Quantifying the relationship between the measurement precision and specifications of a UV/visible sensor on a geostationary satellite

Abstract

To investigate the feasibility of new satellite observations, including air quality (AQ) observations from geostationary (GEO) orbit, it is essential to link the measurement precision (ε) with sensor specifications in advance. The present study attempts to formulate the linkage between ε and specifications of a UV/visible sensor (signal-to-noise ratio (SNR), full width at half maximum (FWHM) of the slit function, and sampling ratio (SR)) on a GEO satellite. A sophisticated radiative transfer model (JACOSPAR) is used to calculate synthetic radiance spectra that would be measured by a UV/visible sensor observing the atmosphere over Tokyo (35.7°N, 139.7°E) from GEO orbit at 120°E longitude. The spectra, modified according to given sensor specifications, are analyzed by the differential optical absorption spectroscopy technique to estimate the ε for slant column densities of O3 and NO2. We find clear relationships: for example, the ε of the O3 slant column density (molecules cm−2) and SNR at 330nm are linked by the equation log(ε)=−1.06·log(SNR)+20.71 in the UV region, and the ε of the NO2 slant column density and SNR at 450nm are linked by log(ε)=−0.98·log(SNR)+18.00, at a FWHM=0.6nm (for the Gaussian slit function) and SR=4. The relationships are mostly independent of other specifications (e.g., horizontal and temporal resolutions), as they affect ε primarily through SNR, providing constraints in determining the optimal SNR (and alternatively FWHM and SR) for similar UV/visible sensors dedicated for AQ studies.