Global carbon monoxide products from combined AIRS, TES and MLS measurements on A-train satellites

J X Warner,Z Sun,R Yang,Z Wei,W Lahoz,F Carminati,B Duncan,A Tangborn,L El Amraoui,J.-L Attié

doi:10.5194/acp-14-103-2014

Abstract

Abstract. This study tests a novel methodology to add value to satellite data sets. This methodology, data fusion, is similar to data assimilation, except that the background model-based field is replaced by a satellite data set, in this case AIRS (Atmospheric Infrared Sounder) carbon monoxide (CO) measurements. The observational information comes from CO measurements with lower spatial coverage than AIRS, namely, from TES (Tropospheric Emission Spectrometer) and MLS (Microwave Limb Sounder). We show that combining these data sets with data fusion uses the higher spectral resolution of TES to extend AIRS CO observational sensitivity to the lower troposphere, a region especially important for air quality studies. We also show that combined CO measurements from AIRS and MLS provide enhanced information in the UTLS (upper troposphere/lower stratosphere) region compared to each product individually. The combined AIRS–TES and AIRS–MLS CO products are validated against DACOM (differential absorption mid-IR diode laser spectrometer) in situ CO measurements from the INTEX-B (Intercontinental Chemical Transport Experiment: MILAGRO and Pacific phases) field campaign and in situ data from HIPPO (HIAPER Pole-to-Pole Observations) flights. The data fusion results show improved sensitivities in the lower and upper troposphere (20–30% and above 20%, respectively) as compared with AIRS-only version 5 CO retrievals, and improved daily coverage compared with TES and MLS CO data.

Highlights

Atmospheric carbon monoxide (CO) is simultaneously measured by three EOS (Earth Observing System) “A-train” satellite sensors: AIRS (Atmospheric Infrared Sounder) (Aumann et al, 2003) on Aqua, and TES (Tropospheric Emission Spectrometer) (Beer, 2006), and MLS (Microwave Limb Sounder) (Waters et al, 2006) on Aura
We show results from applying the data fusion methodology to AIRS and TES CO data
AIRS version 5 (V5) level 2 (L2) and TES V3 CO volume mixing ratios (VMRs) at 500 hPa for 4 March 2006 are shown in Fig. 4 to demonstrate the data coverage pattern, where the squares represent AIRS footprints and the filled circles represent TES pixels

Summary

Introduction

Atmospheric carbon monoxide (CO) is simultaneously measured by three EOS (Earth Observing System) “A-train” satellite sensors: AIRS (Atmospheric Infrared Sounder) (Aumann et al, 2003) on Aqua, and TES (Tropospheric Emission Spectrometer) (Beer, 2006), and MLS (Microwave Limb Sounder) (Waters et al, 2006) on Aura. Other products, such as TES or MLS, are added to the Kalman filter process as observations where the relative weighting is influenced by the observation error covariance from AIRS and TES or MLS retrievals. We show that in the vertical region (lower troposphere and the UTLS) where AIRS has low measurement sensitivity, the AIRS retrievals still provide the correct spatial variability (or spatial patterns), even when they cannot reproduce the correct CO magnitude This system does not require a model to constrain the physics of the geophysical fields, but rather uses AIRS routine measurements to constrain the spatial and temporal variability of the TES and MLS measurements. We used 1-D background error SDs as an approximation for this study, the extension to 2-D error fields and the inclusion of seasonal variations is possible not trivial, and will be the subject of a future study

Objectives

Results

Conclusion