Atmospheric Measurement Techniques (May 2021)

Optimal use of the Prede POM sky radiometer for aerosol, water vapor, and ozone retrievals

  • R. Kudo,
  • H. Diémoz,
  • V. Estellés,
  • V. Estellés,
  • M. Campanelli,
  • M. Momoi,
  • F. Marenco,
  • C. L. Ryder,
  • O. Ijima,
  • A. Uchiyama,
  • K. Nakashima,
  • A. Yamazaki,
  • R. Nagasawa,
  • N. Ohkawara,
  • H. Ishida

DOI
https://doi.org/10.5194/amt-14-3395-2021
Journal volume & issue
Vol. 14
pp. 3395 – 3426

Abstract

Read online

The Prede POM sky radiometer is a filter radiometer deployed worldwide in the SKYNET international network. A new method, called Skyrad pack MRI version 2 (MRI v2), is presented here to retrieve aerosol properties (size distribution, real and imaginary parts of the refractive index, single-scattering albedo, asymmetry factor, lidar ratio, and linear depolarization ratio), water vapor, and ozone column concentrations from the sky radiometer measurements. MRI v2 overcomes two limitations of previous methods (Skyrad pack versions 4.2 and 5, MRI version 1). One is the use of all the wavelengths of 315, 340, 380, 400, 500, 675, 870, 940, 1020, 1627, and 2200 nm if available from the sky radiometers, for example, in POM-02 models. The previous methods cannot use the wavelengths of 315, 940, 1627, and 2200 nm. This enables us to provide improved estimates of the aerosol optical properties, covering almost all the wavelengths of solar radiation. The other is the use of measurements in the principal plane geometry in addition to the solar almucantar plane geometry that is used in the previous versions. Measurements in the principal plane are regularly performed; however, they are currently not exploited despite being useful in the case of small solar zenith angles when the scattering angle distribution for almucantars becomes too small to yield useful information. Moreover, in the inversion algorithm, MRI v2 optimizes the smoothness constraints of the spectral dependencies of the refractive index and size distribution, and it changes the contribution of the diffuse radiances to the cost function according to the aerosol optical depth. This overcomes issues with the estimation of the size distribution and single-scattering albedo in the Skyrad pack version 4.2. The scattering model used here allows for non-spherical particles, improving results for mineral dust and permitting evaluation of the depolarization ratio. An assessment of the retrieval uncertainties using synthetic measurements shows that the best performance is obtained when the aerosol optical depth is larger than 0.2 at 500 nm. Improvements over the Skyrad pack versions 4.2 and 5 are obtained for the retrieved size distribution, imaginary part of the refractive index, single-scattering albedo, and lidar ratio at Tsukuba, Japan, while yielding comparable retrievals of the aerosol optical depth, real part of the refractive index, and asymmetry factor. A radiative closure study using surface solar irradiances from the Baseline Surface Radiation Network and the parameters retrieved from MRI v2 showed consistency, with a positive bias of the simulated global irradiance of about +1 %. Furthermore, the MRI v2 retrievals of the refractive index, single-scattering albedo, asymmetry factor, and size distribution have been found to be in agreement with integrated profiles of aircraft in situ measurements of two Saharan dust events at the Cape Verde archipelago during the Sunphotometer Airborne Validation Experiment in Dust (SAVEX-D) 2015 field campaign.