[论文解读] Analysis of the impact of broad absorption lines on quasar redshift measurements with synthetic observations

Luz Ángela García, Paul Martini+18 more|arXiv (Cornell University)|Apr 12, 2023

Galaxies: Formation, Evolution, Phenomena被引用 4

一句话总结

本研究利用合成的DESI类星体光谱，证明在类星体光谱中对宽吸收线（BALs）进行掩蔽可将红移测量误差降低约1%，并将灾难性红移误差减少约80%。作者主张将自动BAL掩蔽作为DESI等大规模光谱巡天中红移拟合的标准步骤。

ABSTRACT

Accurate quasar classifications and redshift measurements are increasingly important to precision cosmology experiments. Broad absorption line (BAL) features are present in 15-20\% of all quasars, and these features can introduce systematic redshift errors, and in extreme cases produce misclassifications. We quantitatively investigate the impact of BAL features on quasar classifications and redshift measurements with synthetic spectra that were designed to match observations by the Dark Energy Spectroscopic Instrument (DESI) survey. Over the course of five years, DESI aims to measure spectra for 40 million galaxies and quasars, including nearly three million quasars. Our synthetic quasar spectra match the signal-to-noise ratio and redshift distributions of the first year of DESI observations, and include the same synthetic quasar spectra both with and without BAL features. We demonstrate that masking the locations of the BAL features decreases the redshift errors by about 1\% and reduces the number of catastrophic redshift errors by about 80\%. We conclude that identifying and masking BAL troughs should be a standard part of the redshift determination step for DESI and other large-scale spectroscopic surveys of quasars.

研究动机与目标

• 量化宽吸收线（BALs）对大规模光谱巡天中类星体红移测量的影响。
• 研究BAL特征如何降低DESI巡天中红移拟合性能，特别是使用redrock软件时的表现。
• 评估掩蔽BAL区域是否能提高红移精度和分类可靠性。
• 评估在合成DESI数据中，掩蔽效果在不同信噪比和曝光时间下的表现。
• 建议将标准化BAL掩蔽作为未来巡天中类星体红移测定的最佳实践。

提出的方法

• 生成与DESI第一年观测预期信噪比和红移分布相匹配的合成类星体光谱。
• 创建成对的模拟光谱——除在C iv、Si iv、N v和Lyα波长处是否包含BAL特征外，其余完全相同。
• 对包含BAL和不包含BAL的光谱分别应用redrock红移拟合器，比较红移精度、分类成功率和警告标志。
• 实施一种掩蔽策略，在红移拟合前排除已知BAL深谷区域。
• 通过红移误差（dv）、ZWARNING标志出现率和分类成功率等指标衡量性能提升。
• 在不同曝光时间区间和红移范围内分析结果，以评估掩蔽在不同观测条件下的鲁棒性。

Figure 1: Examples of synthetic quasar spectra from DESI Y1 mocks as a function of the observed wavelength. The grey region corresponds to the total flux, whereas the dark blue line shows the smoothed flux. The vertical lines show the main lines used to identify the BAL features. In each panel, BAL

实验结果

研究问题

• RQ1 宽吸收线（BALs）如何影响DESI巡天中类星体光谱的红移测量精度？
• RQ2 BAL特征在多大程度上增加了灾难性红移误差和误分类率（例如，类星体被误认为星系或恒星）？
• RQ3 在光谱中掩蔽BAL区域是否能显著改善使用redrock软件的红移拟合性能？
• RQ4 在合成DESI数据中，掩蔽效果如何随信噪比和曝光时间变化？
• RQ5 与未掩蔽光谱相比，掩蔽BAL区域后红移误差和分类可靠性在数量上提升了多少？

主要发现

• 与未掩蔽光谱相比，掩蔽BAL区域可将红移误差降低约1%。
• 当掩蔽BAL深谷时，灾难性红移误差数量减少约80%。
• 掩蔽后，非BAL类星体的ZWARNING标志（指示红移拟合问题）出现率在基线之上降至0.2%。
• 分类成功率提高约1%，主要归因于更少的误识别和更少的警告标志。
• 掩蔽光谱的红移散射接近非BAL类星体，尤其在低红移时表现更佳。
• 即使经过掩蔽，极高水平的红移弥散（dv > 15,000 km/s）仍无法纠正，表明极端情况下仍存在持续挑战。

Figure 2: Redshift distributions of the 18555 simulated spectra that contain BALs. The histograms show the redshift from the truth file ( $z_{tr;noBAL}$ , black line ), the measurements from redrock ( $z_{rr;noBAL}$ ; dashed red lines ) for the quasars without the BAL templates; 3) the Ly $\alpha$ $