Publications

Luo, BK; Minnett, PJ; Gentemann, C; Szczodrak, G (2019). Improving satellite retrieved night-time infrared sea surface temperatures in aerosol contaminated regions. REMOTE SENSING OF ENVIRONMENT, 223, 8-20.

Abstract
Satellite retrievals of sea surface temperature (SST) have become necessary for many applications. Satellite infrared imaging radiometers passively measure the radiance emitted and reflected by the surface of the Earth and atmosphere. Tropospheric aerosols increase the signal attenuation at the satellite height, degrading the accuracy of SST retrievals. In this study to assess the infrared radiative effects of aerosols on satellite-derived skin SSTs (SSTskin), MODIS (MODerate-resolution Imaging Spectroradiometer) SSTskin retrievals are compared with quality-controlled, collocated SSTskin, measurements from the Marine-Atmospheric Emitted Radiance Interferometer (M-AERI) deployed on research cruises during the Aerosols and Ocean Science Expeditions (AEROSE), and sub-surface temperatures measured by thermistors on drifting buoys. In this region, SSTskin retrievals from the MODIS on the Aqua satellite are generally much cooler than the in-situ measurements. In the Saharan outflow area between 90 W to 90 degrees E and 20 degrees S to 35 degrees N where aerosol optical depths may be > 0.5, satellite SSTskin are often > 1 K cooler than the in situ data. The goal of this research is to determine the characteristics of aerosol effects on satellite retrieved infrared SST, and to derive formulae for improving accuracies of infrared-derived SSTs in aerosol-contaminated regions. A new method to derive a night-time Dust-induced SST Difference Index (DSDI) algorithm based on simulated brightness temperatures and Principal component (PC) analysis (PCA) at infrared wavelengths of 3.8, 8.9, 10.8 and 12.0 mu m, was developed using radiative transfer simulations. The satellite SSTskin biases and standard deviations, derive by comparisons with coincident and collocated surface and in situ measurements, are reduced by 0.263 K and 0.166 K after the DSDI correction. This method can also improve the fraction of useful data available compared to the usual approach of discarding measurements identified as being contaminated by the effects of aerosols.

DOI:
10.1016/j.rse.2019.01.009

ISSN:
0034-4257