Accuracy Analysis of the On-board Data Reduction Pipeline for the Polarimetric and Helioseismic Imager on the Solar Orbiter Mission

Context: Scientific data reduction on-board deep space missions is a powerful approach to maximise science return, in the absence of wide telemetry bandwidths. The Polarimetric and Helioseismic Imager (PHI) on-board the Solar Orbiter (SO) is the first solar spectropolarimeter that opted for this solution, and provides the scientific community with science-ready data directly from orbit. This is the first instance of full solar spectropolarimetric data reduction on a spacecraft. Methods: In this paper, we analyse the accuracy achieved by the on-board data reduction, which is determined by the trade-offs taken to reduce computational demands and ensure autonomous operation of the instrument during the data reduction process. We look at the magnitude and nature of errors introduced in the different pipeline steps of the processing. We use an MHD sunspot simulation to isolate the data processing from other sources of inaccuracy. We process the data set with calibration data obtained from SO/PHI in orbit, and compare results calculated on a representative SO/PHI model on ground with a reference implementation of the same pipeline, without the on-board processing trade-offs. Results: Our investigation shows that the accuracy in the determination of the Stokes vectors, achieved by the data processing, is at least two orders of magnitude better than what the instrument was designed to achieve as final accuracy. Therefore, the data accuracy and the polarimetric sensitivity are not compromised by the on-board data processing. Furthermore, we also found that the errors in the physical parameters are within the numerical accuracy of typical RTE inversions with a Milne-Eddington approximation of the atmosphere. Conclusion: This paper demonstrates that the on-board data reduction of the data from SO/PHI does not compromise the accuracy of the processing. This places on-board data processing as a viable alternative for future scientific instruments that would need more telemetry than many missions are able to provide, in particular those in deep space.