How to export of LandSat 8 radiance values correctly

LandSat 8 image opened in SNAP in Blue band shows the values of the spectral brightness, as shown in the figure (50-70 W/(m^2srµm)).


However, when I am trying to export this band by Band Extractor in .tiff as separate raster the spectral brightness values already have a range of 9000-11000. May I ask to explain why so happened and how I can save this band in the units that SNAP shows initially (W/(m^2sr*µm)?

There is a difference between using “Export View as Image” from , which gives RGB bands, and using “File/Export” from the main menu, which should store geophysical data. Geotiff files lack robust metadata support that is provided by BEAM-DIMAP and NetCDF4-CF. The latter supports compression and is now widely supported by GIS software and languages like R, Matlab, etc.

Dear George, thank you for your answer and some tips. I have tried to export (not from viewer) in the different ways: using Band extractor (from Raster menu) and from the main menu (“File/Export”), - the results were the same. In these variants I have choose for export the Blue and Green bands of LS8 as single.
Followed your tip, I have tried to export in *.nc format also and I have opened it in QGIS. The result was the same: the range of values which are visible in Pixel Info of SNAP (50-70 W/(m^2srµm) are not exported in the same range.
I am not so familiar with R and Mathlab, but I am confident user in ArcGIS and QGIS. And now I have the scientific task to use this soft for calculation by spectral brightness values of some bands of LS8, S2, S3 images.

If you load the NetCDF file in SNAP does it give the correct geophysical values? Band properties gives the data type as uint16, so there should be scale and offset to get the geophysical values. It isn’t always easy to find the values as they are handled transparently by the library (some RS data uses different values depending on the data, so there has been confusion when people try to use the same values for multiple images). See: Using the USGS Landsat Level-1 Data Product

Conversion to TOA Radiance
Landsat Level-1 data can be converted to TOA spectral radiance using the radiance rescaling factors in the MTL file: Lλ=MLQcal+AL where:
= TOA spectral radiance (Watts/( m2 * srad * μm))
ML =Band-specific multiplicative rescaling factor from the metadata (RADIANCE_MULT_BAND_x, where x is the band number)
AL=Band-specific additive rescaling factor from the metadata (RADIANCE_ADD_BAND_x, where x is the band number)
Qcal = Quantized and calibrated standard product pixel values (DN)

Dear George,
thank you very much for your help.

Can I ask where can I find the Qcal values in particular? I am converting DN to TOA radiance in SNAP using landsat images. I still don’t get the meaning of quantized and calibrated standard product pixels. Can you show me how to apply this in the Band Math expression editor?