InfraRed Sounder A Spot

IRS retrievals and applications at high satellite zenith angle

InfraRed Sounder A Spot
InfraRed Sounder A Spot

This study compiled a large dataset of geophysical variables (Temperature (T), Water Vapour (W) and Ozone (O) profiles with associated surface information, representative of the Meteosat Third Generation Infrared Sounder (MTG-IRS) disc in all seasons) and engineered a method aimed at assessing the sensing sensitivity of MTG-IRS at high scanning angles.

Last Updated

16 February 2021

Published on

08 February 2021

This data set and associated analysis addresses the fact that the MTG-IRS measurements at the edge of the disc have high satellite zenith angles. In these cases, the atmosphere observed by the instrument follows a so-called slant path, which means that the geolocation of the field of view can be different from the geolocation of the atmosphere further up on the slant path. Furthermore, the sensitivity of the instrument to the atmosphere at varying heights is affected by the sensing geometry.


The purpose of this study was to characterise the expected impact of the sensing geometry on the averaging kernels and retrieval precision of profiles of T, W and O. This characterisation was based on cloud free simulations using RTTOV-12, for which MTG-IRS coefficients valid for zenith angles up to 85° were available.


The results shown in Figure 1 show that, on average, most of the O, W and T retrieval errors are relatively stable and constant when the satellite zenith angle remains below 60°. Some disparities are observed between tropical regions and mid-latitudes. However, these differences are essentially linked to the meteorological variations of the scenes considered. Indeed, in the tropics the atmosphere is more unstable, with convective phenomena being more strongly pronounced than at mid-latitudes. This stronger convection is associated with higher water vapour content and greater temperature gradients, leading to a higher risk of retrieval error. This is typically what was observed in the results. The ozone error is also correlated to geographic areas, whereby the largest errors were found in the polar region.

Mean retrieval error for temperature, water vapour and ozone w.r.t Satellite Zenith Angle
Figure 1: Mean retrieval error for temperature, water vapour and ozone w.r.t Satellite Zenith Angle.

Of particular note was the evaluation of retrieval errors of profile estimates (Figure 2). It was noted that the bottom of the retrieved profile (from the surface up to 500 hPa) is systematically of lower quality when the satellite zenith angle is very high (SatZA = 85°) compared to that obtained at nadir. On the other hand, the top of the temperature profile (pressure > 500 hPa) is systematically of better quality at a very high satellite zenith angle than at nadir.

Mean averaging kernels over tropical sea
Figure 2: Mean averaging kernels over tropical sea.

This indicates that at high satellite zenith angle, the IRS channels are not sensitive to surface properties. This would explain the larger errors found at the bottom of the profile and the fact that the top of the profile is of the same or higher quality as at nadir. Indeed, if the retrieval were less constrained by the surface, more freedom at the top of the profile would be left, potentially leading to a product more in agreement with channels strongly sensitive at high altitudes.

Further details of this study can be found in the Study Report below.