I have a direct-sun spectrometer that I am using to measure the concentration of a trace gas in the atmosphere. The basis of deriving this concentration is the Lambert-Beer's law. Since my instrument cannot be placed in outer space to collect a blank spectra so that the measured spectra may be compared to calculate the concentration of the respective trace gas, a so-called "reference spectrum" is collected, which may be a single spectrum, although, almost always, it is the mean spectrum of multiple spectrums collected over clear-sky conditions at the zenith-looking geometry, which is done to avoid as much contamination by spectral stray light from Rayleigh scattering, and Mie scattering, as possible. The reference spectrum is a spectrum that is associated with the concentration of the trace gas that is as close as possible to the least count of the spectrometer. This is done in order to generate the least number of negative vertical column densities.
My instrument is giving me a lot of negative vertical column densities for a particular trace gas for which my instrument uses a "standard amount" of 0.5 DU.
I have two questions.
(1) Is the standard amount of a trace gas based on its reference spectrum?
(2) How should I correct the negative values of vertical column densities?