Meteosat Surface Albedo Product: Users Manual Doc.No.
Issue
Date
WBS
:
:
:
:
EUM/OPS/MAN/12/2872
v2H
28 April 2014
EUMETSAT
Eumetsat-Allee 1, D-64295 Darmstadt, Germany
Tel: +49 6151 807-7
Fax: +49 6151 807 555 http://www.eumetsat.int
© EUMETSAT The copyright of this document is the property of EUMETSAT.
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
This page intentionally left blank. Page 2 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Document Change Record
Issue /
Revision
Date
MSA V. 1.1
March 2003
DCN.
No
Changed Pages / Paragraphs
Original document issued jointly by EUMETSAT
and DG JRC SEI
Existed only as PDF document
MSA V. 2.1
May 2010
Original title: Meteosat Surface Albedo (MSA)
Product User’s Manual and Format Guide
Original Document number EUM/FG/13
Original document NOT in Hummingbird DM
system
V. 2A
9 March 2012
Original document transcribed from single PDF to
word document, then created and stored in
Hummingbird DM system.
v.2B
24 Sept 2012
Addition of new table structures. Thorough scientific
review.
v.2C
24 Sept 2012
Addition of algorithm descriptions and Processing
Outline.
v.2D
25 Sept 2012
Addition of text and tables.
v.2E
16 October 2012
Version made read-only for review.
v.2F
16 October 2013
Addition to MSA Product Format Description
v.2G
2 April 2014
Updated after validation campaign.
v.2H
28 April 2014
Algorithm change history updated before release.
Page 3 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Table of Contents
1 2 3 4 5 6 7 Introduction .................................................................................................................................. 6
1.1 Purpose and Scope ............................................................................................................. 6
1.2 Document Structure ............................................................................................................. 6
1.3 Applicable Documents ......................................................................................................... 6
1.4 Reference Documents ......................................................................................................... 7
1.5 Acronyms and Abbreviations Used in this Document ......................................................... 8
1.6 Customer Contact and Product Ordering ............................................................................ 8
Outline of the Retrieval Procedure............................................................................................. 9
2.1 Overview.............................................................................................................................. 9
2.2 Measurement Vector ......................................................................................................... 10
2.3 State vector........................................................................................................................ 12
2.4 Surface Albedo Estimation ................................................................................................ 12
2.5 Product Generation ...........................................................................................................13
Algorithm Description ............................................................................................................... 14
3.1 Processing Outline ............................................................................................................ 14
3.2 Algorithm information input ................................................................................................ 17
3.2.1 Dynamic Input Files .............................................................................................. 17
3.2.2 Static Input Files ................................................................................................... 17
MSA Product Format Description ............................................................................................ 18
4.1 Overview............................................................................................................................ 18
4.2 Product Availability ............................................................................................................ 18
4.3 Retrieving the MSA product from UMARF ........................................................................ 19
4.4 MSA: BUFR format ............................................................................................................ 20
4.5 MSA: HDF4 format ............................................................................................................ 22
4.5.1 Naming convention ............................................................................................... 23
4.5.2 Global attributes .................................................................................................... 23
4.5.3 Scientific dataset................................................................................................... 25
4.6 Ancillary Data File .............................................................................................................. 26
4.6.1 Naming Convention .............................................................................................. 26
4.6.2 Global Attributes ................................................................................................... 26
4.6.3 Scientific data set.................................................................................................. 27
4.7 Static Data Set ................................................................................................................... 28
4.7.1 Naming convention ............................................................................................... 28
4.7.2 Global attributes .................................................................................................... 28
4.7.3 Scientific data set.................................................................................................. 28
Algorithm Assumptions ............................................................................................................ 29
Algorithm limitations and MSA product usage ...................................................................... 30
6.1 RTM in the VIS spectral band............................................................................................ 30
6.2 Cloud Contamination ......................................................................................................... 30
6.3 Conversion to broadband albedo ...................................................................................... 30
6.4 BHR. Error estimation ........................................................................................................31
6.5 Albedo colour palette ......................................................................................................... 33
Algorithm Change History ........................................................................................................ 34
7.1 Changes in Version 1.1 ..................................................................................................... 34
7.2 Changes in Version 1.2 ..................................................................................................... 34
7.3 Changes in Version 2.1 ..................................................................................................... 34
Page 4 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Table of Figures
Figure 1: MSA Retrieval scheme... ........................................................................................................ 10
Figure 2: Meteosat First Generation MVIRI Visible channel Sensor Spectral Response (SSR) ........... 11
Figure 3: Definition of the processing WINDOWS for the 0-degree,057-degree, and 063-degree ....... 14
Figure 4: Daily Accumulation Model (DAM): functionalities and interface ............................................. 15
Figure 5: Data Processing Module (DPM): functionalities and interface ............................................... 16
Figure 6: MSA temporal coverage for 0-DEG and IODC (57 and 63 degree East ). ............................. 18
Figure 7: Example of MSA product (DHR) for the 0-degree mission and the IODC mission ................ 19
Figure 8: Sensor spectral response of Meteosat-7 VIS band. ............................................................... 29
Table of Tables
Table 2-1: Discretisation values of the k, Ө, and r parameters.............................................................. 12
Table 3-1: GSA dynamic input files ........................................................................................................ 17
Table 3-2: Ancillary files GSA static input .............................................................................................. 17
Table 4-2: Operational Meteosat satellite time coverage detailed definition. ........................................ 19
Table 4-3: Header Fields valid for all pixels ........................................................................................... 21
Table 4-4: Retrieved fields valid for each pixel ...................................................................................... 22
Table 4-5: Name and description of the scientific data set. ................................................................... 27
Table 6-1: Empirical coefficients a–d for the DHR conversion .............................................................. 31
Table 6-2: Empirical coefficients a–d for the BHR conversion............................................................... 31
Table 6-3: RGB values per Albedo range used in MSA publication ...................................................... 33
Page 5 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
1
INTRODUCTION
1.1
Purpose and Scope
This document describes the Meteosat Surface Albedo (MSA) product version 2.1 generated on the MTP
MPEF reprocessing environment at EUMETSAT. The original version of this algorithm has been
developed at the Space Applications Institute of DG JRC of the European Commission. It is maintained
and improved by EUMETSAT.
1.2
1.3
Document Structure
Section
Contents
Section 1
Introduction
Section 2
Description of the retrieval strategy
Section 3
Description of the algorithm
Section 4
Description of the product format
Section 5
Algorithm assumptions
Section 6
Algorithm limitations and usage
Section 7
Algorithm change history
Applicable Documents
None
Page 6 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
1.4
Reference Documents
Number
Name, authors, source
RD 1
Engelsen, O.B Pinty, M. Verstraete, and J. Martonchik (1996,).
Parametric bi-directional reflectance factor models: evaluation, improvements and applications.
Technical Report EUR 16426 EN, Space Applications Institute, JRC.
RD 2
Govaerts, Y.M. (1999). Correction of the Meteosat-5 and-6 VIS band relative spectral response
with Meteosat-7 characteristics. International Journal of Remote Sensing 20, 3677–3682.
RD 3
Pinty, B., F. Roveda, M.M. Verstraete, N.Gobron, Y.Govaerts, V.Martonchik, D.J. Diner, and
R.A.Kahn (2000a). Surface albedo retrieval from Meteosat: Part1: Theory. Journal of
Geophysical Research 105, 18099–18112.
RD 4
Rahman, H., B. Pinty, and M.M. Verstraete (1993). Coupled surface-atmosphere reflectance
(CSAR) model.2. Semi-empirical surface model usable with NOAA Advanced Very High
Resolution Radiometer Data. Journal of Geophysical Research 98, 20, 791–20.
RD 5
Govaerts, Y., .Pinty, M. Taberner,and A. Lattanzio (2006). Spectral conversion of surface albedo
derived from Meteosat first generation observations. Geosciences and Remote Sensing Letters,
23–27 doi:10.1109/LGRS.2005.854202.
RD 6
Holben, B.,T.Eck, I.Slutsker, D.Tanre, J.Buis, A.Setzer, E.Vermote, J.Reagan, Y.Kaufman,
T.Nakajima, F.Lavenu, I.Jankowiak, A.Smirnov(1998). AERONET-a federated instrument
network and data archive aerosol characterization. Remote Sensing of Environment 66, 1–16.
RD 7
Lattanzio, A.,Y.Govaerts, and B.Pinty (2006). Consistency of surface anisotropy
characterization with Meteosat observations. Advanced Space Research,
doi:10.1016/j.asr.2006.02.049.
RD 8
Lattanzio, A. Meteosat Surface Albedo Retrieval Algorithm Theoretical Base Document
(ATBD), 2013. EUM/OPS-MSG/SPE/12/3367
RD 9
Loew, A. And Y.Govaerts (2010). Towards multi-decadal consistent meteosat surface albedo
time series. Remote Sensing 2(4), 957–967.
RD 10
Govaerts, Y., and Lattanzio, A. (2007). Retrieval Error Estimation of Surface Albedo Derived
from Geostationary Large Band Satellite Observations: Application to Meteosat-2 and -7 Data.
Journal of Geophysical Research 112, doi:10.1029/2006JD007313
RD 11
Pinty,B., F.Roveda, M.M. Verstraete, N.Gobron, Y.Govaerts, J.V. Martonchik, D.J.Diner, and
R.A.Kahn (2000a). Surface albedo retrieval from Meteosat: Part1: Theory. Journal of
Geophysical Research 105, 18099–18112.
RD 12
Pinty, B., F.Roveda, M.M.Verstraete, N.Gobron, Y.Govaerts, J.V.Martonchik, D.J.Diner,
andR.A.Kahn (2000b). Surfaceal bed or retrieval from Meteosat:
Part 2: Applications Journal of Geophysical Research 105, 18113–18134.
RD 13
Pinty, B. and D. Ramond (1987). A method for the estimate of broadband directional surface
albedo from a geostationary satellite. Journal of Climate and Applied Meteorology 26, 1709-1722.
RD 14
Diner, D. J., W. A. Abdou, A. T. P., K. Crean, H. R. Gordon, R. A. Kahn. J. V.Martonchik, S. R.
Paradise, B. Pinty, 1'1. M. Verstraete, M. Wang, and R. A. West (1997). 1'1ISR Level 2 Aerosol
Retrieval Algorithm Theoretical Basis. Technical Report JPL D-11400, Rev. C, NASA Jet
Propulsion Laboratory
Page 7 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
1.5
1.6
Acronyms and Abbreviations Used in this Document
Acronym
Meaning
AOT
Aerosol Optical Thickness
API
Application Programming Interface
ASM
Atmosphere Scattering Module
BHR
Bi-Hemispherical Reflectance
BRF
Bi-directional Reflectance Factor
DAM
Data Accumulation Module
DCP
Data Consistency Procedure
DG
Directorate General
DDHR
Daily averaged Directional Hemispherical Reflectance
DHR
Directional Hemispherical Reflectance
DIM
Data Interpretation Module
ECMWF
European Centre for Medium-range Weather Forecasts
HDF
Hierarchical Data Format
IODC
Indian Ocean Data Coverage
JRC
Joint Research Centre
MARF
Meteorological Archive Retrieval Facility
MPEF
Meteorological Product Extraction Facility
MSA
Meteosat Surface Albedo
MTP
Meteosat Transition Programme
RPV
Rahman Pinty Verstraete
RTM
Radiative Transfer Model
SDS
Scientific Data Sets
TAM
Space-Time Averaging Module
TOA
Top Of Atmosphere
UMARF
Unified Meteorological Archive and Retrieval Facility
VIS
Visible
Customer Contact and Product Ordering
All enquiries related to this document or to the Meteosat product should be directed here:
UMARF Customer Enquiries Eumetsat-Allee 1, D-64295 Darmstadt, Germany Tel: +49 6151 807-7 Fax: +49 6151 807-555 E-mail: archive@eumetsat.int
Page 8 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
2
OUTLINE OF THE RETRIEVAL PROCEDURE
2.1
Overview
The MSA algorithm relies on an approach proposed by Pinty et al. described in [RD 12] and [RD 13].
During the course of the day, observations acquired in the visible part of the electromagnetic spectrum at
different illumination conditions are accumulated. Ultimately, it retrieves the surface anisotropy and the
atmospheric aerosol load through the inversion of a Radiative Transfer Model (RTM) as shown in
Figure 1. The elements of the model are an absorbing layer (gases absorption), a scattering layer (aerosol
scattering) and an anisotropic surface (coupled with the scattering layer). This retrieval scheme relies on
the applicability of the reciprocity principle at a spatial scale of several kilometres (see Lattanzio and
Govaerts, 2006 [RD 7].
Assuming that the geophysical properties that control the radiance field emerging from a given pixel do not
evolve much over a day, the acquisition of radiance data over such a period of time corresponds to an
angular sampling of the same radiance field for various solar geometries. Such a strategy had already been
explored by Pinty and Ramond [RD 13] to study the seasonal variation of surface albedo over the Sahelian
regions. The lack of spectral information and the radiometric noise limit the possibility to unequivocally
discriminate among the various solutions that could fit the measurement vector at a given level of
confidence. This level of confidence depends on the size of the measurement vector that can change from
pixel to pixel according to cloud and illumination conditions. Hence, a probability is assigned to each
solution. This probability specifically depends on the number of degrees of freedom and the value of the
cost function. For details refer to [RD 8]. Because of the nature of the retrieval, it is necessary to further
constrain the surface retrievals by taking advantage of other currently available knowledge on the coupled
surface-atmosphere system. The proposed algorithm will use ozone and water vapour contents from other
sources as input data to reduce the problem: from a radiation transfer problem to a surface-aerosol
scattering problem. The most critical variables of such a system are then the aerosol optical depth and the
surface brightness. Indeed, it is assumed that actual atmospheric situations matches one of the limited
numbers of standard atmospheric models which can be prescribed from experience, and that the only free
atmospheric property to be estimated in the retrieval process is the aerosol load. The surface is described
with the Rahman Pinty Verstraete (RPV) model (see Rahman et al., 1993 [RD 4]. The surface brightness
is estimated during the same retrieval process from a set of predefined solutions (parameters k, Ө in Table
1) which describe the anisotropic properties of typical surfaces. The reflectance level 0 presented in
Section 2.2 is a not constraint parameter. The approach followed in solving this surface-aerosol scattering
problem is an extension of the MISR algorithm for retrieving aerosol optical depth values over dark
surfaces (Diner et al., 1997 [RD 14].
The algorithm also estimates the retrieval parameter error. The method for such estimation relies on a
statistical analysis of the solution ensemble that satisfies the inversion scheme given the measurement error
(for details refer to [RD 8]). A 10-day temporal compositing technique is applied to maximise the spatial
coverage of cloud-free pixels. Finally, the retrieved surface state variables are used to derive the
Directional Hemispherical Reflectance (DHR) corresponding to a sun position of 30 ° together with its
respective error. The estimated retrieval error and the probability of the solution are the two key elements
that permits a meaningful comparison of surface albedo derived from different radiometers.
Page 9 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
During the retrieval process, four assumptions are made:
1. surface and atmospheric scattering properties are constant throughout the day,
2. continental aerosol type can be applied everywhere and all year long,
3. surface anisotropy can be represented with the simple Bi-directional Reflectance Factor (BRF)
model proposed by Rahman et al., 1993 [RD 4],
4. the reciprocity principle is valid over terrestrial surfaces at a spatial resolution of a few kilometres
(Lattanzio and Govaerts, 2006 [RD 7]).
Figure 1: MSA Retrieval scheme. The observations accumulated during the day are used as an angular sampling of
the surface. The elements of the model are an absorbing layer, a scattering layer and an anisotropic surface.
2.2
Measurement Vector
Meteosat image acquisition, line (row) and pixel (column), results from a combination of the main mirror
rotation and the satellite’s spin. These images are acquired by the Meteosat Visible and Infrared Imager
(MVIRI) in three channels that includes a solar band (referred to as the VIS band), ranging from 0.4 µm up
to 1.1 µm. The spectral responses for the first generation Meteosat are shown in Figure 2.
Page 10 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Figure 2: Meteosat First Generation MVIRI Visible channel Sensor Spectral Response (SSR)
This VIS band is systematically calibrated against simulated radiances over stable bright desert targets
[RD2]. The technical characteristics of the Meteosat series of instruments and their orbital positions can be
found in the Meteosat First Generation User Handbook:
http://www.eumetsat.int/groups/ops/documents/document/pdf_td06_marf.pdf
The measurement vector consists of a time series of VIS band top-of-atmosphere (TOA) BRF accumulated
under different illumination (Ωs) angles and a stacic viewing (Ωv) condition during the course of one day.
This vector contains only clear-sky observations. Obvious cloudy conditions are screened by setting a
threshold value equal to 0.6 on the TOA BRF measurements. Also pixels with a value below 0.05,
considered as “water for sure”, are not considered for the retrieval. In order to eliminate unscreened events
such as remaining clouds, topography shadows, errors in the data geo-rectification process, and/or
significant sub-daily variations in the aerosol load and type, a Data Consistency Procedure produces an
angularly smooth TOA BRF series which accounts for hot spot conditions. This procedure checks the
consistency of the pre-screened TOA BRF values by attempting fitting the data series against a modified
version of the RPV model [RD1]. An iterative process eliminates the observed BRF value that exhibits the
largest absolute departure with respect to the model prediction. The result of the fit,
, between the
modelled TOA BRF and the remaining measurements provides an estimation of the filtering process cost,
accounting for an uncertainty (σDCP) between the data and the model. The number of cloud-free
observations can clearly change from place to place and time to time according to the duration of the day,
the cloud condition and, the actual number of available observations. A minimum of six clear-sky
observations per day are necessary to perform an inversion.
Page 11 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
2.3
State Vector
The state vector represents the ensemble of radiative parameters that describes the observed medium in the
VIS spectral band. It is actually divided into two categories. The first category represents the parameters
that are retrieved through the inversion of the measurement vector, namely the Aerosol Optical Thickness
(AOT) τ and three parameters {ρ0, Θ, k} characterising the state of the surface anisotropy. The second
category includes all the parameters, part of the observed medium, that are not retrieved due to the lack of
spectral information. They are the water vapour and ozone total column concentration noted, respectively
UH2O and UO3. The total column water vapour and the total column ozone are taken from the European
Centre for Medium-Range Weather Forecasts (ECMWF) analysed data.
The model used to describe the surface is the RPV [RD 4] and it reads:
Equation 1
where Ωs and Ωv are the illumination and viewing direction and z0 denotes the bottom of the atmosphere. ρ0
and
s (z0,
Ωs, Ωv; Θ, k) describe the amplitude and the angular field of the surface BRF, respectively.
Parameters
Values
k
0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0
Ө
2
τ
0.1, 0.2, 0.3, 0.4, 0.6, 0.8, 1.0
Table 1: Discretisation values of the k, Ө, and r parameters.
2.4
Surface Albedo Estimation
The DHR is defined as follows:
Equation 2
and is calculated for a sun position of 30 °. Assuming that the errors on {ρ0, Θ, k} are not correlated, the
non-systematic error on the estimation of the DHR is expressed with the following:
Equation 3
Page 12 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Since the variable of integration (the solid angle) is not derived, the derivation and integration operators
can be exchanged, and Equation 4 is produced:
Equation 4
Additional information on this error estimation can be found in [RD 11].
2.5
Product Generation
For each 10-day compositing periods, the following two physical quantities are calculated:
 The Directional Hemispherical Reflectance (DHR) estimated with Equation 2 using the solution
{ρ0, Θ, k}.
 The isotropic Bi-Hemispherical Reflectance BHRiso estimation with the following:
Equation 5
Refer to the ATBD [RD 7], Equation 46 for the definition of α0.
Page 13 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
3
ALGORITHM DESCRIPTION
3.1
Processing Outline
The MSA algorithm is organised into two main processing steps, responsible respectively for the daily
accumulation of the required input data and for the retrieval. The accumulation and retrieval is performed
dividing the processing area into smaller portions called windows. The windows for the 0-degree,
057-degree and 063-degree missions are shown in Figure 3.
0-degree
057-degree
063-degree
Figure 3: Illustration of the processing windows.
Page 14 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
For each window, an internal file containing the necessary input (DAM file) and an internal file containing
the results from the retrieval (SAM file) is generated. The definition and usage of those internal files is
completely transparent to the user; this information is included only for clarification.
In the first part of the processing, called Data Accumulation Module (DAM), the TOA measurements for
each pixel are stored in files called DAM files together with any other dynamical or static ancillary
information needed for the retrieval.
Figure 4: Daily Accumulation Model (DAM): functionalities and interface
The retrieval is performed in the second step, called Data Processing Module (DPM), after this daily
accumulation of the observations. To speed up the retrieval, the field of solutions for the RTM is
discretized and all the necessary integrals for of the radiative transfer equations are stored in Look-up
Tables. [RD 8] has detailed specifications. After a solution for each pixel is determined, the retrieval
information is stored in the Space Averaging Module (SAM) files. Finally, among all the solutions
retrieved for the 10-day period, one is selected and stored in the final product.
This retrieval approach is implemented in the following four steps in the DPM:
Page 15 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Before proceeding with the retrieval, two thresholds are applied to the accumulated TOA BRF and
measurements with a value below 0.05 and greater than 0.6 are excluded.
1. Data Consistency Procedure (DCP): This module is responsible for screening out the not
cloud-free pixels and it must be executed before entering the Atmospheric Scattering Module. This
procedure runs for every pixel. If a cloud mask is available (optional input), it is applied prior to the
DCP process.
2. Atmospheric Scattering Module (ASM): In this module, the gas absorption contribution to the Top
of Atmosphere reflectance is removed and the inversion of the RTM representing the scattering layer
and surface reflectance for all the possible values of the parameterised fields (see Govaerts and
Lattanzio, 2007 [RD 10]) is performed.
3. Data Interpretation Module (DIM): In this module, the most likely solution among all the possible
ones estimated in the previous step for each pixel is chosen.
4. Space-Time Averaging Module (TAM): Steps 1 to 3 are applied after the daily accumulation and
the solutions stored in the SAM temporary files for the subsequent 10-day temporal compositing. In
this latter module, the best solution for the 10-day period is selected.
Figure 5: Data Processing Module (DPM): functionalities and interface
Page 16 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
3.2
Algorithm Information Input
The MSA input can be divided in dynamic and static ancillary input files. The dynamic input files are
listed in Table 2.
3.2.1
Dynamic Input Files
File
Meaning
Radiance
Units
Radiance at pixel resolution retrieved in the instrument visible band
Cloud mask at pixel level.
If not present, the algorithm assumes all pixels are cloud-free.
0: Cloud free
1: Cloudy
Total Column Ozone (TCO3) and
Total Column Water Vapour (TCWV).
Cloud Mask
NWP Model
Reanalysis Data
If not present, default values are used :
-2
Wm sr
-1
1
TCO3: cm atm
TCWV: gcm-2
TCO3 : 0.3cm atm
TCWV : 2.0 gcm-2
Table 2: GSA dynamic input files
The NWP data, the Total Column Ozone (TCO3) and Total Column Water Vapour (TCWV) are used in
one of the Look-Up Tables (LUT) in order to invert the Radiative Transfer Model (RTM). Several ozone
and water vapour bands are located within the Meteosat visible spectral response and the effects of these
gases on radiation transfer processes must be considered. The sensitivity of the proposed retrieval schemes
with respect to those atmospheric parameters has been analysed in [RD 8].
3.2.2
Static Input Files
The Ancillary static input files are listed in Table 3. The Latitude and Longitude input files are generated
according to the MVIRI rectified image definition. Look-Up Tables (LUT) are defined in the
ATBD [RD 8].
File
Meaning
Look Up Table (LUT)
Binary files. The LUTs contains pre-computed integrals
used for the RTM inversion
Latitude
MVIRI rectified image
Longtitude
MVIRI rectified image
Table 3: Ancillary files GSA static input
Page 17 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
4
MSA PRODUCT FORMAT DESCRIPTION
4.1
Overview
One MSA product contains surface albedos estimated in the Meteosat VIS band spectral range. One
product generated every 10-day compositing periods. These compositing periods are defined with the
Julian day number. The first period runs from day 1 to day 10, the second from day 11 to day 20, and so
on. The last period of the year is therefore slightly shorter than the other periods and runs from day 361 to
365 (or to day 366 in a leap year). There is a maximum of 37 products per year.
The MSA product is not derived over the entire Meteosat disk and no sea mask is applied during the
generation process so that values over sea are also available is this area. The spatial resolution of the MSA
product is equal to the one of the Meteosat VIS band instrument. The distance between two adjacent pixels
at the sub-satellite point is equal to 2.5 km.
4.2
Product Availability
The MSA algorithm has been used to process Meteosat observations from 1982 (Meteosat-2) up
to 2006 (Meteosat-7) over the 0° position (0 degrees). Since summer 1998, EUMET has operated a second
geo-stationary satellite (Meteosat-5) over the Indian Ocean at position of 63° east (IODC). In 2006, a
satellite (Meteosat-7) has been moved at 571° east for keeping on covering the Indian Ocean region.
A detailed list of the availability of the different Meteosat First Generation satellites for the generation of
the MSA product is given in Figure 6 and Table 4
Figure 6: MSA temporal coverage for 0-DEG and IODC (57 and 63 degree East ).
Page 18 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Coverage
0 DEG
IOD
SSP
Satellite
Start Date
End Date
M2
16/08/1981
11/08/1988
M3
11/08/1988
25/01/1991
M4
19/06/1989
04/02/1994
M5
02/05/1991
13/02/1997
M6
21/10/1996
20/01/2000
M7
03/06/1998
19/07/2006
63° E
M5
01/07/1998
16/04/2007
57° E
M7
01/11/2006
ongoing
0° E
Table 5: Operational Meteosat satellite time coverage detailed definition.
Retrieving the MSA product from the EUMETSAT archive
The product can be ordered from EUMETSAT web page at the following URL:
http://archive.eumetsat.org/umarf/
Access to the product is only granted to registered users. In the list of products, the MSA product is
identified as the MTP Mean Surface Albedo 0100. Make sure you select the format you want.
The MSA product is available in BUFR and HDF4 formats.
Figure 7: Example of MSA product (DHR) for the 0-degree mission and the IODC mission.
Page 19 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
4.3
MSA: BUFR format
For a description of the fields stored in the BUFR file refer to Section 4.4.2 and Section 4.4.3. The World
Meteorological Organization maintains a complete code description and reference website:
http://www.wmo.int/pages/prog/www/WMOCodes.html
Table 6 and Table 7 identify the fields stored in the MSA product in BUFR format.
4.3.1
Header Fields valid for all pixels
Serial
Name
1
1007
SATELLITE IDENTIFIER
2
1031
IDENTIFICATION OF ORIGINATING/GENERATING CENTRE (SEE NOTE 10)
3
2020
2020 SATELLITE CLASSIFICATION
4
4001
YEAR
5
4043
DAY OF THE YEAR
6
4001
YEAR
7
4043
DAY OF THE YEAR
8
4004
HOUR
9
4005
MINUTE
10
4006
SECOND
11
49001
49001 TOTAL PRODUCTS (MSA)
12
49014
DAYS AVAILABLE TO DERIVE THE PRODUCT (MSA)
13
49002
MAJOR VERSION (MSA)
14
49003
PATCH LEVEL (MSA)
15
4001
YEAR
16
4043
DAY OF THE YEAR
17
49018
NUMERIC CALIBRATION VERSION (MSA)
18
49019
WATER REFLECTANCE THRESHOLD (MSA)
19
49020
CLOUD THRESHOLD (MSA)
20
49021
CLOUD SCREENING SMOOTHNESS (MSA)
21
6001
LONGITUDE (HIGH ACCURACY)
22
49022
PROBABILITY OF ERROR CONFIDENCE (MSA)
23
49023
AUTOCORRELATION COEFFICIENT (MSA)
24
49024
PERCENT GOOD PIXELS (MSA)
25
49025
MEAN REL RADIOMETRIC ERROR (MSA)
26
49026
MEAN AVAILABLE SLOTS (MSA)
27
49027
MEAN PROCESSED SLOTS (MSA)
28
49028
MEAN VALID PIXELS (MSA)
29
49029
MEAN WEAK SOLUTIONS (MSA)
30
49030
MEAN DUBIOUS SOLUTIONS (MSA)
31
49031
MEAN NUMBER SOLUTIONS (MSA)
32
49032
MEAN OPTICAL THICKNESS (MSA)
Page 20 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Serial
Name
33
5001
LATITUDE (HIGH ACCURACY)
34
6001
LONGITUDE (HIGH ACCURACY)
35
49033
MEAN PROBABILITY THRESHOLD (MSA)
36
49034
MEAN DHR30 (MSA)
37
49035
MEAN REL ERROR DHR30 (MSA)
38
49036
DHR SCALING FACTOR (MSA)
39
49037
BHRISO SCALING FACTOR (MSA)
40
49038
R0 SCALING FACTOR (MSA)
41
49039
TAUAVG SCALING FACTOR (MSA)
42
49040
TAUSTD SCALING FACTOR (MSA)
43
49041
CHI2ASM SCALING FACTOR (MSA)
44
49042
CHI2DCP SCALING FACTOR (MSA)
45
49043
RADRELERR SCALING FACTOR (MSA)
46
49044
DHRERR SCALING FACTOR (MSA)
47
49045
R0ERR SCALING FACTOR (MSA)
48
49046
KERR SCALING FACTOR (MSA)
49
49047
THETAERR SCALING FACTOR (MSA)
50
49048
TAUERR SCALING FACTOR (MSA)
51
49066
HIGH RES ROW NUMBER 1-5000
52
49066
HIGH RES ROW NUMBER 1-5000
53
49067
HIGH RES COLUMN NUMBER 1-5000
54
49067
HIGH RES COLUMN NUMBER 1-5000
Table 6: Header Fields valid for all pixels.
4.3.2
Retrieved fields valid for each pixel
Serial
Name
55
49049
BHRISO BYTE-CODED DAILY INTEGRATED DHR (MSA)
56
49050
DHR BYTE-CODED DIRECTIONAL HEMISPHERICAL REFLECTANCE (MSA)
57
49006
QUALITY FLAG (MSA)
58
49064
SOLUTIONS (MSA)
59
49065
INPUT SLOTS (MSA)
60
49065
INPUT SLOTS (MSA)
61
49068
SURFACE K PARAM (MSA)
62
49069
SURFACE THETA PARAM (MSA)
63
49069
SURFACE THETA PARAM (MSA)
64
49070
OPTICAL THICKNESS (MSA)
65
49051
R0 BYTE-CODED SURFACE REFLECTIVITY INTENSITY (MSA)
66
49052
R0ERR BYTE-CODED STANDARD ERROR (MSA)
67
49053
DHR30 BYTE-CODED ERROR OVER PERIOD (MSA)
Page 21 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Serial
Name
68
49014
DAYS AVAILABLE TO DERIVE THE PRODUCT (MSA)
69
49015
BEST DAY OF THE PERIOD (MSA)
70
49054
X2ASM BYTE-CODED COST OF ATMOSPHERIC SCATTERING MODULE (MSA)
71
49055
X2DCP BYTE-CODED COST OF DATA CONSISTENCY MODULE (MSA)
72
5001
LATITUDE (HIGH ACCURACY)
73
6001
LONGITUDE (HIGH ACCURACY)
74
49056
PROBABILITY (MSA)
75
49057
DHRERR10D BYTE-CODED DHR30 ERROR BEST DAY (MSA)
76
49058
K ERROR BYTE-CODED (MSA)
77
49059
T ERROR BYTE-CODED (MSA)
78
49060
TAU ERROR BYTE-CODED (MSA)
79
49061
TAU MEAN BYTE-CODED (MSA)
80
49062
TAU STD BYTE-CODED (MSA)
81
49063
RADIOMETRIC NOISE BYTE-CODED (MSA)
82
49066
HIGH RES ROW NUMBER 1-5000
83
49067
HIGH RES COLUMN NUMBER 1-5000
Table 7: Retrieved fields valid for each pixel
4.4
MSA: HDF4 format
Each format contains the same fields. Field names of the HDF format are described here. Each 10-day
product is composed of three files:
First file
Contains the albedo values and its associated quality indicator.
Example: MSA_Albedo_L2.0_V2.01_000_2000_311_320.HDF
Second file
Contains all the ancillary information.
Example: MSA_Ancillary_L2.0_V2.01_000_2000_311_320.HDF
Static file
Contains the geographical location of each pixel.
Example: MSA_Static_L2.0_V2.10.HDF
Go to this website for complete specifications on the HDF format:
http://www.hdfgroup.org/products/hdf4/
Page 22 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
4.4.1
Naming convention
The file name has the following structure:
MSA_Albedo_L2.0_Vm.nn_sss_yyyy_fff_lll.HDF
where:
m
=
MSA algorithm major version number
nn
=
MSA algorithm minor version number
sss
=
Sub-satellite point code (000,063)
yyyy =
four digit of the year
fff
=
First Julian day of the period
lll
=
Last Julian day of the period
4.4.2
Global attributes
The table on the two pages that follow contains all the global attributes for the MSA Albedo file.
Page 23 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Attribute
Description
BUFR Field Index
Nominal SSP
Center of the reference grid in the Meteosat projection.
21
Start line
The line number (row) of the first pixel of the datasets in the Meteosat Projection for the visible resolution
(5000 × 5000 pixels). Indexing starts at 1.
53
Height
The y dimension of the datasets.
54
Start pixel
The pixel number (column) of the first pixel of the datasets in the Meteosat projection for the visible
resolution (5000 × 5000 pixels). Indexing starts at 1.
51
Width
The x dimension of the datasets.
52
HDF Conversion time
The satellite identifier (in the range [1–7]).
Not applicable
Satellite number
The satellite identifier (in the range [1–7]).
Start year
The starting year (number) of the compositing period.
4
Start Julian day
Start Julian day of the period.
5
End year
The ending year (number) of the compositing period.
6
End Julian day
End Julian day of the period.
7
Number of products
The total number of valid pixels in each data set.
11
Actual Nbr Day
The actual number of days in the compositing period (in the range [1–10]).
12
MSA major version
Version of the MSA algorithm.
13
MSA minor version
Patch level of the major version number.
14
Calibration version
Version of the calibration datasets.
17
Water Reflectance threshold
Minimum TOA BRF below which observations have been rejected.
18
Cloud for Sure threshold
MaximumTOA BRF above which observations have been rejected.
19
Cloud screening smooth
TOA smoothness used for the cloud screening
20
Probability alpha
Confidence level for the error estimation.
22
Autocorrelation coefficient
Coefficient α in Equation (16).
23
Percent good pixels
Percentage of pixels with a valid solution among the total number of processed pixels.
24
Mean Relative Radiometric Error
Mean relative radiometric error (Equation 5) in per cent.
25
Page 24 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product Users Manual
Attribute
Description
BUFR Field Index
Mean Number Slots
Mean number of available observation per pixel prior to the cloud screening.
26
Mean Number Processed Slots
Mean value of Ny, the size of the measurement vector ym.
27
Mean Valid Pixels
Percentage of valid pixels.
28
Mean Weak Solutions
Percentage of pixels with a probability between 10 % and 50 %.
29
Mean per cent Dubious
Percentage of pixels with a probability smaller than 10 %.
30
Mean per cent no solutions
Percentage of pixels with no solution.
31
Mean Optical Thickness
Mean optical thickness.
32
Actual SSP latitude
Mean value of the actual spacecraft SSP latitude.
33
Actual SSP longitude
Mean value of the actual spacecraft SSP latitude.
34
Mean probability threshold
Mean value of the probability P(σy).
35
Mean DHR 30
Mean value of the DHR.
36
Mean DHR 30 relative error
Mean value of the DHR relative error in percent.
37
MSA version string
MSA algorithm version number as a string.
Not Applicable
Start day
The starting day of the month (number) of the compositing period.
Not Applicable
Start month
The starting month (number) of the compositing period.
Not Applicable
End day
The end day of the month (number) of the compositing period.
Not Applicable
End month
The end month (number) of the compositing period.
Not Applicable
Start date string
The starting day of the 10-day compositing period. The format is dd/mm/yyyy plus a space character, the
letter J and a three-digit number indicating the Julian day of the year.
Not Applicable
End date string
The end date of the 10-day compositing period. The format is dd/mm/yyyy plus a space character, the letter
J and a three-digit number indicating the Julian day of the year.
Not Applicable
Period index
The number obtained by dividing the ending Julian day by 10.
Not Applicable
Processing time
Date and time when the product has been generated.
Not Applicable
Satellite generation
Always MET
Not Applicable
Figure 8: Global attributes: MSA Albedo file
Page 25 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
4.4.3
Scientific dataset
Each SDS is a bi-dimensional matrix of pixels corresponding to an image in the Meteosat projection with
the same east-west and south-north scanning mode. These values are coded on one byte with the value 255
used to indicate invalid pixels. Decoded values are calculated as with DV = CAL × (CV – Offset) where
DV is the decoded value and CV is stored value. CAL and OFFSET fields are available in the
CALIBDATA HDF structure. The conversion of the DHR 30 and BHRiso fields into broadband
albedo is given in Section 4.3.
Name
Description
The isotropic Bi-Hemispherical Reflectance field contains the
surface albedo in the Meteosat VIS band that would have been
observed under isotropic illumination conditions for the best
BHRiso
solution of day with
(d) =
BUFR Field Index
55
of the compositing period.
The Directional Hemispherical Reflectances field (DHR30)
contains the surface albedo value for the best solution of day d
DHR30
with (d) = of the compositing period. It represents the
spectral albedo in the Meteosat sensor VIS band spectral
interval assuming a sun zenith angle of µ0 = 30°. Since the angle
µ0 is the same for all pixels throughout the year, the DHR is
appropriate for the monitoring of the spatial or temporal
changes of the surface radiative properties.
56
DHR30 error 10D
This field represents the estimated DHR error.
74
Probability(%)
This field contains the probability of the solution of the
selected day d of the compositing period. Pixels with a
probability smaller than 80 % or 90 % should not be considered
when the MSA product is analysed.
73
4.5
Ancillary Data File
4.5.1
Naming Convention
The file name has the following structure:
MSA_Ancillary_L2.0_V m.nn_sss_yyyy_ff_flll.HDF
where:
m
=
MSA algorithm major version number
nn
=
MSA algorithm minor version number
sss
=
Sub-satellite point code (000,063)
yyyy =
four digit of the year
fff
=
First Julian day of the period
lll
=
Last Julian day of the period
4.5.2
Global Attributes
See Section 4.4.2.
Page 26 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
4.5.3
Scientific data set
Each scientific data set is a bi-dimensional matrix of pixels corresponding to an image in the Meteosat
projection with the same east-west and south-north scanning mode. These values are coded on one byte
with the value 255 used to indicate invalid pixels. Decoded values are calculated as done in
DV = CAL × (CV – Offset), where DV is the decoded value and CV is stored value. CAL and OFFSET
fields are available in the CALIBDATA HDF structure.
Name
QUALITYFLAG
Description
BUFR Field
Index
This flag takes the following values:
Val
0
1
2
3
4
Meaning
OK
No valid days in the period
No valid samples in the period
No likely day
Invalid solution index
5
Dubious solution , Pa( (d)) = 0.1
6
Weak solution, 0.1 < Pa( (d)) ≤ 0.5
57
NUMSOL
The number of acceptable solutions. The value 255 is set for invalid pixels.
58
NSLOT
Number of inputs lots before cloud screening.
59
NSLOTASM
Number of clear sky input slots.
60
K
Parameter describing the shape of the surface BRF in the RPV model for
the best fit.
61
THETA
Parameter describing the asymmetry of the surface BRF in the RPV model
for the best fit.
62
AER_OPT_THICK
Estimated equivalent aerosol optical thickness.
63
Name
Description
BUFR Field
Index
R0
Amplitude of the surface BRF in the RPV model for best fit.
64
ERR_R0
Estimated error of
65
NUMDAYS
Selected day during the compositing period.
0.
67
2
BESTDAY
Normalised cost function of the best solution given by Equation (4) χ /Ny.
68
Chi2DCP
Cost function of the cloud screening.
70
Chi2ASM
Cost function of the invertion.
69
DHRError
Estimated error of the DHR for the best day d.
66
ERR_K
Estimated error of
75
ERR_T
Estimated error of
76
ERR_OPT
Estimated error of
77
AVGOPT
Average value of
ERR_AVG_ERR
Standard deviation of AVGOPT.
79
RADIOMETRIC
NOISE
Mean daily radiometric noise of the best day d calculated with
(ΣNy σy(t))/Ny
80
during the compositing period.
Table 8: Name and description of the scientific data set.
Page 27 of 34
78
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
4.6
Static Data Set
4.6.1
Naming convention
The file name has the following structure:
MSA_Static_L2.0_Vm.nnsss.HDF
Where:
m
=
MSA algorithm major version number
nn
=
MSA algorithm minor version number
sss
=
Sub-satellite point code (000, 057, 063)
4.6.2
Global attributes
Contains a subset of the fields listed in Section 4.4.2.
4.6.3
Scientific data set
Name
Description
Symbol
Units
Min
Max
latitude
Pixel latitude
λ
degree
-90
+90
longitude
Pixel longtitude
φ
degree
-180
+180
Page 28 of 34
Scaled
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
5
ALGORITHM ASSUMPTIONS
The approach relies on a daily accumulation of Meteosat observations acquired at about 20 different
illumination conditions to characterise the scattering properties of the surface and the atmosphere,
assuming that:
1. Surface and atmospheric scattering properties are constant throughout the day.
2. A US62 vertical profile and continental aerosol type applied everywhere and throughout the year.
3. Surface anisotropy can be represented with the simple BRF model proposed by Rahman et al.
[RD 4].
4. The reciprocity principle is valid over terrestrial surfaces at a spatial resolution of a few kilometres.
See also [RD 9].
Figure 9: Sensor spectral response of Meteosat-7 VIS band (dash-dotted line) and of Terra-MISR bands (solid lines).
The dotted line represents the total transmittance in a US standard atmosphere with an aerosol optical thickness of
0.2 at 0.5 µm. The solid line with the ⋄symbols illustrates typical vegetated surface reflectance.
Page 29 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
6
ALGORITHM LIMITATIONS AND MSA PRODUCT USAGE
6.1
RTM in the VIS spectral band
The radiometer on board the Meteosat First Generation (MFG) satellites acquires radiances twice per hour
in a single large solar spectral band ranging from approximately 0.4 µm up to 1.1 µm. This is referred to as
the VIS band (Fig.1). This interval contains some strong gas absorption bands and is also subject to aerosol
scattering-absorption processes whose magnitude varies with wavelength. Consequently, the decoupling
between absorption and scattering processes introduces some inaccuracies when the spectral integration is
performed over such a large interval. This assumption is implemented in the retrieval algorithm to speed up
the atmospheric correction scheme. Additionally, vegetated surface reflectance exhibits quite strong and
fast spectral variations over this spectral region as a consequence of the differences in the radiation transfer
regimes occurring on both sides of 0.7 µm. This is mainly governed by absorption (scattering) at
wavelengths shorter (larger) than 0. 7 µm (Figure 9). These spectral variations cannot be explicitly taken
into account in the atmospheric-correction scheme because observations occur only in one single band
under different illumination conditions. Consequently, surface albedo derived from geostationary satellite
observations in the VIS band with the MSA algorithm is subject to some systematic biases depending on
the shape of the surface spectra, the aerosol load, and the absorbing gas concentration. Under such
circumstances, these spectral effects that occur within the VIS band need to be corrected before performing
the spectral conversion, or any other quantitative analysis of the MSA product.
6.2
Cloud Contamination
The cloud detection method described in Pinty et al. [RD 3] is based on the temporal analysis of the ToA
BRF, and does not perform optimally when the cloud cover remains stable during an entire day.
Consequently, some surface albedo pixels might still be contaminated by undetected clouds. Applying
conservative filtering on the probability field and the DHR estimated error could minimize cloud
contamination. We recommended restriction of the analysis of the DHR or BHR• fields to those pixels with
a probability Pa larger than 80 % or 90 % and a σDHR/DHR relative error over 10 days smaller than 50.
6.3
Conversion to Broadband albedo
The method to transform the DHR derived in the Meteosat VIS band DHR VIS in shortwave broadband
albedo DHRBB is described in Govaerts et al.[RD 10]. This relationship is defined with a third-order
polynomial written as follows:
Equation 6
Recently, Loew and Govaerts [RD 10] found some temporal discrepancies when the method described in
[RD 2] is used for the spectral conversion of Meteosat-2 to Meteosat-4. These authors developed an
empirical method to derive the spectral coefficients for the Meteosat-2 to Meteosat-4 satellites.
The empirical coefficients a–d for the DHR conversion take the following values according to the
Meteosat number [RD 10]:
Page 30 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
Met
a
b
c
2
-2.95364443e-05
1.22636437e+00
-1.45464587e+00
1.27798259e+00
3
-2.95364443e-05
1.32036722e+00
-1.52968502e+00
1.25365901e+00
4
-2.95364589e-05
1.22655797e+00
-1.07426369e+00
8.96015048e-01
5
-2.95364443e-05
1.25341415e+00
-1.09384084e+00
8.89843404e-01
6
-2.95364443e-05
1.30573940e+00
-1.31526375e+00
1.05711114e+00
7
-2.95364589e-05
1.26273489e+00
-1.11476350e+00
9.00940299e-01
d
Table 9: Empirical coefficients a–d for the DHR conversion
.
For the conversion of the BHR , the empirical coefficients are as follows in Table 10:
Met
a
b
c
2
-2.85976712e-05
9.81895685e-01
-8.48408699e-01
7.43798614e-01
3
-2.85976712e-05
1.09896255e+00
-1.07471538e+00
9.11732554e-01
4
-2.85976712e-05
1.00361478e+00
-6.55005634e-01
6.47315860e-01
5
-2.85976712e-05
1.04928327e+00
-7.66418219e-01
7.47902989e-01
6
-2.85976712e-05
1.15992260e+00
-1.13301563e+00
9.98916626e-01
7
-2.85976712e-05
1.03751910e+00
-6.88233614e-01
7.00615168e-01
d
Table 10: Empirical coefficients a–d for the BHR conversion
6.4
BHR. Error estimation
The BHR• error is currently not stored in the distributed product. This error can be calculated with the
following expression:
Equation 7
Page 31 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
Where α0 is tabulated with the following value:
h
Θ
k
α0
h
Θ
k
α0
h
Θ
k
α0
0.15
-0.30
0.40
3.29568
0.15
-0.20
0.40
3.01252
0.15
-0.10
0.40
2.74655
0.15
-0.30
0.50
2.91138
0.15
-0.20
0.50
2.64497
0.15
-0.10
0.50
2.39410
0.15
-0.30
0.60
2.62286
0.15
-0.20
0.60
2.36857
0.15
-0.10
0.60
2.12919
0.15
-0.30
0.70
2.40092
0.15
-0.20
0.70
2.15551
0.15
-0.10
0.70
1.92501
0.15
-0.30
0.80
2.22700
0.15
-0.20
0.80
1.98812
0.15
-0.10
0.80
1.76452
0.15
-0.30
0.90
2.08885
0.15
-0.20
0.90
1.85469
0.15
-0.10
0.90
1.63641
0.15
-0.30
1.00
1.97802
0.15
-0.20
1.00
1.74715
0.15
-0.10
1.00
1.53294
0.15
-0.25
0.40
3.15165
0.15
-0.15
0.40
2.87767
0.15
-0.05
0.40
2.61871
0.15
-0.25
0.50
2.77600
0.15
-0.15
0.50
2.51782
0.15
-0.05
0.50
2.27346
0.15
-0.25
0.60
2.49365
0.15
-0.15
0.60
2.24720
0.15
-0.05
0.60
2.01425
0.15
-0.25
0.70
2.27618
0.15
-0.15
0.70
2.03856
0.15
-0.05
0.70
1.81463
0.15
-0.25
0.80
2.10550
0.15
-0.15
0.80
1.87455
0.15
-0.05
0.80
1.65780
0.15
-0.25
0.90
1.96964
0.15
-0.15
0.90
1.74369
0.15
-0.05
0.90
1.53264
0.15
-0.25
1.00
1.86037
0.15
-0.15
1.00
1.63808
0.15
-0.05
1.00
1.43151
h
Θ
k
α0
0.15
0.00
0.40
2.49373
0.15
0.00
0.50
2.15556
0.15
0.00
0.60
1.90210
0.15
0.00
0.70
1.70718
0.15
0.00
0.80
1.55420
0.15
0.00
0.90
1.43218
0.15
0.00
1.00
1.33363
and
Page 32 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
6.5
Albedo colour palette
The colour palette used in MSA publication is given in this section.
Albedo range
rgb value
0.000 – 0.020
(0, 0, 050)
0.020 – 0.040
(0, 0, 100)
0.040 – 0.050
(0, 0, 200)
0.050 – 0.060
(0, 0, 255)
0.060 – 0.070
(80, 20, 10)
0.070 – 0.080
(100 , 30, 20)
0.080 – 0.090
(100, 50, 30)
0.090 – 0.100
(80, 60, 40)
0.100 – 0.110
(60, 80, 40)
0.110 – 0.120
(20, 80, 20)
0.120 – 0.130
(30, 100, 30)
0.130 – 0.140
(40, 120, 40)
0.140 – 0.150
(50, 140, 50)
0.150 – 0.160
(60, 160, 50)
0.160 – 0.170
(80, 180, 30)
0.170 – 0.180
(100, 160, 20)
0.180 – 0.190
(110, 150, 10)
0.190 – 0.200
(120, 140, 10)
0.200 – 0.210
(140, 120, 0)
0.210 – 0.220
(130, 110, 0)
0.220 – 0.230
(125, 99, 0)
0.230 – 0.240
(120, 80, 0)
0.240 – 0.250
(111, 75, 0)
0.250 – 0.260
(120, 82, 7)
0.260 – 0.280
(126, 91, 14)
0.280 – 0.300
(141, 108, 28)
0.300 – 0.325
(156, 125, 42)
0.325 – 0.350
(171, 142, 56)
0.350 – 0.375
(186, 159, 71)
0.375 – 0.400
(201, 176, 85)
0.400 – 0.450
(216, 193, 99)
0.450 – 0.500
(231, 210, 113)
0.500 – 0.550
(240, 220, 120)
0.550 – 0.600
(246, 225, 135)
0.600 – 0.650
(246, 235, 155)
0.650 – 0.700
(240, 240, 180)
0.700 – 0.750
(250, 250, 210)
0.750 – 0.800
(230, 253, 200)
0.800 – 0.900
(220, 230, 240)
Table 11: RGB values per Albedo range used in MSA publication
Page 33 of 34
EUM/OPS/MAN/12/2872
v2H, 28 April 2014
Meteosat Surface Albedo Product: Users Manual
7
ALGORITHM CHANGE HISTORY
7.1
Changes in Version 1.1
The following changes have been implemented in version 1.1 with respect to version 1.0: 1.
1. The processed area over the zero degree mission has been increased.
2. A processing area over the 63° emission has been defined.
3. The computation accuracy of the sun angles has been improved.
4. The computation of the viewing angles now accounts for the actual position of the spacecraft at the
acquisition time.
5. The discretisation range of the predefined conditions for the , , and
= 0.1
0.2
0.4
0.6
1.0
7.2
=
HG
0.5
0.6
0.7
0.8
0.9
1.0
parameters is as follows:
HG
= −0.3
−0.2
−0.1
−0.0
Changes in Version 1.2
The following changes have been implemented in version 1.2.
1. An error has been corrected in the access of the ECMWF total water column field.
7.3
Changes in Version 2.1
Version 2.1 contains major changes as it includes now an a priori estimation of the retrieval error of each
parameter and the DHR.
1. The minimum number of slots is set to six.
2. χ2 is estimated are real χ2 function.
3. The radiometric error is not prescribed anymore but is estimated for each pixel.
4. The retrieval mechanism accounts for the aerosol load auto-correlation during the day.
5. The threshold on the χ value is not prescribed anymore but calculated as a function of its
probability.
6. The computation of the viewing angles accounts now for the actual position of the spacecraft
at the acquisition time.
7. The error of the four retrieved parameters is estimated.
8. The 10-day compositing technique is based on a minimising of the error.
9. The discretisation range of the τ, k, and Θ parameters. See Table 1.
Page 34 of 34