1. Masaya Takahashi (JMA) and Sebastien Wagner (EUMETSAT)
Format for GSICS VIS/NIR products and requirements for GSICS Plotting Tool
Masaya Takahashi (JMA)
and
Sebastien Wagner (EUMETSAT)

2. Overview and Purpose Overview Purpose of this talk
GEO-LEO-VISNIR product(s) will enter Demonstration-Phase in 2016
SEVIRI vs. Aqua/MODIS product: in GPPA
At present, NASA DCC method is only available, but products (netCDF) will contain multiple calibration methods’ results in one file in future
NetCDF convention: almost fixed, but need to discuss the use of enhanced data model (grouping function on netCDF-4) within GRWG/GDWG
Requirement documents for GSICS Plotting Tool should be prepared by GRWG
Purpose of this talk
To refresh your memory on past discussions and current proposal
To discuss / get consensus on remaining issues
GDWG discussion results will be reported on Friday

3. Topic-1: NetCDF convention for GSICS VIS/NIR products

4. Current Convention for the GEO-LEO-VISNIR (DCC) product
Basic idea: to follow existing GEO-LEO-IR convention
File naming
E.g., W_XX-EUMETSAT-Darmstadt,SATCAL+RAC+GEOLEOVNIR,MSG3+SEVIRI-Aqua+MODIS_C_EUMG_ _demo_01.nc
netCDF Convention
Global attributes
Nothing special issues to be discussed
Variables
Basic variables (e.g. date, channel name)
Official calibration coefficients (This will also be discussed for GEO-LEO-IR)
Specific variables to GSICS VISNIR products in order to support multiple calibration methods

5. VIS/NIR netCDF Convention - Global Attributes (1)
:atbd_doc_doi = "N/A" ;
:atbd_doc_url = " ;
:conventions = "CF-1.6" ;
:creator_ = ;
:creator_name = "EUMETSAT”;
:creator_url = " ;
:date_created = " T13:09:12Z" ;
:date_modified = " T13:09:12Z" ;
:geospatial_lat_max = 60.f ;
:geospatial_lat_min = -60.f ;
:geospatial_lat_units = "degrees_north" ;
:geospatial_lon_max = 60.f ;
:geospatial_lon_min = -60.f ;
:geospatial_lon_units = "degrees_east" ;

6. VIS/NIR netCDF Convention - Global Attributes (2)
:history = "Blend T13:09:12Z v1.0.0; …" ;
:id = "W_XX-EUMETSAT-Darmstadt,SATCAL+RAC+GEOLEOVNIR,MSG3+SEVIRI-AQUA+MODIS_C_EUMG_ _demo_01.nc" ;
:institution = "EUMETSAT" ;
:keywords = "GSICS, satellites, inter-calibration, reflective solar bands" ;
:license = "Calibration information delivered as a GSICS operational product is generated in accordance with GSICS principles…" ;
:local_data_subcategory = 3s ;
:metadata_conventions = "Unidata Dataset Discovery v1.0" ;
:monitored_instrument = "SEVIRI" ;
:monitored_instrument_wmo_code = "57 207" ;
:naming_authority = "int.eumetsat.gsics" ;
:processing_level = "demonstration/v1.0.0" ;

7. VIS/NIR netCDF Convention - Global Attributes (3)
:reference_instrument = "MODIS" ;
:reference_instrument_wmo_code = " " ;
:references = "ATBD, Unidata NetCDF, Climate Format Conventions" ;
:standard_name_vocabulary = "CF Standard Name Table (Version 19, 22 March 2012)" ;
:summary = "Coefficients of the GSICS Correction for the reflective solar bands of a GEOstationary imager using a LEO imager reference instrument. The current version includes deseasonalisation in the processing." ;
:time_coverage_end = " T13:09:11Z" ;
:time_coverage_start = " T13:09:11Z" ;
:title = "MSG3+SEVIRI vs Aqua+MODIS GSICS Re-Analysis Correction" ;
:window_period = "P-15D+15D" ;
:wmo_data_category = 30s ;
:wmo_international_data_subcategory = 5s ;
:project = "Global Space-based Inter-Calibration System " ;
:srf_url = “SRF (netCDF) URL”; // Optional Global Attribute, to be discussed in GDWG session
GDWG_2015.6a5:The GEOLEOVNIR or LEOLEOVNIR shall include an optional global attribute referencing the location of the associated SRF netCDF file.

8. VIS/NIR netCDF Convention - Variables (1)
Example of SEVIRI DCC product
float central_wavelength(chan) ;
:long_name = "nominal channel central wavelength" ;
:units = "m" ;
char channel_name(chan, chan_strlen) ;
:long_name = "channel identifier" ;
double date(date) ;
:bounds = "validity_period" ;
:units = "seconds since T00:00:00Z" ;
float mon_official_offset(date, chan) ;
:long_name = "official calibration offset" ;
:units = "W m-2 sr-1 um-1" ;
float mon_official_slope(date, chan) ;
:long_name = "official calibration slope" ;
double validity_period(date, validity) ;
:long_name = "correction validity period" ;
central_wavelength(1) = ;
channel_name(1,5) = "VIS06" ;
date(2) = , ;
mon_official_offset(2,1) =
, ;
mon_official_slope(2,1) =
, ;
validity_period(2,2) =
, ,
, :
date[1]
date[2]

9. VIS/NIR netCDF Convention - Variables (2)
Example of SEVIRI DCC product
char method_name(method, method_strlen) ;
:long_name = "method identifier" ;
float weight_method(date, chan, method) ;
:long_name = "weight of the methods used for the blend in each channel" ;
:units = "1" ;
float mon_offset(date, chan, method) ;
:long_name = "calibration offset with respect to reference instrument" ;
:units = "W m-2 sr-1 um-1" ;
float mon_slope(date, chan, method) ;
:long_name = "calibration slope with respect to reference instrument" ;
method_name(2) = "DCC", "BLEND" ;
weight_method(2,1,2) =
1, 1,
1, 1 ; a
mon_offset(2,2) =
_, _,
_, _ ;
mon_slope(2,1,2) =
, ,
, ;
date[1]
date[2]
DCC BLEND

10. Discussion on products updating frequency
Current VISNIR netCDF Convention works if calibration updating frequency is the same amongst inter-calibration methods (e.g. DCC, Moon, Rayleigh).
double date(date) ;
:bounds = "validity_period" ;
:units = "seconds since T00:00:00Z";
char method_name(method, method_strlen) ;
:long_name = "method identifier" ;
float weight_method(date, chan, method) ;
:long_name = "weight of the methods used for the blend in each channel" ;
:units = "1" ;
float mon_slope(date, chan, method) ;
:long_name = "calibration slope with respect to reference instrument" ;
:units = "W m-2 sr-1 um-1" ;
date(2) = , ;
method_name(3) =
"DCC", “Moon”, "BLEND" ;
weight_method(2,1,3) = a
0.6, 0.4, 1,
0.6, 0.4, 1 ;
mon_slope(2,1,3) =
0.5828, , ,
0.5633, , ;
Same dates for all the calibration methods
date[1]
date[2]
DCC Moon BLEND

11. Summary on products updating frequency at GSICS web meeting in Nov
It was clarified that this is distinct from the window period. The preferred approach was to issue frequent updates, together with uncertainties. This was to avoid step functions or discontinuities in the GSICS calibrations. Since real time users have scripts to download files, a consistent update frequency is desired. If there is an unscheduled calibration update, this will need to be a manual update for users, which is burdensome.
It was agreed that the GPRC would decide their own windowing period, outlier removal and trend analysis of the DCC observations to provide users the resulting calibration coefficients. The windowing periods are unique to every GEO domain.

12. Discussion on products updating frequency
Do we agree to use consistent updating frequency (i.e. use the same date) for all calibration methods?
It would be possible, but depends on blending method…
Solution-1: to use netCDF-4 “grouping” function (enhanced format)
SEVIRI vs. Aqua/MODIS VISNIR product
Combined result (DCC+Moon+…)
Global attributes, Dimensions, Variables
Method1 (Deep Convective Cloud)
Global attributes, Dimensions, Variables
Method2 (Moon)
Global attributes, Dimensions, Variables
Method3 (Desert)
Global attributes, Dimensions, Variables
Yu and Wu (2015), Remote Sensing of Environment
This is NOT the GDWG recommendation, but if GRWG really need, GDWG would resume the discussion/

13. Solution-2 To add calibration method names to variable names
float blend_date(date, chan) ;
float blend_mon_slope(date, chan) ;
float blend_mon_offset (date, chan) ;
float dcc_date(date, chan) ;
float dcc_mon_slope(date, chan) ;
float dcc_mon_offset (date, chan) ;
float moon_date(date, chan) ;
float moon_mon_slope(date, chan) ;
float moon_mon_offset(date, chan) ;
float moon_irr_giro (date, chan) ; …

14. Any other issues to be discussed On VISNIR netCDF Convention?
All the VISNIR netCDF Convention will be summarized on:

15. Topic-2: Requirements for GSICS (Bias) Plotting Tool

16. GSICS Calibration Products Plotting Tool
GPRCs have their own product monitoring/plotting websites
“GSICS Plotting Tool” has been developed at EUMETSAT
Common look and feel (easy comparison among multiple products)
Dynamical plotting (no need to prepare lots of figures)
At present, GEO-LEO-IR is only supported
Primary user of the plotting tool: product users, NOT developers

17. Summary on plotting tool at GSICS web meeting in Nov. 2015
Plotting Tool: it was confirmed that the plotting tool should be able to generate time series plots of nominal and corrected calibration coefficients (gains) and biases, as well as their uncertainties. It was felt that while GPRCs should also monitor the PDFs of DCC distributions as intermediate products, these should not be necessary for the users.
Action: Masaya Takahashi and Sebastien Wagner will draft requirements for application of plotting tool to DCC products for presentation at 2016 GRWG meeting.
Units would be important. For example:
Calibration gain in [W/m2/sr/um] not easy-to-understand calibration trends
Standardized values [%] at “Day-1” of the result would be user-friendly
-> Progress: behind schedule…

18. Units of plotting variables
Time series of MSG-2/SEVIRI vs. Aqua/MODIS DCC RAC product by EUMETSAT (S. Wagner)
Monitored Instrument Gain
[W/m2/sr/um]
Option 1)
Inter-calibration gain
[W/m2/sr/μm]
Option 2)
Inter-calibration gain, scaled at the first result [%]
Monitored Instrument Gain
[%]

19. Discussion on GSICS Plotting Tool
GRWG requirement on the Plotting Tool:
“time series plots of nominal and corrected calibration coefficients (gains) and biases, as well as their uncertainties”
Any other variables useful for GSICS product users?