• Spectroscopy and Spectral Analysis
  • Vol. 40, Issue 8, 2332 (2020)
PEI Guo-chao1、2, LI Yuan3、4、*, and BAI Ting-zhu1、2
Author Affiliations
  • 1[in Chinese]
  • 2[in Chinese]
  • 3[in Chinese]
  • 4[in Chinese]
  • show less
    DOI: 10.3964/j.issn.1000-0593(2020)08-2332-07 Cite this Article
    PEI Guo-chao, LI Yuan, BAI Ting-zhu. Temporal Variation Model of Ultraviolet Hyperspectral Solar Reference Spectrum[J]. Spectroscopy and Spectral Analysis, 2020, 40(8): 2332 Copy Citation Text show less

    Abstract

    The solar reference spectrum is the basis for radiometric calibration and wavelength calibration in the field of on-board calibration. During the solar cycle, the amplitude of the solar irradiance spectrum fluctuates periodically, changing by more than 10% in the ultraviolet band. However, the on-board absolute calibration requirement is 3%. Temporal variation characteristics of the sun cannot be ignored. First of all, we studied the influence of calculation of MgII when given spectra, which are based on interpolation and convolution intervals, with different resolutions and sampling. After selecting appropriate interpolation and convolution interval, we model the periodic variation of the solar-based on the MgII of Climate spectrum and SORCE spectrum. Secondly, we select 80 sets (there are two dates for every set) of amplitude variation of MgII and irradiance at a different wavelength from August 21, 2003, to April 15, 2012. By first order fitting instead of ratio method, the first order fitting method decreases error from 1.167% to 0.125% for Climate spectrum, from 1.057% to 0.558% for SORCE spectrum, when compared to true irradiance data from August 21, 2003, to April 15, 2012. The results show that the first-order fitting method has higher inversion precision and prediction accuracy than the ratio method. Finally, based on time series of MgII we fulfil normalization of time and format of the spectrum. By eliminating large deviated data over the different spectrum, we got a solar reference spectrum with the resolution of 1 nm and sampling of 0.1 nm on June 25, 2008. In addition, the spectrum has an absolute error of 0.982% compared to the average of six spectra without eliminating deviated data in all wavelength. The spectrum can also be converted to any day using solar cycle activity model based on MgII and conversion factor. By promotion of resolution and sampling for this reference spectrum based on KNMI hyperspectral, as a result, we get the solar reference spectrum with the spectrum range of 250~500 nm, resolution of 0.1 nm and sampling of 0.01 nm, which can provide reference spectrum for on-orbit observation and on-board calibration of China’s FY-3 ultraviolet hyperspectral ozone detector (OMS).
    PEI Guo-chao, LI Yuan, BAI Ting-zhu. Temporal Variation Model of Ultraviolet Hyperspectral Solar Reference Spectrum[J]. Spectroscopy and Spectral Analysis, 2020, 40(8): 2332
    Download Citation