Articles | Volume 9, issue 12
Atmos. Meas. Tech., 9, 6013–6023, 2016
Atmos. Meas. Tech., 9, 6013–6023, 2016

Research article 14 Dec 2016

Research article | 14 Dec 2016

Deriving clear-sky longwave spectral flux from spaceborne hyperspectral radiance measurements: a case study with AIRS observations

Xiuhong Chen and Xianglei Huang Xiuhong Chen and Xianglei Huang
  • Department of Climate and Space Sciences and Engineering, University of Michigan, Ann Arbor, MI 48109, USA

Abstract. Previous studies have shown that longwave (LW) spectral fluxes have unique merit in climate studies. Using Atmospheric Infrared Sounder (AIRS) radiances as a case study, this study presents an algorithm to derive the entire LW clear-sky spectral fluxes from spaceborne hyperspectral observations. No other auxiliary observations are needed in the algorithm. A clear-sky scene is identified using a three-step detection method. The identified clear-sky scenes are then categorized into different sub-scene types using information about precipitable water, lapse rate and surface temperature inferred from the AIRS radiances at six selected channels. A previously established algorithm is then used to invert AIRS radiances to spectral fluxes over the entire LW spectrum at 10 cm−1 spectral interval. Accuracy of the algorithms is evaluated against collocated Clouds and the Earth's Radiant Energy System (CERES) observations. For nadir-view observations, the mean difference between outgoing longwave radiation (OLR) derived by this algorithm and the collocated CERES OLR is 1.52 Wm−2 with a standard deviation of 2.46 Wm−2. When the algorithm is extended for viewing zenith angle up to 45°, the performance is comparable to that for nadir-view results.

Short summary
We explore algorithms of estimating spectral flux over the entire longwave spectrum solely from hyperspectral radiance observations using AIRS data as an example. This is different from the traditional approach of estimating broadband flux from satellite observations in two ways: (1) no other remote sensing data sets are needed, and (2) the spectral details of the broadband flux can be derived. This study shows that the hyperspectral radiances can be used to directly obtain spectral flux.