Algorithm Input and Output Data Definition (IODD)

Input data

The input to the retrieval is using the L1B data product from the CIMR instrument, which includes the brightness temperatures of the channels 1.4, 6.9, 10.7, 18.7, and 36.5 GHz and their uncertainties. Technically, missing values are allowed, in case of malfunctioning channels, but the retrieval uncertainties will be larger. In addition, for a better constrain of the solution space, ECMWF analysis data are highly recommended as additional input (see Auxiliary data below). The \(\mathbf{y}\) and \(\mathbf{S}_e\) from (1) are set with the brightness temperatures and their uncertainties from the L1B data product. The input data for the multi-parameter retrieval includes all channels at horizontal and vertical polarization and their uncertainties coming from the L1b processor. In addition, time and location is required for the resampling of the variables. The algorithm is then applied on the C-band footprints, with all other channels being resampled to it. The uncertainties are assumed Gaussian in the retrieval.

Field	Description	Shape/Amount
L1B TB	L1B Brightness Temperature of all channels (both H and V polarization)	full swath or section of it (Nscans, Npos)
L1B NeΔT	Random radiometric uncertainty of all channels	full swath or section of it (Nscans, Npos)

Output data

The output data will include the retrieved geophysical parameters listed in equation (2) and their posterior uncertainties. In addition, quality flags are derived for each quantity and the retrieval procedure in general. The parameters are are: WS, TWV, CLW, SST, IST, SIC, MYI and SIT, SSS, as well as time and location for the resampling. The uncertainties are assumed Gaussian in the retrieval and are provided along with the parameters. The complete list of all output variables output in the EASE2 grid for each hemisphere is given in the table below.

Abbreviation	CF variable name	Description	Shape/Amount
WS	wind_speed	Wind speed	(1440,1440)
WSerr	wind_speed_error	Wind speed error	(1440,1440)
TWV	total_water_vapour	Total water vapour	(1440,1440)
TWVerr	total_water_vapour_error	Total water vapour error	(1440,1440)
CLW	cloud_liquid_water	Cloud liquid water	(1440,1440)
CLWerr	cloud_liquid_water_error	Cloud liquid water error	(1440,1440)
SST	sea_surface_temperature	Sea surface temperature	(1440,1440)
SSTerr	sea_surface_temperature_error	Sea surface temperature error	(1440,1440)
IST	ice_surface_temperature	Ice surface temperature	(1440,1440)
ISTerr	ice_surface_temperature_error	Ice surface temperature error	(1440,1440)
SIC	sea_ice_concentration	Sea ice concentration	(1440,1440)
SICerr	sea_ice_concentration_error	Sea ice concentration error	(1440,1440)
MYI	multi_year_ice_fraction	Multi-year ice fraction	(1440,1440)
MYIerr	multi_year_ice_fraction_error	Multi-year ice fraction error	(1440,1440)
SIT	sea_ice_thickness	Sea ice thickness	(1440,1440)
SITerr	sea_ice_thickness_error	Sea ice thickness error	(1440,1440)
SSS	sea_surface_salinity	Sea surface salinity	(1440,1440)
SSSerr	sea_surface_salinity_error	Sea surface salinity error	(1440,1440)

Auxiliary data

ECMWF surface analysis data is used as background values for the retrieval. The variables used are WS, TWV, CLW, T2M, TSK. They are used to fill the \(\mathbf{S}_a\) matrix and the \(\mathbf{x}_a\) in equation (1). For near real-time retrieval, the \(\mathbf{S}_a\) and \(\mathbf{x}_a\) can use monthly or seasonal values, as the retrieval is not sensitive to the exact values of the background variables with the rather wide covariance matrix.

Note

In this document a retrieval with fixed background values and TB error is used (execpt in First assessment of the algorithm). This may alter the results of the retrieval on the test cards compared to actual data where the background values are from ECMWF analysis data.