Uncertainty quantification in the Infrared surface Emissivity Model (ISEM)

Tanvir Islam, P. K. Srivastava, George Petropoulos

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)
128 Downloads (Pure)

Abstract

Accurate modeling of surface emissivity is imperative for accurate radiative transfer simulation and forward modeling of satellite radiance observations. The Radiative Transfer for (A)TOVS (RTTOV) fast radiative transfer model uses the Infrared Surface Emissivity Model (ISEM) for the computation of sea surface emissivity in the infrared (IR) spectrum. However, the model does not incorporate the effect of surface-emitted surface reflected (SESR) radiation and dependence of wind speed in the emissivity calculation. This paper investigates the uncertainty in the ISEM model caused by ignoring the SESR radiation and wind speed effects in the 3 IR bands, 3.7, 11, and 12 μm. First, we develop a new model called Surface Emissivity Model in IR with SESR (SEMIS) that takes the SESR radiation and wind speed effects into account. The uncertainty in the ISEM model is then quantified by comparing the ISEM emissivity against SEMIS derived emissivity. The comparison results suggest that two models are in excellent agreement below \sim! 60^\circ emission angle, implying no notable uncertainty in the ISEM model at smaller angles. Nevertheless, uncertainty tends to significantly increase with increasing emission angle above \sim !60^\circ , which is even more notable at high wind speed ( {\rm{\sim! 15, m/s}} ). Two models are further compared against emissivity measurements from a radiometer. The ISEM model has produced large errors as opposed to the SEMIS
Original languageEnglish
Pages (from-to)5888-5892
JournalIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Volume9
Issue number12
Early online date25 May 2016
DOIs
Publication statusPublished - 01 Dec 2016

Fingerprint

Dive into the research topics of 'Uncertainty quantification in the Infrared surface Emissivity Model (ISEM)'. Together they form a unique fingerprint.

Cite this