Algorithm Stability and the Long-Term Geospace Data Record From TIMED/SABER

The ability of satellite instruments to accurately observe long-term changes in atmospheric temperature depends on many factors including the absolute accuracy of the measurement, the stability of the calibration of the instrument, the stability of the satellite orbit, and the stability of the numerical algorithm that produces the temperature data. We present an example of algorithm instability recently discovered in the temperature data set from the SABER instrument on the NASA TIMED satellite. The instability resulted in derived temperatures that were substantially colder than anticipated from mid-December 2019 to mid-2022. This algorithm-induced change in temperature over 1 to 2 years corresponded to the expected temperature change over several decades from increasing anthropogenic CO2. This paper highlights the importance of algorithm stability in developing Geospace Data Records (GDRs) for Earth's mesosphere and lower thermosphere. A corrected version (Version 2.08) of the temperatures from SABER is described.

Automated summary

The accuracy of long-term atmospheric temperature observations by satellite instruments relies on various factors, such as measurement accuracy, instrument calibration stability, satellite orbit stability, and numerical algorithm stability. A recent example of algorithm instability in the temperature data from the SABER instrument on the NASA TIMED satellite is presented. This instability resulted in significantly colder temperatures than anticipated for a period from mid-December 2019 to mid-2022. The paper emphasizes the importance of algorithm stability in developing Geospace Data Records (GDRs) for Earth's mesosphere and lower thermosphere, and introduces a corrected version (Version 2.08) of the SABER temperatures.