The sky spectra output files (Sky-###.fits) are given in units of counts per square pixel (on the CCD) divided by a scaled flat field image. It is possible to convert the sky spectrum to digital numbers by multiplying Sky-###.fits by the flat spectrum used (Flat-###.fits) and dividing by the scale factor. You can look in the screen output from the reduction (or the .log file) to see which flat field was used and what the scale factor was. You can also divide DN-###.fits by Flux-###.fits and then multiply the result by Sky-###.fits.

You can then get a rough estimate the true flux by dividing the the total DN at each wavelength by the total fractional Keck+ESI throughput (see ESI Observing ).

You can then convert to photons/second/cm^2/Angstrom/(square arcsecond), given the exposure time, a 10 meter diameter aperture, the Angstroms per pixel, the electrons per digital number (0.5 for ESI high gain) and the area of sky light in each spectrum pixel in square arcseconds (this is the width of a pixel (about 0.15") times the width of the slit in arcseconds).