Unlocking Galaxy and AGN Co-evolution via SKA Radio Continuum Surveys
By the peak of cosmic star formation, roughly 85% of all star-forming activity was hidden behind thick clouds of cosmic dust. This dust acts like a heavy fog, absorbing ultraviolet and optical light. Because of this, traditional telescopes struggle to see the most energetic processes in the early universe. A new study outlines how the upcoming Square Kilometre Array (SKA) radio telescope will bypass this fog, providing a dust-unbiased view of how galaxies and their central black holes grow together.
The limitations of dust-obscured views
Current understanding of galaxy formation faces a significant hurdle. As galaxies form stars, they produce vast amounts of dust. This dust absorbs light, hiding the most intense periods of star formation and the most heavily obscured active galactic nuclei (AGN)—the energetic regions surrounding a central black hole.
This "dusty" problem is acute during "cosmic noon." This is the period between redshifts $z \sim 1$ and $z \sim 3$ (a measure of how far back in time we are looking). During this era, star formation and black hole accretion reached their peak. Relying on optical or infrared data alone provides an incomplete census. Scientists are left unable to fully map the relationship between a galaxy's stellar mass and its dark matter (DM) halo (the invisible gravitational scaffolding governing galaxy growth). Without a way to peer through this shroud, the regulatory role of AGN feedback remains unclear. This feedback is the process where a black hole injects energy to suppress star formation.
A multi-tier, multi-frequency architecture
To bypass the dust, the authors propose leveraging the SKA's radio continuum surveys. Radio waves pass through dust relatively unimpeded. The proposed strategy uses a "wedding cake" survey design. This involves a tiered approach with Ultra-Deep, Deep, and Wide layers. Like a layered geological survey, this ensures the telescope catches both rare, luminous objects and faint, abundant populations.
The mechanism relies on three pillars:
- Tiered Depth: The "Wide" tier covers large areas to catch rare, powerful AGN. The "Deep" and "Ultra-Deep" tiers focus on smaller areas. These reach the sensitivity needed to detect typical, low-mass galaxies.
- Multi-frequency Imaging: The authors advocate for a matched-resolution approach from 0.15 to 15 GHz. This allows for the reconstruction of the radio Spectral Energy Distribution (SED). An SED is essentially a fingerprint of a source's energy output across different frequencies.
- High Angular Resolution: By using interferometry (combining signals from many antennas to simulate a massive single dish), the surveys aim for sub-arcsecond resolution. This is critical to prevent "source confusion." Confusion occurs when multiple distant objects appear as a single blurred blob.
Quantifying the SKA advantage
The authors use the T-RECS simulated radio catalogues to forecast survey effectiveness. They report that the SKA-Mid Band 2 "Deep" tier, with a $5\sigma$ sensitivity of $1\text{ }\mu\text{Jy/bm}$, is essential for a census of star-forming galaxies (SFGs) and radio-quiet (RQ) AGN. The "Ultra-Deep" tier reaches $0.05\text{ }\mu\text{Jy/bm}$. This extreme sensitivity allows the telescope to detect the faintest, most common galaxies and quiet black holes across cosmic time.
Comparing these capabilities to existing technology reveals a significant jump. As shown in, the proposed SKA surveys outperform current legacy surveys and planned precursors like LOFAR 2.0.
Current surveys like EMU (ASKAP) excel at tracing luminous AGN. However, they cannot reach the faint populations the SKA is designed to capture. Furthermore, shows that to reach a sample of 100 objects across various AGN sub-populations, the SKA's tiered approach is necessary.
It bridges the gap between bright outliers and the faint, mainstream galaxy population.
Engineering trade-offs and constraints
The roadmap involves significant technical costs. A primary constraint is the trade-off between angular resolution and integration time. As shown in, achieving fine resolution requires exponentially more observing time. The telescope must spend more effort "sharpening" the image.
Practitioners must also manage the phased rollout of the observatory. During the initial AA* phase, the SKA will have roughly four times poorer image resolution than its final AA4 configuration. This degradation increases the risk of confusion noise. This noise can drown out faint signals. The authors note this makes the Band 2 Ultra-Deep tier unfeasible during early stages. They suggest early operations prioritize wider, shallower surveys. Alternatively, they suggest high-frequency Band 5 observations where resolution is easier to maintain. Additionally, the Band 5b tier has a smaller field of view. Covering large areas will require intensive "mosaicking"—stitching together many individual pointings.
The verdict: A necessary panchromatic synergy
The SKA's radio surveys are not a standalone solution. They are a vital piece of a larger puzzle. The authors argue that the true scientific value comes from "panchromatic" synergies (using the full electromagnetic spectrum). Combining SKA's dust-unbiased radio maps with high-resolution infrared data from JWST or optical/NIR surveys from Euclid is essential. This helps characterize the physical properties of galaxies.
If the tiered strategy is implemented, the SKA will become a precision laboratory for galaxy evolution. It moves the field from asking if black holes regulate star formation to asking how they do it. For researchers, the SKA provides the statistical backbone that current, biased surveys lack. This is true provided it is integrated into a multi-wavelength observational framework.
Figures from the paper
How this was made
Model: nvidia/Gemma-4-26B-A4B-NVFP4
Persona: academic_accessible
Template: engineering_deepdive
Refinement: 0
Pipeline: forge-1.1
Evaluator: nvidia/Gemma-4-26B-A4B-NVFP4
Score: 91% (passed)
Model: nvidia/Gemma-4-26B-A4B-NVFP4
NVIDIA GB10 · 128 GB unified · NVFP4 · 100% local · $0 cloud
Tokens: 142,166
Wall-time: 262.8s
Tokens/s: 541.0