About this ArticleT. Fischer (tfischer[at]stsci.edu) and B. James (bjames[at]stsci.edu)
The Space Telescope Science Institute's Spring 2022 Science Workshop, ''Large-Volume Spectroscopic Analyses of AGN and Star-Forming Galaxies in the Era of JWST,'' was dedicated to discussing the tools that astronomers currently use to analyze large numbers of complex spectroscopic data in active galactic nuclei (AGN) and star-forming galaxies. The workshop was held virtually across three platforms; BlueJeans (talks), Gather.town (posters/social interactions), and Slack (questions and discussions). The focus of the workshop was to:
- Inform users on spectroscopic tools that currently exist and have them avoid 'reinventing the wheel'
- Discuss ways to apply these tools for analysis of users' own data
- Aid users in scaling the sometimes daunting initial learning curve when employing these tools
- Determine major issues that still exist in spectroscopic analysis
The abstract selection process was completely anonymous to avoid any potential biases the reviewers might have had. The submissions were processed through a Google form and abstracts were randomly assigned to members of the Scientific Organizing Committee (SOC) for reviewing. This resulted in a diverse program with a gender balance that was representative of the submitted abstracts.
The workshop consisted of four days, with six hours each day of morning talks and afternoon workshop sessions. There were a total of 31 talks: 7 invited (20 minutes each, including questions), 24 contributed (15 minutes each, including questions), and 35 posters. Each day had a poster session before and after the talks to facilitate participants across multiple time zones. Similar to previous virtual conferences held by the Institute, poster sessions were held in Gather.town, an online platform that allows you to create an avatar and navigate a virtual environment designed to imitate the spontaneous interactions of an in-person conference. The virtual conference space was set up to resemble the Muller building, with three poster rooms scattered throughout the virtual building.
Each morning was formatted such that we would hear from two Invited Speakers on the background of the tool they had a hand in developing, along with six contributed talks that pertained to additional spectroscopic tools and/or AGN or star-forming galaxy research. Each afternoon included lightning talks from poster presenters (60s per poster) and workshop sessions with the days' Invited Speakers (50-minute sessions), where they were asked to walk though how to use their tool, from initial dataset to analysis output.
At the end of the last day, we performed a data analysis 'rodeo.' Before the workshop, each Invited Speaker was given the same 30 x 30 spaxel Very Large Telescope/Multi Unit Spectroscopic Explorer (VLT/MUSE) data cube of nearby AGN NGC 1386 to fit using their spectroscopic analysis tool. The speakers were asked to fit H-alpha, H-beta, [OⅢ] λ5007 and [NⅡ] λ6584 emission lines, and provide their fit results to us prior to the beginning of the workshop. At the beginning of the rodeo, H-alpha emission-line flux, width, and centroid maps were presented for each tool and the affiliated Speaker was given the opportunity to discuss the trials and tribulations they experienced during their analysis. The largest differences between fits occurred where emission lines exhibited either A) multiple kinematic components or B) absorption features. As a result, we found that flux measurements could exhibit severe variations between tools. Several other properties were derived and compared from each tool, including metallicity maps (derived from the [NⅡ]/H-alpha flux ratio) and BPT diagnostic plots (derived from [NⅡ]/H-alpha vs [OⅢ]/H-beta). Again, non-negligible differences existed between these properties, mostly due to differences between the assumptions employed while fitting the cubes.
After discussing the rodeo results, we reflected on the workshop and what would be the best inclusions to tools for future users. We concluded that the following would be most useful:
- Standardized outputs. Clearly providing the key measurement parameters of integrated flux, line width, and line centroid are paramount to users quickly digesting analysis results. The example preferred during the discussion was .fits files that were easy to navigate (i.e., individual files for each parameter or clearly labeled extensions in a single file).
- Clearly stated modeling assumptions. Assumptions made during spectroscopic fitting should be made clear to the user at the beginning of the analysis. Each tool contained caveats within their fits that, if unknown to the user, could produce misleading results, as illustrated in the flux and galaxy property variations from the rodeo.
- Instructions and examples of how to use each tool. The participants noted that the recordings produced throughout the workshop of each Invited Speaker walking through using their tool are likely invaluable to the spectroscopic community. As such, websites for tools are encouraged to include such videos to provide first-hand information from the creators on how to implement their tools in spectroscopic analysis.
- Standardized stellar templates. Several fitting tools account for the underlying stellar continuum before emission line analysis. From our discussion, it appears that there is not a consensus on reliable/consistently employed stellar templates.
- Data Quality Checks. Users expressed that optional signal-to-noise cuts on individual fitting components would be useful in assessing the reliability of their line-fitting results. A clear understanding of how errors are accurately propagated from data cube to line-fit results would be essential in providing such cuts.
By the end of the conference, we were able to reach 275 participants across 37 countries.
As a final note, the chairs would like to personally thank all of the members of the SOC, the Event Planning Group, and IT staff at the Institute who made this workshop a huge success.