We present what we believe to be a novel approach to simulating the spectral fine structure \left(<1\mathrm{nm}\right) in measurements of spectrometers such as the Global Ozone Monitoring Experiment (GOME). GOME measures the Earth's radiance spectra and daily solar irradiance spectra from which a reflectivity spectrum is commonly extracted. The high-frequency structures contained in such a spectrum are, apart from atmospheric absorption, caused by Raman scattering and by a shift between the solar irradiance and the Earth's radiance spectrum. Normally, an a priori high-resolution solar spectrum is used to simulate these structures. We present an alternative method in which all the required information on the solar spectrum is retrieved from the GOME measurements. We investigate two approaches for the spectral range of 390–400\mathrm{nm} . First, a solar spectrum is reconstructed on a fine spectral grid from the GOME solar measurement. This approach leads to undersampling errors of up to \mathrm{0.5\%} in the modeling of the Earth's radiance spectra. Second, a combination of the solar measurement and one of the Earth's radiance measurement is used to retrieve a solar spectrum. This approach effectively removes the undersampling error and results in residuals close to the GOME measurement noise of \mathrm{0.1\%} .