On a limitation of Zeeman polarimetry and imperfect instrumentation in representing solar magnetic fields with weaker polarization signal

Alexei Pevtsov, Yang Liu, Ilpo Virtanen, Luca Bertello, К. Мурсула, K. D. Leka, Anna L. H. HughesJournal of Space Weather and Space Climate 2021 · DOI

TL;DR

This paper investigates a systematic discrepancy observed between two major solar vector magnetograph instruments (SOLIS/VSM and SDO/HMI) when measuring weaker magnetic fields. The authors propose that this disagreement arises from the inherent differences in how longitudinal and transverse magnetic field components are derived from Zeeman polarization signals, exacerbated by noise and unresolved magnetic structures, leading to potential misinterpretations of field orientation.

What Problem Does This Paper Address?

Understanding the Sun's magnetic field is crucial for space weather forecasting and understanding solar activity. While longitudinal magnetograms provide a basic measure, full-disk vector magnetograms offer a more complete picture of the magnetic field's orientation. However, observations of these vector fields, particularly in regions with weaker signals (like plage areas), have shown disagreements between different instruments. Specifically, the orientation of the transverse (East-West) and meridional (North-South) components can appear to have opposite signs. This paper aims to explain the physical reasons behind this discrepancy, which can impact our interpretation of solar magnetic field morphology and evolution.

Methods

The authors analyze data from two prominent full-disk vector magnetographs: the Vector Stokes Magnetograph (VSM) on the Synoptic Optical Long-term Investigation of the Sun (SOLIS) platform, and the Helioseismic and Magnetic Imager (HMI) on the Solar Dynamics Observatory (SDO). They compare vector magnetic field maps derived from these instruments, focusing on areas where disagreements are most pronounced.

The core of their investigation involves theoretical arguments and modeling to explain how noise and instrumental limitations affect the derivation of magnetic field components. They highlight the differing signal-to-noise ratios between longitudinal (line-of-sight) and transverse (in-plane) magnetic field measurements, with transverse fields being more susceptible to noise due to their quadratic dependence on linear polarization signals. They also consider the impact of the "filling factor" – the fraction of a pixel occupied by magnetized plasma – which is often assumed to be unity but can be less in reality, especially in weaker field regions.

Datasets

The paper utilizes observational data from:

  • SOLIS/VSM (Vector Stokes Magnetograph): Provides full-disk vector magnetograms.
  • SDO/HMI (Helioseismic and Magnetic Imager): Also provides full-disk vector magnetograms.

The authors specifically compare data from these instruments, focusing on observations taken around November 19, 2015, and February 10, 2015, to analyze discrepancies in active regions and plage areas.

Key Insights and Ideas

  • Differential Sensitivity to Noise: The longitudinal magnetic field (B||) is derived from circular polarization (Stokes V), while the transverse field (B⊥) is derived from linear polarization (Stokes Q and U). B|| has a linear relationship with Stokes V, while B⊥ has a quadratic relationship with Q and U. This means that for the same underlying field strength, the signal-to-noise ratio for B⊥ is significantly lower than for B||, making it more susceptible to noise.
  • Azimuthal Ambiguity and Noise: The 180° ambiguity in determining the transverse magnetic field's azimuth is a known challenge. When noise is present, especially in weaker field regions, it can systematically bias the chosen orientation. The authors show that noise added to the transverse field can lead to a systematic inclination of the derived vector field away from the solar disk center.
  • The Filling Factor Matters: The assumption of a unity filling factor (meaning the entire pixel is filled with magnetic field) is common. However, in regions with weaker fields or mixed magnetic and non-magnetic plasma, this assumption may not hold. The paper suggests that a non-unity filling factor can influence the sign of the derived transverse field components, potentially explaining sign reversals observed across the central meridian.
  • Disagreement in Weak Field Regions: The observed disagreements between SOLIS/VSM and SDO/HMI are most prominent in areas with weaker polarization signals, such as plage regions, and less so in sunspots where the magnetic field is stronger and the filling factor is likely closer to unity.

Key Results

  • The paper provides a theoretical framework and observational examples demonstrating how noise and the differential sensitivity of instruments to longitudinal versus transverse magnetic field components can lead to systematic errors in the derived vector magnetic field orientation.
  • Figure 1 visually illustrates disagreements in the meridional and zonal components between SDO/HMI and SOLIS/VSM for a specific active region.
  • The authors show through modeling (Figure 2) that noise added to the transverse field can cause a systematic tilt in the inferred field direction, and this effect can be significant (up to 20 degrees inclination error for a 200 G field with 20 G noise).
  • Figure 3 shows how a bipolar region's zonal field pattern changes as it moves across the solar disk, consistent with the proposed effect of noise and orientation.
  • Appendix A presents a test case with SOLIS/VSM data showing that when the magnetic filling factor is assumed to be less than unity, the transverse field components can reverse sign as the region crosses the central meridian, an effect that disappears when the filling factor is set to unity.

Limitations and Future Work

  • Unresolved Structures: The paper acknowledges that unresolved magnetic structures within pixels (non-unity filling factor) are a significant factor contributing to the observed discrepancies. Accurately modeling these mixed magnetic and non-magnetic contributions in pixels outside sunspots is extremely difficult.
  • Instrumental Differences: While the paper focuses on the physics of Zeeman polarimetry, subtle differences in instrumental characteristics and data processing pipelines between SOLIS/VSM and SDO/HMI could also play a role.
  • Correction Challenges: The systematic nature of the inclination error, stemming from noise and the inherent physics of polarization signals, makes it difficult to correct for without potentially introducing other errors or increasing noise in other components.
  • Future Work: The authors suggest that further investigation into spatially coupled inversion techniques, which consider the spatial distribution of magnetic and non-magnetic components within a pixel, might be necessary. Additionally, exploring methods to equalize noise levels between longitudinal and transverse field measurements, possibly through longer integration times, could be beneficial, though it might come at the cost of increased noise in the longitudinal field.

Summary updated

Abstract

Full disk vector magnetic fields are used widely for developing better understanding of large-scale structure, morphology, and patterns of the solar magnetic field. The data are also important for modeling various solar phenomena. However, observations of vector magnetic fields have one important limitation that may affect the determination of the true magnetic field orientation. This limitation stems from our ability to interpret the differing character of the Zeeman polarization signals which arise from the photospheric line-of-sight vs. the transverse components of the solar vector magnetic field, and is likely exacerbated by unresolved structure (non-unity fill fraction) as well as the disambiguation of the 180° degeneracy in the transverse-field azimuth. Here we provide a description of this phenomenon, and discuss issues, which require additional investigation.
BibTeX
@article{pevtsov2021limitation,
  author = {Alexei Pevtsov and Yang Liu and Ilpo Virtanen and Luca Bertello and К. Мурсула and K. D. Leka and Anna L. H. Hughes},
  title = {{On a limitation of Zeeman polarimetry and imperfect instrumentation in representing solar magnetic fields with weaker polarization signal}},
  journal = {Journal of Space Weather and Space Climate},
  year = {2021},
  doi = {10.1051/swsc/2021003},
}
Figure