Open AccessRetrieval of cloud properties from near‐ultraviolet, visible, and near‐infrared satellite‐based Earth reflectivity spectra: A comparative study
Author(s) -
van Deelen R.,
Hasekamp O. P.,
van Diedenhoven B.,
Landgraf J.
Publication year - 2008
Publication title -
journal of geophysical research: atmospheres
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.67
H-Index - 298
eISSN - 2156-2202
pISSN - 0148-0227
DOI - 10.1029/2007jd009129
Subject(s) - remote sensing , ultraviolet , cloud computing , satellite , infrared , wavelength , trace gas , cloud fraction , spectral line , materials science , optics , environmental science , physics , cloud cover , geology , computer science , meteorology , astronomy , operating system
In this study, we investigate the capability to retrieve cloud parameters from near‐ultraviolet, visible, and near‐infrared satellite‐based reflectivity measurements. These parameters are essential to enable trace gas retrievals for cloud‐contaminated satellite scenes. We compare the retrieval of cloud top pressure, cloud fraction, and cloud optical thickness from simulated reflectivity measurements in three wavelength ranges: (1) 350–400 nm, which includes pronounced Ring effect structures; (2) 460–490 nm; and (3) 755–775 nm, which contain absorption features of O 2 ‐O 2 and O 2 , respectively. Retrieval simulations are performed for both a typical noise level of present‐day space‐borne spectrometers and additional random‐like measurement biases. Furthermore, we investigated the importance of the spectral continuum for the retrieval of cloud properties. It is found that reflectivity measurements in the wavelength ranges 350–400 and 755–775 nm provide complementary information on cloud properties. Both spectral windows provide more information on clouds than the 460–490 nm window. The best results are obtained for the combination of the 350–400 and 755–775 nm windows. In this case all three cloud parameters can be retrieved independently, and a high robustness is obtained with respect to random‐like measurement biases. Here it is not required to resolve the Ring effect structures in the near‐ultraviolet window. For this combination of spectral windows the error on retrieved NO 2 columns is reduced considerably.