Meteorological satellite data rescue: Assessing radiances from Nimbus-IV IRIS (1970-1971) and Nimbus-VI HIRS (1975-1976)

This report presents an example of valorisation of two historical radiance datasets. In 1970 and 1971, the InfraRed Interferometer Spectrometer (IRIS) operated from the Nimbus-IV satellite. Even by today’s standards, this Michelson interferometer was a hyperspectral sounder, with 862 channels. It covered wavenumbers between 400 and 1600 cm−1(or 25.00–6.25 mm). Though without cross-scanning, this instrument predated by more than 30 years the current hyperspectral sounders such as AIRS on EOS-Aqua, IASI on Metop-A and -B, and CrIS on Suomi NPP.

The data collected by Nimbus-4 IRIS have so far not been used in global, multi-decadal, atmospheric reanalyses. Yet, these radiance data contain spectrally detailed information about our atmosphere’s vertical structure and its constituents. Also, owing to the nature of the calibration problem acting on measurements of narrow spectral intervals, these radiance data have a great potential to serve as stable references in an assimilation scheme during the time period when they are available, or can be used to assess the quality of other atmospheric datasets, once the IRIS data quality has been understood and characterized.

The data from the Nimbus-4 IRIS experiment have recently been rescued from tapes by the NASA Goddard Earth Sciences Data and Information Services Center (GES DISC).With the aid of a state-of-the-art radiative transfer model, the EUMETSAT NWP-SAF RTTOV, we revisit this radiance dataset by comparing it with ECMWF reanalyses (ERA-40 and ERA-20C).

After cloud detection, we find in some spectral regions (e.g., 15 microns band) that the fit to the ERA-40 and ERA-20C reanalyses is below 1 K standard deviation in brightness temperatures. However, there are residual biases around 1 K in the window region which require further investigation.

The other dataset evaluated are radiances collected by Nimbus-VI HIRS, between August 1975 and March 1976. After applying simple quality controls, we find that the temperature sounding channels in the 15 microns band present no unexpected cross-scan field-of-view dependence. The data record is too short for assimilation and additional data from this experiment should be recovered.

Last, the report proposes that the series of infra-red sounders and imagers back to 1970 could help compare the first (IRIS) and current (IASI, CrIS) interferometers, with the help of the simultaneous nadir overpass technique. The magnitude of the discrepancy between forward- and backward-propagation of calibration corrections would then help refine the confidence placed in long-term trends derived from infra-red radiance data. ERA