Intermediate-Band Imaging Survey (IBIS)

• Intermediate-Band Imaging Survey (IBIS) is a dual-purpose astronomical survey that uses medium-band filters for both high-resolution solar spectropolarimetry and extragalactic BAO mapping.
• The extragalactic program employs five contiguous filters on DECam, reaching a depth of 25.5 magnitudes and achieving a photometric redshift precision of approximately 0.05.
• By analyzing angular power spectra with Fisher matrix forecasts, IBIS constrains BAO dilation parameters with an uncertainty near 2.6%, aiding in precise cosmic distance measurements.

The Intermediate-Band Imaging Survey (IBIS) encompasses two distinct but technically connected efforts: (1) the solar physics IBIS instrument and archive at the Dunn Solar Telescope, which delivers high-resolution narrowband spectropolarimetric imaging of the solar photosphere and chromosphere, and (2) the extragalactic IBIS medium-band optical imaging program using DECam on the Blanco 4m telescope that enables large-scale structure mapping and cosmic distance measurements via photometric selection of high-redshift galaxies. This article focuses primarily on the DECam-based IBIS survey and its cosmological applications, while providing brief context for the solar IBIS where relevant.

1. Survey Architecture and Instrumentation

IBIS employs five contiguous medium-band filters on DECam, each with FWHM $\Delta\lambda\simeq250~\textrm{\AA}$, spanning central wavelengths $\lambda_n=(4110, 4380, 4640, 4900, 5170)$ Å and tracking Ly$\alpha$ over $2.2 < z < 3.5$ (Feder et al., 6 Dec 2025, Ebina et al., 30 Sep 2025). The survey footprint is planned at $5000~\textrm{deg}^2$, with initial clustering analysis conducted on a $10~\textrm{deg}^2$ pilot region. Deep fields (e.g., COSMOS, XMM-LSS) reach medium-band $5\sigma$ depths of $25.5$ magnitudes via $\sim$10 hours total integration per band, using 92 independent dithers for imaging uniformity. Instrumental flux calibration is maintained via nightly standards and cross-registration with external catalogs (e.g., HSC).

Extensive source-modeling is achieved with Tractor, which fits morphological profiles and delivers both total and fiber fluxes, accompanied by covariance-based uncertainty estimates. The survey strategy targets five overlapping redshift bins, each defined by filter bandpass and Ly$\alpha$–redshift mapping, yielding $\Delta z \simeq 0.2$ per shell.

2. Target Selection and Sample Properties

IBIS selects primary tracers via medium-band photometric excess, emphasizing Ly$\alpha$ emitters (LAEs) and, in some analyses, Lyman break galaxies (LBGs) (Ebina et al., 30 Sep 2025). The dominant selection criteria employ color cuts measuring flux boosts in a central medium band relative to adjacent bands or synthetic broadband constructs. Example criteria for Ly$\alpha$ in M464 include:

• $M438 - M464 > 0.6$
• $M464 - M490 < 0$
• $M438 - M464 > 1 + 0.6\,(M464 - M490)$, augmented by FRACFLUX and blending rejection. These color excesses correspond to photometric redshift precision $\sigma_z/(1+z)\approx0.05$ (validated against spectroscopic redshifts), producing thin redshift distributions which are crucial for angular BAO measurements.

Comoving number densities per bin range between $1.5–3.0 \times 10^{-4}~h^3~\textrm{Mpc}^{-3}$ (125–250 deg$^{-2}$ per shell), with galaxy bias rising linearly from $b_g=2.0$ (at $z=2.41$) to $b_g=2.5$ (at $z=3.26$) (Feder et al., 6 Dec 2025).

Redshift Bin [z]	Mean $z$	Bias $b_g$	$n_{LAE} [10^{-4}~h^3~\textrm{Mpc}^{-3}]$	Surface Density [deg$^{-2}$]
[2.26, 2.56]	2.41	2.0	1.5–3.0	125–250
[3.10, 3.41]	3.26	2.5	1.5–3.0	125–250

Interloper contamination is characterized in forecasts; the fiducial $f_{\text{int}}=10\%$ is included in signal modeling, while cluster analyses confirm a mixture of LAEs and LBGs with $b\sim1.8-2.5$ (Ebina et al., 30 Sep 2025).

3. Clustering Analysis and Cosmological Methodology

IBIS characterizes large-scale structure through measurement of the angular power spectrum $C_\ell^{ij}$ between bins $i$ and $j$, incorporating both radial and transverse modes by employing the flat-sky approximation:

$$C_\ell^{ij} = \frac{1}{\pi \chi_i \chi_j} \int_{k_\parallel^{\min}}^{k_\parallel^{\max}} {\rm d}k_\parallel\, P_g^{ij}(k_\perp, k_\parallel; z)\, |\tilde{\phi}_i(k_\parallel)||\tilde{\phi}_j(k_\parallel)|$$

with $k_\perp = \ell/\sqrt{\chi_i \chi_j}$ and where the radial window $\tilde{\phi}_i(k_\parallel)$ is the Fourier transform of the survey selection function (Feder et al., 6 Dec 2025).

Shot-noise–dominated covariance matrices are constructed as:

$$\text{Cov}\left[C_\ell^{ij}, C_\ell^{mn}\right] = \frac{1}{(2\ell + 1) f_{\rm sky}} \left( \tilde{C}_\ell^{im} \tilde{C}_\ell^{jn} + \tilde{C}_\ell^{in} \tilde{C}_\ell^{jm} \right)$$

where $\tilde{C}_\ell^{ij} = C_\ell^{ij} + N_\ell^{ij}$ and $N_\ell^{ij} = \delta_{ij}/\bar{n}_i^{2D}$.

The BAO dilation parameter $\alpha$ rescales the "wiggle" component of the matter power spectrum:

$$P_w(k; z; \alpha) = P_w\left(\frac{k}{\alpha}; z\right)$$

with $$\alpha \equiv \frac{D_A(z)/r_d}{(D_A(z)/r_d)_{\rm fid}}$$.

Fisher matrix formalism is used to forecast uncertainties on $\alpha$ by modeling $C_\ell$ as a sum of rescaled BAO and smooth spline components, marginalizing over the latter to isolate broadband systematics (Feder et al., 6 Dec 2025).

4. BAO Forecasts and Survey Sensitivity

Using a fiducial LAE density $n = 2 \times 10^{-4}~h^3~\textrm{Mpc}^{-3}$, bias $b=2.2$ at $z_{\rm eff} = 2.8$, $f_{\text{int}}=10\%$, and survey area $A_{\rm survey}=5000~\textrm{deg}^2$ over multipoles $100 \leq \ell \leq 1500$, IBIS achieves $\sigma(\alpha)=2.6\%$ (Feder et al., 6 Dec 2025). This directly yields a BAO angular diameter distance constraint with

$\sigma(D_A) = 2.6\% \times (D_A/r_d)_{\rm fid}$

Sensitivity in $\sigma(\alpha)$ scales approximately as $(A_{\rm survey}\,\bar{n})^{-1/2}$ under shot-noise domination. Increasing bias by $\pm0.5$ shifts $\sigma(\alpha)$ by $\pm15\%$. Interloper fraction impacts are modeled as $\sigma(\alpha)$ degradation by $(1-f_{\text{int}})^{-1}$, e.g., $f_{\text{int}}=20\%$ increases $\sigma(\alpha)$ by $\sim25\%$.

Comparative analyses show that IBIS imaging alone positions between spectroscopic three-dimensional surveys (e.g., DESI Lyman-$\alpha$ forest at sub-percent $D_A$) and traditional photometric BAO (with $\sigma(D_A)\sim5\%$ at $z<1$). For Stage-V medium-band upgrades (e.g., Rubin MB), even lower $\sigma(\alpha) \lesssim 1\%$ is forecast with greater area and higher density (Feder et al., 6 Dec 2025).

5. Halo Occupation, Large-Scale Structure, and Simulation Implications

Clustering analyses leverage both HOD and perturbation theory. Measured correlation lengths are $r_0 \simeq 3–4~h^{-1}\textrm{Mpc}$; linear bias $b \sim 1.8–2.5$ matches external LAE and LBG samples (Ebina et al., 30 Sep 2025). Selection functions yield redshift shells of width $\Delta z \sim 0.2$, optimal for minimizing Limber projection and maximizing angular BAO detectability. Simulations are recommended for validation of mock catalogs and for the propagation of selection function uncertainties into cosmological inference.

The overlapping nature of LAE/LBG samples and robust clustering signal in IBIS medium-band selection underscore their suitability as DESI-II/Spec-S5 targets and foreground calibration tools.

6. Scientific Applications and Impact

IBIS medium-band imaging provides angular BAO constraints at $z>2.2$, bridging conventional photometric ($z<1$) and spectroscopically-selected BAO surveys (Feder et al., 6 Dec 2025). As a "forward scout," IBIS photometry calibrates selection functions, redshift distributions, and systematics ahead of spectroscopic follow-up, optimizing fiber assignment and exposure strategies.

The methodology supports multiwavelength studies and joint analysis with other cosmological probes, including SDO/HMI and ALMA for solar IBIS data (Ermolli et al., 2022). Over 150 refereed papers have employed IBIS data, focusing on topics from solar magnetism and MHD wave propagation (via spectropolarimetric imaging) to extragalactic structure and dark energy constraints (via clustering and BAO).

7. Archive Structure, Data Access, and Future Prospects

The IBIS archive (for solar data) is organized into hierarchical levels:

• Level 0: Raw NB/BB frames (25 TB)
• Level 1: Calibrated NB cubes (4 TB)
• Level 1.5: Science-ready polarization and velocity maps (15 GB)
• Level 2: Milne–Eddington inversions with VFISV (20 GB)

All levels are searchable via a Django/MySQL front-end with SOLARNET-compliant metadata. Science-ready products include CP, LP, NCP, and $V_{\rm los}$ maps, with quick-look movies and seeing diagnostics available for user assessment (Ermolli et al., 2022).

Planned upgrades include extended area ($A>3000~\textrm{deg}^2$), improved sensitivity, and higher-level products such as NLTE inversion maps (e.g., DeSIRe). For the cosmological IBIS, future Stage-V survey implementations such as Rubin MB and HSC-like extensions up to $z=5$ promise sub-percent BAO precision and expanded parameter space (Feder et al., 6 Dec 2025).

A plausible implication is that the integration of IBIS imaging surveys with space-based and deep spectroscopy programs will enhance the precision of $D_A(z)$ measurements at cosmic noon, calibrate photometric selection at scale, and provide critical leverage for next-generation cosmological analyses.