Are the meteorological annotations (i.e., solar radiation, temperature, skin temperature, wind speed, latent heat, sensible heat, snow albedo) products of SENTINEL-3 SLSTR L1 RBT feasible to use for research?
Why some of them have five different _tx e.g., solar_radiation_tx1 to solar_radiation_tx5? And what “_tx” stand for?
What are the difference between them?
How can I make one single product (e.g., solar radiation) from these five?
It will be a great help to know these answers.
without having the details of S-3 processing present, I suspect these are not “products” (i.e. derived from the S-3 data,), but - as their name “annotations” suggest - auxiliary data taken from other sources (like weather models/reanalysis), provided in order to facilitate/easy analysis of the S3-3 data.
at least i’m fairly sure, one can not derive atmospheric temperature from S-3 with an accuracy that makes it a meaningful parameter (or, I would be be very surprised if one can). similar for wind speed (at other than at cloud top height).
ok, short google check brings up:
“The meteorological parameters data file contains ECMWF forecast or analysis fields, regridded onto tie points.” (from here, p37)
but no idea, what the different txN are for. maybe they refer to different models, the data is coming from. did you check and compare the contents of attributes of these variables? according to the documentation linked above, each *_tx variable e.g. contains a ‘model’ attribute. is this the same for tx1-tx5?
Are these really the names the bands have? E.g., solar_radiation_tx1 and not solar_radiation_tx_time_1_tx ? Because the latter is what I get when I open a SLSTR L1B product with the S3TBX.
Now, for solar_radiation the different numbers stand for different times, so the bands actually form a time series. The same is true for east_west_stress, latent_heat, north_south_stress, sensible_heat, thermal_radiation, u_wind and v_wind. For specific_humidity and temperature_profile, there exist 25 bands for different pressure levels. You can look up the time and pressure values in the metadata (variable attributes of met_tx for t_series and p_atmos, respectively).
The ending tx indicates that the data is on a tie-point grid, opposed to other data in the product that is on oblique or nadir view grids on 500- or 1000-m resolutions.
If you want to extract the bands to a single product you can create a subset (Raster->Subset). Either open the product with the 500- or 1000-m resolution reader or resample it to a single size, then in the subset dialog deselect all bands and all tie-point grids except for the ones you want.