Stratospheric Temperature Trends and Anomalies in Modern Satellite Data
Abstract
Space-based data records sensitive to temperature of the Earth's middle atmosphere are highly calibrated and are becoming long enough to be useful for detection changes to the structure of the atmosphere due to global warming. Two such records are those being obtained by high spectral resolution infrared sounders and by GPS radio occultation, both records commencing circa 2003. Proper interpretation of their measurements as middle atmospheric temperature trends requires that their retrieval algorithms not induce any time-dependent bias. Comparing the 15-year temperature record of the upper troposphere and lower stratosphere obtained by these two data types reveals time-dependent biases due to retrieval error, and the spatial signature bears out which retrieval system is at fault for every bias detected.
Comparing the temperature records obtained by the Atmospheric Infrared Sounder (AIRS) and the CHAMP and COSMIC radio occultation (RO) missions reveals several time-dependent biases in each of their retrieval systems. We have found that a retrieval bias is temperature trends is incurred in infrared remote sensing depending on whether or not collocated microwave data is included in retrieval. We also have found that AIRS temperature retrievals are susceptible to null-space error, are associated with the seasonal cycle in southern high latitudes and with the stratospheric quasi-biennial oscillation, and can be as large as 4 K. There are indications of a long-term time-dependent retrieval error of at least 0.1 K decade-1 in AIRS data. In RO data we have found a time-dependent retrieval error due to initialization of the RO retrieval integrals induced by differences in GPS receiver performance. We have also found that faulty prior information on upper tropospheric humidity can induce a large error in the annual cycle of upper tropospheric temperature as retrieved from RO data. Overall, uncertainty in temperature trends in the lower stratosphere is approximately 0.2 to 0.5 K decade-1 from 2003 through 2015. We anticipate that all of the time-dependent biases found in infrared retrieval and RO retrieval, with further research, will be corrected in future version of their respective retrieval algorithms.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.A51O2418L
- Keywords:
-
- 0305 Aerosols and particles;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 0322 Constituent sources and sinks;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 0368 Troposphere: constituent transport and chemistry;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 3311 Clouds and aerosols;
- ATMOSPHERIC PROCESSES