Virtual Observatory for Astronomers: Where Are We Now?

After several years of intensive technological development Virtual Observatory resources have reached a level of maturity sufficient for their routine scientific exploitation. The Virtual Observatory is starting to be used by astronomers in a transparent way. In this article I will review several research projects making use of the VO at different levels of importance. I will present two projects going further than data mining: (1) studies of environmental effects on galaxy evolution, where VO resources and services are used in connection with dedicated observations using a large telescope and numerical simulations, and (2) a study of optical and near-infrared colours of nearby galaxies complemented by the spectroscopic data.

💡 Research Summary

The paper “Virtual Observatory for Astronomers: Where Are We Now?” provides a comprehensive overview of the current state of the Virtual Observatory (VO) and demonstrates how it has progressed from a data‑access framework to a platform capable of supporting full‑scale scientific investigations that combine data mining, dedicated observations, and numerical simulations. After a brief introduction that likens the VO to a “World‑Wide Web for astronomers,” the author outlines the suite of IVOA standards (VO Table, Resource Metadata, Spectrum Data Model, TAP, SIA/SSA, authentication mechanisms, etc.) and the ecosystem of client tools (TOPCAT, STILTS, AstroGrid VO Desktop, AstroGrid Runtime, VO‑enabled SExtractor services, etc.) that make heterogeneous multi‑wavelength data interoperable.

The paper then surveys the early scientific successes of the VO, such as the discovery of optically faint obscured quasars and AGN, and notes that most of these studies relied primarily on data discovery, access, and mining. The central question posed is whether the VO can now support research that goes beyond mining. To answer this, two case studies are presented in detail.

Case Study 1 – Compact Elliptical (cE) and Tidally Stripped Galaxies
The authors designed an automated workflow that uses VO services to identify candidate cE galaxies in HST WFPC2 images of galaxy clusters. The steps include: (1) querying VizieR for clusters with redshift z < 0.055; (2) retrieving precise coordinates and Galactic extinction from NED; (3) accessing reduced HST images via the Hubble Legacy Archive’s Simple Image Access Protocol (SIAP); (4) running a remote SExtractor service on the images without downloading them; (5) selecting compact, low‑ellipticity objects with effective radii < 0.7 kpc and high surface brightness; and (6) cross‑matching with NED to obtain redshifts and additional photometry. Applying this pipeline to the entire WFPC2 archive yielded dozens of candidates across 63 clusters. Follow‑up spectroscopy with the 6‑m BTA telescope (SCORPIO multi‑slit mode) provided high‑resolution spectra, which were analysed with PEGASE.HR stellar population models and the NBURSTS full‑spectral fitting technique. All eight observed candidates were confirmed as cluster members with old (≥ 8 Gyr) stellar populations and metallicities around Z⊙/2. Complementary N‑body simulations using GADGET‑2 demonstrated that tidal stripping of intermediate‑luminosity disc galaxies can produce remnants with the observed structural parameters, supporting the hypothesis that cE galaxies are the stripped bulges of larger progenitors. The authors also discuss practical VO limitations encountered (lack of a public NED VO service, private SIAP URLs, need for custom SExtractor configuration, and lengthy telescope proposal cycles).

Case Study 2 – Optical and Near‑Infrared Colours of Nearby Galaxies
In this project the authors combined SDSS DR7 optical photometry and spectra with UKIDSS DR4 LAS near‑infrared photometry to study colour‑age‑metallicity relations in nearby galaxies (0.03 < z < 0.3). Using SDSS CasJobs they retrieved ~170 000 spectroscopic galaxies, then cross‑matched them with UKIDSS within a 5″ radius via the AstroGrid VO Desktop (which handled authentication for the restricted UKIDSS service). TOPCAT/STILTS were employed for large‑table joins, and the VO‑enabled AstroGrid Desktop facilitated the retrieval of UKIDSS aperture magnitudes. Rest‑frame magnitudes were obtained by fitting the combined optical‑NIR spectral energy distributions with PEGASE.2 models to compute K‑corrections. Spectra were processed with the NBURSTS full‑spectral fitting code (run as a standalone, non‑VO service) to derive velocity dispersions, stellar ages, and metallicities within 3″ apertures. The authors discovered that using Petrosian magnitudes from the two surveys leads to systematic colour biases because of differing surface‑brightness sensitivities; they therefore adopted consistent 3″ aperture fluxes for both datasets. The final analysis yields robust colour‑stellar‑population correlations and demonstrates the feasibility of constructing homogeneous multi‑wavelength datasets entirely through VO tools, apart from the final spectral fitting step.

The paper concludes that the VO has matured to a level where it can support end‑to‑end scientific workflows that start with data discovery, proceed through targeted observations, and finish with theoretical modelling. However, the authors stress that true “VO‑enabled science” still requires non‑VO components (e.g., custom data‑reduction pipelines, proprietary simulation codes, and external fitting tools). To become a fully integrated e‑Science environment, future development should focus on exposing analysis and simulation services through standard VO protocols, improving interoperability with legacy archives such as NED, and simplifying the deployment of remote processing services. In summary, the Virtual Observatory is no longer a mere catalogue browser; it is an emerging research infrastructure that, when combined with traditional observational and theoretical resources, can significantly accelerate discovery in modern astrophysics.