I am using a untraviolet-visible spectrometer to measure Vertical Column Densities (VCDs) of trace gases up to the stratosphere. It is physically impossible to have negative VCDs. I have observed some of the measurements of slant column densities (SCDs) to be negative. mathematically, VCD = SCD/AMF. I was wondering whether AMF could be negative or the negative SCDs should be filtered out.
No, AMF is always positive. Depending on the retrieval algorithm, negative SCDs can occur as artifacts. You should look at your retrieval algorithm but in general I'd treat negative values in such a situation as "below detection threshold" values. Depending on what you are doing replacing them by 0 might be a valid course of action.