2016-08-05: New paper published in AMT: How big is an OMI pixel?
By averaging the collocated MODIS reflectances over the FoV of OMI in different ways, the optimal 'footprint' of OMI was determined. The results confirmed the lab tests that were performed by our department before the launch of OMI. The shape of the OMI FoV is not quandrangular, but rather super-Gaussian shaped and overlapping with neighbouring pixels. Furthermore, the shape is dependent on viewing angle and on the scene, because the measurements are slightly sensitive to polarisation.
The correlation of OMI reflectances with collocated MODIS reflectances revealed interesting dependencies, especially with respect to clouds. The time difference between Aura (with OMI) and Aqua (with MODIS) is only in the order of 8 minutes, but this is enough to significantly change the reflectance of a (partly) clouded scene. For partly clouded scenes the optimal overlap function can be signicantly distorted from the true FoV, because of clouds moving in and out of the scene. The correlation decreases significantly, while the shape changes: the middle of the scene has a much higher correlation than the edges, which become blurred. This knowledge is important for interpreting cloud masks from high-spatial-resolution imagers for instruments flying on a different platform, like the one planned for TROPOMI from VIIRS.
Read the full paper here