Earthdata Forum

I have been working with TEMPO L3 subsetted files for NO2 and HCHO (last downloaded from NASA EarthData Search around 8/4/2024), and used xarray in python to open them and access the data. When I open and plot the data, some of the data is negative (I believe the units are molecules/cm^2, so negatives don't make sense to me). The negative data is on similar orders of magnitude to the rest of the data (~1e14). I've attached a screenshot of part of the opened array below. Is this indicative of those data cells being missing, or how do I go about interpreting/handling this data?

@shrinidhin9

Thank you for contacting us about your interest in TEMPO data.

A Subject Matter Expert has been notified of your question and will reply shortly.

Warm Regards

Most negative values in TEMPO observations are normal and expected. In some cases, these may result from nonphysical retrievals, but these nonphysical observations will usually be flagged as “bad” or “suspect” in the main data quality flag. In fact, when calculating averages, users should not exclude negative values flagged as "good" or they may risk causing a high bias in the background level of the gas of interest.

TEMPO trace gases are retrieved from the ratio of two spectra (ratio of spectrum of the Earth radiance to a reference spectrum, which can either be a direct Sun measurement or a radiance reference). These spectra are collected with CCD detectors which have a contribution of signal from noise. If there is no gas of interest in the measurement (i.e., a “clean” atmosphere), we would expect the distribution of multiple retrieved measurements to be distributed around zero. In reality, there is always a bit of background in the gases that we look at (like NO2 and HCHO) so the distribution will be shifted in the positive direction, but some negative values may remain. If we look at a gas like total O3 which has a very large signal in the spectra and a high amount in the background, this is less likely to happen.