Accurate photometric calibration by fitting the system transmission

TL;DR

This paper introduces a new photometric calibration method that simultaneously accounts for instrumental and atmospheric effects by fitting system transmission per image, achieving sub-1% accuracy using Gaia spectra and data from LAST.

Contribution

The paper presents a novel approach to absolute photometric calibration that improves accuracy by fitting system transmission on an image-by-image basis, leveraging Gaia spectra.

Findings

Residuals between observations and synthetic photometry are <1% per image.

Zero-point accuracy is between 3-5 mmag across images.

Method shows high stability over long timescales.

Abstract

Transforming the instrumental photometry of ground-based telescopes into a calibrated physical flux in a well-defined passband is a major challenge in astronomy. Along with the intrinsic instrumental difference between telescopes sharing the same filter, the effective transmission is continuously modified by the effects of the variable atmosphere of the Earth. We have developed a new approach to the absolute photometric calibration that simultaneously treats instrumental and atmospheric effects on an image-by-image basis by fitting the system transmission. This approach aims at breaking the 1% absolute photometric accuracy which limits current calibration methods for ground-based observatories. We fit the transmission, as a function of wavelength, for each image. The fit is done by comparing the instrumental fluxes of stars in the image to the synthetic photometry of the stars given their spectrum and the transmission function which have free parameters. A key element that enables this approach is the set of about 220 million low-resolution spectra measured by Gaia, which provides a large number of stellar calibrators in the image. We demonstrate the method using data from the Large Array Survey Telescope (LAST). We show that the residuals between observations and synthetic photometry of the Gaia spectra in the fitted transmission have a standard deviation $<$1% on an image-by-image basis, with no spatial and color dependencies. The median accuracy of the zero-point throughout the image is between 3-5 mmag, depending on the total image exposure. Furthermore we show that this method provides high stability over long temporal scales.