Viking UV cameras: Calibration using model calculations and long-wavelength light leakage

Abstract

In this paper we present the results of a comparison of Viking UV images of the dayglow with model calculations of the dayglow in order to assess the uncertainties in the calibration of the cameras. The procedure that was adopted was to calculate the dayglow emissions using proven photoelectron and radiative transfer codes, and convolve these results with the prelaunch camera transfer functions to estimate the emission strength for the two cameras (1304 Å (1 Å = 10−10 m) and Lyman–Birge–Hopfield (LBH)). We found quite good agreement between the model and measured intensities for the 1304 Å camera. However, the results for the LBH camera suggest that there is substantial leakage of far UV (FUV) light at 2000 Å and "white" light around 3400 Å: only about 20–30% of the emission comes from the LBH dayglow. To evaluate the contamination from the longer wavelengths, we assumed that it was due to Rayleigh scattering off air molecules and reflection off surfaces. A multilayer multiple-scattering radiative transfer code allowing for Rayleigh scattering and absorption by O2 and O3 was used. The model intensities were then convolved with "known" and extrapolated camera filter response characteristics in the region from 1800 to 7500 Å to obtain model camera signal values due to long-wavelength light leakage. A preliminary estimate of the scattering contribution to the observed intensities using the first-order guess of the camera response function (CRF) beyond 2000 Å resulted in a substantial overestimation of the LBH camera intensities. To obtain a better estimate of the CRF we used an LBH image where clouds were known to be present. This allowed us to distinguish between the 2000 and 3400 Å contributions in order to obtain a modified CRF that reproduces the data for the LBH camera. The fact that both the 1304 Å dayglow camera intensities and the model results are comparable and the LBH intensities may be modelled allowing for Rayleigh scattering suggest that the calibration of the cameras in the FUV are defined to better than about 30% and so should reproduce nighttime auroral intensities at 1304 Å and for the LBH system with comparable accuracy.