Development of a nighttime shortwave radiative transfer model for remote T...

Wang, J., M. Zhou, X. Xu, S. Roudini, S. P. Sander, T. Pongetti, S. D. Miller, J. S. Reid, E. Hyer, and R. J. D. Spurr (2020), Development of a nighttime shortwave radiative transfer model for remote T sensing of nocturnal aerosols and fires from VIIRS, Remote Sensing of Environment, 241, 111727, doi:10.1016/j.rse.2020.111727.

The launch of the Visible Infrared Imaging Radiometer Suite (VIIRS) on board the Sumo-NPP satellite in 2011 ushered in a new era of using visible light and shortwave radiation at night to characterize aerosol and fire distributions from space. In order to exploit the full range of unprecedented observational capabilities of VIIRS, we have developed a nighttime shortwave radiative transfer model capability in the UNified and Linearized Radiative Transfer Model (UNL-VRTM). This capability is based on the use of additional source functions to treat illumination from the Moon, from fires, and from artificial lights. We have applied this model to address fundamental questions associated with the VIIRS sensing of aerosol and fire at night. Detailed description of model developments and validation (either directly with surface measurements of lunar spectra or indirectly through cross validation) are presented. Our analysis reveals that: (a) when convolution with the broad-range (500–900 nm) relative spectral response (RSR) function of the VIIRS Day-Night Band (DNB) is omitted, AOD retrieval from the DNB have uncertainties up to a factor of two in conditions with low or moderate AOD (<0.5 in mid-visible); (b) using a wavelength independent spectrum for the surface illumination source can lead to an AOD bias of −10% over surfaces illuminated by light-emitting diodes and fluorescent lamps, and −30% illuminated by high-pressure sodium lamps; and (c) a DNB-equivalent narrow band for AOD retrieval over the surfaces illuminated by the three types of bulbs studied in this paper is found to be centered at 585 nm at which the look-up table can be generated for AOD retrieval from DNB. Furthermore, while uncertainty in AOD retrievals from the DNB decreases as AOD increases, fire characterization can be affected by AOD; for a smokescenario AOD of 2.0, the DNB and SWIR (1.6 μm) radiances can be reduced by 50% depending on the fire area fraction and temperature within VIIRS pixel. DNB is overall more sensitive to smaller and cooler fires than SWIR and can be used to retrieve AOD over bright surfaces. Finally, three-dimensional (3D) radiative transfer effects and the non-collimated nature of most artificial light sources are neglected in this 1D radiative transfer (planeparallel) model, resulting in possibly large uncertainties (e.g., the inability to reproduce side-illumination of clouds by city lights) that should be studied in future.

PDF of Publication: 
Download from publisher's website.
Research Program: 
Atmospheric Composition
Atmospheric Composition Modeling and Analysis Program (ACMAP)
Radiation Science Program (RSP)
Tropospheric Composition Program (TCP)