each satellite product
A. Kniffka,
H. Le
Gleau,AsM.
Lockhoff,
–J.
hadJoro,
its own
cloud mask
process.
cloud
mask isH.also
Luts, for
J. most
F. Meirink,
P. Minnis,
Plikonda,
R.
required
Himawari-8
products, R.
MSC
decided to
develop
CMPA.asThoss,
a resource
available
for and
Himawari-8
Roebeling,
S. Platnik,
P. Watts,
G. Wind,
products in common. CMP is a part of the Fundamental
Cloud Product (FCP), which contains cloud phase, type
and top height for Himawari-8/AHI pixels (Mouri et al.
2016a, 2016b). FCP has been produced since Himawari-8
Nieman S. J., J. Schmetz, and W. P. Menzel, 1993: A
began
operation.of Several Techniques to Assign Heights
Comparison
This
paperTracers.
describes
the algorithm
theoretical
basis for
to Cloud
Journal
of Applied
Meteorology,
32,
CMP.
1560.
（一行）
Schmetz
J.,
K.
Holmlund,
J. Hoffman, and B.Strauss,
2. Algorithm
1993: Operational cloud motion winds from Meteosat
images. Jounal of Applied Meteorology, 32,
2.1infrared
Overview
Cloud detection in the CMP algorithm is based on
1207-1225.
comparison
observation
data 1978:
and clear
sky data
Smith, W. L.,ofand
C. M. R. Platt,
Intercomparison
calculated
from
numerical
weather
prediction
of radiosonde, ground based laser, and (NWP)
satellite
data. If observation data differ from clear sky data, the
deduced cloud heights.
J. Appl. Meteor., 17,
pixel is cloudy. To discriminate between cloudy and clear
1796-1802.
pixels,
it is desirable for observation variables in clear
Suzuecloudy
H., T. conditions
Imai, and K.
2016:forHigh-resolution
and
to Mouri,
differ and
the physical
Cloud
Analysis
Information
derived
from
Himawari-8
reason behind the difference to be explicit
(e.g.,
temperature,
reflectance Satellite
and emissivity
the top Note,
and
data. Meteorological
Center of
Technical
transmittance
61, 43-51. relating to clouds and the atmosphere).
Rather
than
of these
five
Winker, D.
M., encompassing
C. A. Hostetler, retrieval
M. A. Vaughan,
and A.
H.
variables, most of the tests in the CMP algorithm involve
Omar, 2006: Caliop Algtorithm Theoretical Basis
the use of brightness temperature or reflectance data with
Document Part 1: CALIOP Instrument, and Algorithms
similar tendencies. Cloud mask tests are based on the
Overveiw.
cloud
mask 20-23.
algorithm of the NoWCasting (NWC)
Satellite Application Facility (SAF) (Meteo-France,
2012) and the National Oceanic and Atmospheric
Administration
(NOAA)/National
Environmental
Satellite, Data, and Information Service (NESDIS)
(Heidinger and Straka III, 2012).
Radiative transfer calculation using NWP data for
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
- 41 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 41 －
METEOROLOGICAL SATELLITE CENTER TECHNICAL NOTE, No.61, MARCH, 2016
（二行）
雲頂高度アルゴリズム記述書
Algorithm Theoretical Basis for Himawari-8 Cloud Mask Product
（二行）
毛利
浩樹*、鈴江 寛史*、吉田 良*、泉
IMAI Takahito* and YOSHIDA Ryo**
敏治**
（三行）
要旨
雲頂高度プロダクトは、基本雲プロダクトの一要素である。基本雲プロダクトは雲マスク、
Abstract
雲タイプ、雲頂高度プロダクトからなる。気象庁気象衛星センターでは、ひまわり８号を機に、
This paper describes the algorithm theoretical basis for the Himawari-8 Cloud Mask Product
様々な２次プロダクトが取り扱う雲マスク、雲タイプ・相、雲頂高度を独立させ、基本雲プロダ
(CMP) developed by the Meteorological Satellite Center. CMP is part of the Fundamental Cloud
クトとしてまとめ上げた。高分解能雲情報が現在基本雲プロダクトを使用している。
Product, which incorporates cloud phase, type and top altitude from Himawari-8/AHI and has been
雲頂高度プロダクトは、雲頂高度、雲頂温度、雲頂気圧の３つの要素から成り、アルゴリズムは
operational since 7 July 2015.
欧州気象衛星機関の
NWCSAF
という枠組みで開発されたものを使用している。入力データはひ
The CMP algorithm
is based
on the cloud mask technique of NWC-SAF for MSG/SEVIRI and
まわり
8
号に搭載されている
AHI（Advanced
Imager）によって観測されたデータ、放
NOAA/NESDIS for GOES-R/ABI. Most cloudHimawari
detection tests
involve threshold methods based on
射伝達計算済みの鉛直プロファイルデータ、
radiative transfer calculation using NWP dataおよび雲タイププロダクトによって出力された雲タ
as atmospheric profiles. The thresholds are modified
イプデータである。
2015 年on
7月
日から
8 月 2 日の with
2 週間における雲頂高度の評価では、
MODIS
using offsets determined
the20basis
of comparison
the MODIS cloud mask product. Initial
雲頂高度および
Calipso
雲頂高度に対して、本プロダクトの雲頂高度はやや低い傾向であった。
results showed that the CMP hit ratio derived from such comparison was more than 85%. The CMP
algorithm was applied to SEVIRI data, with results showing hit ratios with the MODIS product
were around 85% for all seasons.
（三行）
1. Introduction
（一行）
Himawari-8 is the new geostationary satellite of the
Japan Meteorological Agency (JMA). It was launched on
7 October 2014 and began operation on 7 July 2015. The
satellite carries the Advanced Himawari Imager (AHI),
which is greatly improved over past imagers in terms of
its number of bands and its temporal/special resolution
(Bessho et al. 2016). Using AHI data, JMA’s
Meteorological Satellite Center (MSC) developed the
Himawari-8 Cloud Mask Product (CMP).
Cloud mask is used to discriminate cloudy pixels
from clear ones in satellite data. Numerous satellite
products require cloud mask information. By way of
example, Sea Surface Temperature (SST) (Yasuda and
Shirakawa 1999), Clear Sky Radiance (CSR) (Uesawa
2009) and Aerosol Detection (Okawara et al. 2003) are
derived in clear pixels. Conversely, Cloud Grid
Information (Tokuno 2002) is processed in cloudy pixels.
There were no MSC cloud mask products before the
start of Himawari-8’s operation; each satellite product
had its own cloud mask process. As cloud mask is also
required for most Himawari-8 products, MSC decided to
develop CMP as a resource available for Himawari-8
products in common. CMP is a part of the Fundamental
Cloud Product (FCP), which contains cloud phase, type
and top height for Himawari-8/AHI pixels (Mouri et al.
2016a, 2016b). FCP has been produced since Himawari-8
began operation.
This paper describes the algorithm theoretical basis for
CMP.
（一行）
2. Algorithm
2.1 Overview
Cloud detection in the CMP algorithm is based on
comparison of observation data and clear sky data
calculated from numerical weather prediction (NWP)
data. If observation data differ from clear sky data, the
pixel is cloudy. To discriminate between cloudy and clear
pixels, it is desirable for observation variables in clear
and cloudy conditions to differ and for the physical
reason behind the difference to be explicit (e.g.,
temperature, reflectance and emissivity of the top and
transmittance relating to clouds and the atmosphere).
Rather than encompassing retrieval of these five
variables, most of the tests in the CMP algorithm involve
the use of brightness temperature or reflectance data with
similar tendencies. Cloud mask tests are based on the
cloud mask algorithm of the NoWCasting (NWC)
Satellite Application Facility (SAF) (Meteo-France,
2012) and the National Oceanic and Atmospheric
Administration
(NOAA)/National
Environmental
Satellite, Data, and Information Service (NESDIS)
(Heidinger and Straka III, 2012).
Radiative transfer calculation using NWP data for
* 気象衛星センターデータ処理部システム管理課
* Office of Observation Systems Operation, Observation Department, Japan Meteorological Agency
** 気象研究所気象衛星・観測システム研究部
- 34 ** System Engineering Division, Data Processing Department, Meteorological
Satellite Center
(Received September 1, 2015, Accepted 20 November 2015)
-1-
－ 42 －