About atmospheric remote sensors

[DielsAlder]

Hi

I am currently involved in atmospheric remote sensors on board a satellite which collect back-scattered and reflected solar radiation to provide atmospheric composition measurements through an algorithm based on Differential Spectroscoy. I am currently involved in studying the physical principles behind the operation principle and retrieval step for these instruments. Examples are TROPOMI on board the Sentinel-5P and OMI (Aura). I would like to clarify some general aspects through some questions regarding their viewing. I would like to post an exercise and I would appreciate it very much if you could assess if my conclusions are correct . Thanks very much.

 

[jim mcnamara]

It is okay to post some data. I'm not sure by what you mean by 'viewing', I am guessing analysis and interpretation. If need be, we can move your question into the forum best suited for it.

 

[DielsAlder]

Thank you. Probably the term "viewing" is not properly used in my . I meant the swath viewed on Earth by the satellite so you are right and I was talking as well about analysis and interpretation of the collected solar radiation during the operation. It will be better explained if I make the statement of the problem, which is as follows:

Suppose that we have a linear array sensor (like GOME-2 for instance) whose assembly (spectrometer, telescope, ports and scanning mirror) is designed to fly oriented into the flight direction, meaning that through a downward looking that sensor scans a swath in the Earth surface perpendicular to the flight direction. It can be seen in the following figure

All past and current sensors operate in that Nadir looking mode (some of them also in limb or occultation mode)
However there is a change on the instrument accomodation and as a result it is oriented with a deviation of 30 degrees with respect to the flight direction:

The sensor is actually pointing downward with an angle of 30 degrees offset from the flight direction. It is not measuring a swath perpendicular to the flight direction but an inclined one as a result of the 30 degrees angle offset from that direction

My questions are related to the impact of that deviation on instrument performance in comparison with the former orientation into the flight direction when the rest of parameters (satellite and orbit) remain unchanged. In other words how crucial is that orientation regarding that it is unprecedented if we check all past and ongoing sensor missions

1) I do not have clear if there is an impact on the coverage
2) I assume that the ground pixel size depends on its distance along swath (is that correct in linear array sensors?) so the highest spacially-resolved measuremets of the instrument (just below the sensor) do not coincide with pixels imaged just below the satellite (as it would be if the sensor is accommodated oriented into the flight direction). Therefore if satellite orbit is selected as if the sensor is accommodated oriented into the flight direction, resolution will be smaller with the deviated accommodation
3) Regarding observation and calibration, Sun-instrument geometry depends on the selected orbit but also on the instrument accommodation. Sunlight capture will be impacted.
4) As long as Solar Zenith Angle (SZA) varies along the inclined swath, algorithms and radiative transfer model are impacted. I do not have clear if vertical profiles could be achieved as if the sensor is accommodated without the deviation. Algorithms (based on Differential Spectroscopy) should include a correction including the deviation, that is, it must accounted for by on-ground software processing. However I do not know if it is feasible. In other words, since experimental measurements must be corrected by onground modelling, it raises the question of whether experimental data from satellite measurements really worth it. The deviation may lead to a significant uncertainty in the gas retrieval step.

There may be more parameters which should be considered (I am not an expert at all in this field) so I would appreciate it very much if you have more ideas about this problem or suggest any correction in the above statements.

 

[Tom.G]

This from a decidedly non-expert! Given that the sensor is nadir-pointing and you are imaging a sphere, any pixel-size correction would remain the same across the field-of-view regardless of the sensor-orbit orientation. The only difference would be a rotation of parallax correction, if any is used. And of course any ground-truth locations would also need the parallax rotation correction.

Cheers,
Tom