A review of differential absorption algorithms utilized at NOAA for measuring sea surface temperature with satellite radiometers

[摘要] In a 1988 study, a generalized non-linear SST algorithm was derived from the radiative transfer equation. An empirical approximation to this algorithm provided the simplified NLSST algorithm for computing sea surface temperatures with satellite radiometer data. It has remained a favorite of the remote sensing community for the past 25 years. However, a 1998 study demonstrated that the functional form of the generalized non-linear algorithm, which is here designated the GNLSST, is very much data dependent. In this paper we explore the various functional forms that the GNLSST may take. Rather than using actual satellite data, we model the effects of atmospheric absorption by water vapor, under saturation conditions, on the satellite split-window temperature difference, AT, in the 11-13 pm infra-red spectral region. The parameters used in the GNLSST algorithm are adjusted to demonstrate the various forms that the GNLSST can take. One key finding is that the familiar algorithms used for measuring SST, such as the MCSST, the NLSST and the QSST are all special solutions to the more general GNLSST algorithm. Additionally, the GNLSST may take other forms, which are described graphically. The simplifying assumptions which yield the NLSST solution represent an approximation to reality when working with actual satellite data. It is demonstrated that minor modifications to the regression procedures, which are used to derive the NLSST with satellite split-window data, may reduce the temperature errors resulting from these approximations. Alternative forms of the GNLSST algorithm may be useful for users of regional split-window satellite data. (C) 2016 Elsevier Inc. All rights reserved.

[发布日期] 2016-12-15

[关键词] Sea surface temperature;MCSST;QSST;NLSST;CPSST;GNLSST;AVHRR;NOAA;Radiative transfer theory;Water vapor saturation