"Advanced" data reduction for the AMBER instrument

TL;DR

This paper introduces an enhanced data reduction pipeline for the AMBER instrument that improves calibration accuracy by accounting for instrumental and atmospheric variations, resulting in more reliable data products.

Contribution

The authors developed a modular pipeline extension that incorporates additional calibration steps, addressing limitations of the standard amdlib software.

Findings

Improved calibration accuracy with correction of jitter and coherence loss.

Enhanced data quality assessment through overnight data structure analysis.

Production of fully calibrated AMBER data products with new pipeline features.

Abstract

The amdlib AMBER data reduction software is meant to produce AMBER data products from the raw data files that are sent to the PIs of different proposals or that can be found in the ESO data archive. The way defined by ESO to calibrate the data is to calibrate one science data file with a calibration one, observed as close in time as possible. Therefore, this scheme does not take into account instrumental drifts, atmospheric variations or visibility-loss corrections, in the current AMBER data processing software, amdlib. In this article, we present our approach to complement this default calibration scheme, to perform the final steps of data reduction, and to produce fully calibrated AMBER data products. These additional steps include: an overnight view of the data structure and data quality, the production of night transfer functions from the calibration stars observed during the night, the correction of additional effects not taken into account in the standard AMBER data reduction software such as the so-called "jitter" effect and the visibility spectral coherence loss, and finally, the production of fully calibrated data products. All these new features are beeing implemented in the modular pipeline script amdlibPipeline, written to complement the amdlib software.