New aerosol robust sea surface temperature algorithms for the along‐track scanning radiometer

1997 
The along-track scanning radiometer (ATSR) was the first spaceborne radiometer to possess a dual view angle capability designed to improve the accuracy of the remote sensing of sea surface temperature (SST). We present the results of comparing SST retrieved by using "single-look" and "dual-look" atmospheric correction algorithms applied to ATSR data collected around the British Isles from April to July 1992. We find that the retrievals using the single-look algorithms are consistently cooler (mean value 0.67 K) than the dual-look retrieved temperatures, the rms spread of the differences being 0.2 K. Potential causes of these differences could be either a previously predicted wind-induced (surface emissivity related) error for the dual-look algorithm or stratospheric volcanic aerosols, produced by the eruption of Mt. Pinatubo in June 1991, mainly affecting the single-look algorithm. To determine the dominant cause of the observed differences, we develop three new dual-look algorithms: one robust against surface emissivity variations, one against aerosol variations, and one against both of these effects. Comparing the results yielded by these algorithms indicates that the retrieved SST differences are due to aerosol variations affecting the single-look retrievals and that the surface emissivity effects on the dual-look algorithm are less than 0.1 K. We use this result to develop global dual-look ATSR SST algorithms that are very robust against aerosols to a precision of 0.08 K under a wide range of aerosol loadings. The results are particularly important for the early part of the ATSR mission, when the current dual-look SST retrievals could be affected by Pinatubo aerosols. Furthermore, the high spatial and temporal variations in stratospheric aerosols seen in our study indicate that such algorithms are essential for long-term bias-free global SST monitoring.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    31
    References
    21
    Citations
    NaN
    KQI
    []