KICC papers

arXiv:2504.18637v1 Announce Type: new Abstract: The Sunyaev-Zeldovich (SZ) effect is a window into the astrophysical processes of galaxy clusters, and relativistic corrections (the "rSZ") promise to provide a global census of the gas feedback within clusters. Upcoming wide-field millimeter-wave surveys such as the Simons Observatory (SO), Fred Young Submillimeter Telescope, and CMB-S4 will make increasingly precise measurements of the SZ effect and its relativistic corrections. We present simulated full-sky maps of the rSZ effect and a fast code to generate it, for use in the development of analysis techniques and pipelines. As part of the websky simulation suite, our mock observations have semi-realistic cross-correlations with other large-scale structure tracers, offering insights into the formation and evolution of galaxy clusters and large-scale structure. As a demonstration of this, we examine what an SO-like experiment can learn from the rSZ effect. We find that high significance detections will be possible, provided that the instrumental systematics are under control, and that the evolution of cluster temperatures with mass and redshift can be probed in a manner complementary to X-ray measurements.

arXiv:2504.18384v1 Announce Type: new Abstract: While mounting observational evidence suggests that intermediate mass black holes (IMBHs) may be important in shaping the properties of dwarf galaxies both at high redshifts and in the local Universe, our theoretical understanding of how these IMBHs grow is largely incomplete. To address this, we perform high-resolution simulations of an isolated dwarf galaxy with a virial mass of $10^{10}~{\rm M}_{\odot}$ harbouring a $10^4~{\rm M}_{\odot}$ IMBH at its centre at a peak spatial resolution of $\lesssim 0.01$ pc. Within the fully multi-phase interstellar medium (ISM), we incorporate explicit sampling of stars from the initial mass function, photo-ionization, photoelectric heating, individual supernovae (SNe), as well as a Shakura-Sunyaev accretion disc model to track the evolution of BH mass and spin. We find that a nuclear star cluster (NSC) effectively captures the ISM gas and promotes formation of a circumnuclear disc (CND) on scales of $\lesssim 7$ pc. Simultaneously, gravitational torques from the NSC reduce CND angular momentum on (sub-)parsec scales, circularizing the gas onto the $\alpha$-accretion disc and promoting sustained IMBH growth at $\sim 0.01$ of the Eddington rate. While in the innermost regions ($\lesssim 0.5$ pc), star formation is highly suppressed, the CND is susceptible to fragmentation, leading to the formation of massive, young stars. Interestingly, despite an in-situ SN rate of $0.3~{\rm Myr}^{-1}$, the dense CND persists, sustaining BH accretion and leading to its net spin-up. Our study demonstrates the complexity of IMBH accretion within a multi-phase ISM, and paves the way for next-generation studies where IMBH growth in a fully cosmological context can be captured.

arXiv:2504.18245v1 Announce Type: new Abstract: The multi-tracer (MT) technique has been shown to outperform single-tracer analyses in the context of galaxy clustering. In this paper, we conduct a series of Fisher analyses to further explore MT information gains within the framework of non-linear bias expansion. We examine how MT performance depends on the bias parameters of the subtracers, showing that directly splitting the non-linear bias generally leads to smaller error bars in $A_s$, $h$, and $\omega_{\rm cdm}$ compared to a simple split in $b_1$. This finding opens the door to identifying subsample splits that do not necessarily rely on very distinct linear biases. We discuss different total and subtracer number density scenarios, as well as the possibility of splitting into more than two tracers. Additionally, we consider how different Fingers-of-God suppression scales for the subsamples can be translated into different $k_{\rm max}$ values. Finally, we present forecasts for ongoing and future galaxy surveys.

arXiv:2407.18349v2 Announce Type: replace Abstract: Understanding the impact of baryonic physics on cosmic structure formation is crucial for accurate cosmological predictions, especially as we usher in the era of large galaxy surveys with the Rubin Observatory as well as the Euclid and Roman Space Telescopes. A key process that can redistribute matter across a large range of scales is feedback from accreting supermassive black holes. How exactly these active galactic nuclei (AGN) operate from sub-parsec to Mega-parsec scales however remains largely unknown. To understand this, we investigate how different AGN feedback models in the Fable simulation suite affect the cosmic evolution of the matter power spectrum (MPS). Our analysis reveals that AGN feedback significantly suppresses clustering at scales $k \sim 10\,h\,cMpc^{-1}$, with the strongest effect at redshift $z = 0$ causing a reduction of $\sim 10\%$ with respect to the dark matter-only simulation. This is due to the efficient feedback in both radio (low Eddington ratio) and quasar (high Eddington ratio) modes in our fiducial Fable model. We find that variations of the quasar and radio mode feedback with respect to the fiducial Fable model have distinct effects on the MPS redshift evolution, with the radio mode being more effective on larger scales and later epochs. Furthermore, MPS suppression is dominated by AGN feedback effects inside haloes at $z = 0$, while for $z \gtrsim 1$ the matter distribution both inside and outside of haloes shapes the MPS suppression. Hence, future observations probing earlier cosmic times beyond $z \sim 1$ will be instrumental in constraining the nature of AGN feedback.

arXiv:2504.17549v1 Announce Type: new Abstract: The co-evolution of massive black holes (BHs) and their host galaxies is well-established within the hierarchical galaxy formation paradigm. Large-scale cosmological simulations are an ideal tool to study the repeated BH mergers, accretion and feedback that conspire to regulate this process. While such simulations are of fundamental importance for understanding the complex and intertwined relationship between BHs and their hosts, they are plagued with numerical inaccuracies at the scale of individual BH orbits. To quantify this issue, taking advantage of the $(100 \, h^{-1}\,\text{cMpc})^3$ FABLE simulation box, we track all individual BH mergers and the corresponding host galaxy mergers as a function of cosmic time. We demonstrate that BH mergers frequently occur prematurely, well before the corresponding merger of the host galaxies is complete, and that BHs are sometimes erroneously displaced from their hosts during close galaxy encounters. Correcting for these artefacts results in substantial macrophysical delays, spanning over several Gyrs, which are additional to any microphysical delays arising from unresolved BH binary hardening processes. We find that once the macrophysical delays are accounted for, high-mass BH merger events are suppressed, affecting the predictions for the BH population that may be observable with LISA and pulsar timing arrays. Furthermore, including these macrophysical delays leads to an increase in the number of observable dual active galactic nuclei, especially at lower redshifts, with respect to FABLE. Our results highlight the pressing need for more accurate modelling of BH dynamics in cosmological simulations of galaxy formation as we prepare for the multi-messenger era.

arXiv:2504.16988v1 Announce Type: cross Abstract: Precision observations of orbital systems have recently emerged as a promising new means of detecting gravitational waves and ultra-light dark matter, offering sensitivity in new regimes with significant discovery potential. These searches rely critically on precise modeling of the dynamical effects of these signals on the observed system; however, previous analyses have mainly only relied on the secularly-averaged part of the response. We introduce here a fundamentally different approach that allows for a fully time-resolved description of the effects of oscillatory metric perturbations on orbital dynamics. We find that gravitational waves and ultra-light dark matter can induce large oscillations in the orbital parameters of realistic binaries, enhancing the sensitivity to such signals by orders of magnitude compared to previous estimates.

arXiv:2504.17007v1 Announce Type: new Abstract: The James Webb Space Telescope (JWST) discovered 79 transients out to $z$$\sim$4.8 through the JADES Transient Survey (JTS), but the JTS did not find any $z$$>$5 transients. Here, we present the first photometric evidence of a $z$$>$5 transient/variable source with JWST. The source, AT 2023adya, resides in a $z_{\mathrm{spec}}$$=$5.274 galaxy in GOODS-N, which dimmed from $m_{\rm F356W}$$=$26.05$\pm$0.02 mag to 26.24$\pm$0.02 mag in the rest-frame optical over approximately two rest-frame months, producing a clear residual signal in the difference image ($m_{\rm F356W}$$=$28.01$\pm$0.17 mag; SN$_\mathrm{var}$$=$6.09) at the galaxy center. Shorter-wavelength bands (F090W/F115W) show no rest-frame ultraviolet brightness change. Based on its rest-frame V-band absolute magnitude of M$_\mathrm{V}$$=$$-$18.48 mag, AT 2023adya could be any core-collapse supernova (SN) subtype or an SN Ia. However, due to low SN Ia rates at high redshift, the SN Ia scenario is unlikely. Alternatively, AT 2023adya may be a variable active galactic nucleus (AGN). However, the JWST NIRCam/Grism spectrum shows no broad H$\alpha$ emission line (FWHM$=$130$\pm$26 km s$^{-1}$), disfavoring the variable AGN scenario. It is also unlikely that AT 2023adya is a tidal disruption event (TDE) because the TDE models matching the observed brightness changes have low event rates. Although it is not possible to determine AT 2023adya's nature based on the two-epoch single-band photometry alone, this discovery indicates that JWST can push the frontier of transient/variable science past $z$$=$5 and towards the epoch of reionization.

arXiv:2504.16791v1 Announce Type: new Abstract: Radiometers are crucial instruments in radio astronomy, forming the primary component of nearly all radio telescopes. They measure the intensity of electromagnetic radiation, converting this radiation into electrical signals. A radiometer's primary components are an antenna and a Low Noise Amplifier (LNA), which is the core of the ``receiver'' chain. Instrumental effects introduced by the receiver are typically corrected or removed during calibration. However, impedance mismatches between the antenna and receiver can introduce unwanted signal reflections and distortions. Traditional calibration methods, such as Dicke switching, alternate the receiver input between the antenna and a well-characterised reference source to mitigate errors by comparison. Recent advances in Machine Learning (ML) offer promising alternatives. Neural networks, which are trained using known signal sources, provide a powerful means to model and calibrate complex systems where traditional analytical approaches struggle. These methods are especially relevant for detecting the faint sky-averaged 21-cm signal from atomic hydrogen at high redshifts. This is one of the main challenges in observational Cosmology today. Here, for the first time, we introduce and test a machine learning-based calibration framework capable of achieving the precision required for radiometric experiments aiming to detect the 21-cm line.

arXiv:2502.02647v2 Announce Type: replace Abstract: The field of high redshift galaxy formation has been revolutionised by JWST, which is yielding unprecedented insights on galaxy assembly at early times. Our key aim is to study the physical mechanisms that can explain the unexpected abundance of bright galaxies at $z \geq 11$, as well as their metal enrichment and spectral properties. We also use recent data to determine the key sources of reionisation. To do so, we implement cold gas fractions and star formation efficiencies derived from the SPHINX20 high-resolution radiation-hydrodynamics simulation into DELPHI, a semi-analytic model that tracks the assembly of dark matter halos and their baryonic components from $z \sim 4.5-40$. In addition, we explore two different methodologies to boost galaxy luminosities at $z \geq 11$: a stellar initial mass function (IMF) that becomes increasingly top-heavy with decreasing metallicity and increasing redshift (eIMF model), and star formation efficiencies that increase with increasing redshift (eSFE model). Our key findings are: (i) both the eIMF and eSFE models can explain the abundance of bright galaxies at $z \geq 11$; (ii) dust attenuation plays an important role for the bright-end of the UV LF at $z \leq 11$; (iii) the mass-metallicity relation is in place as early as $z \sim 17$ in all models although its slope is model-dependent; (iv) within the spread of both models and observations, all of our models are in good agreement with current estimates of $\beta$ slopes at $z \sim 5-17$ and Balmer break strengths at $z \sim 6-10$; (v) in the eIMF model, galaxies at $z\geq12$ or with $\rm{M_{UV}}\geq-18$ show values of $\xi_{\rm{ion}} \sim 10^{25.55}~{\rm [Hz~erg^{-1}]}$, twice larger than in other models; (vi) star formation in galaxies below $10^{9}\rm{M_{\odot}}$ is the key driver of reionisation, providing the bulk ($\sim 85\%$) of ionising photons down to its midpoint at $z \sim 7$.

arXiv:2504.15346v1 Announce Type: new Abstract: The dust attenuation of galaxies is highly diverse and closely linked to stellar population properties and the star dust geometry, yet its relationship to galaxy morphology remains poorly understood. We present a study of 141 galaxies ($9<\log(\rm M_{\star}/\rm M_{\odot})<11.5$) at $1.7

arXiv:2504.15334v1 Announce Type: new Abstract: In the presence of a weak gravitational wave (GW) background, astrophysical binary systems act as high-quality resonators, with efficient transfer of energy and momentum between the orbit and a harmonic GW leading to potentially detectable orbital perturbations. In this work, we develop and apply a novel modeling and analysis framework that describes the imprints of GWs on binary systems in a fully time-resolved manner to study the sensitivity of lunar laser ranging, satellite laser ranging, and pulsar timing to both resonant and nonresonant GW backgrounds. We demonstrate that optimal data collection, modeling, and analysis lead to projected sensitivities which are orders of magnitude better than previously appreciated possible, opening up a new possibility for probing the physics-rich but notoriously challenging to access $\mu\mathrm{Hz}$ frequency GWs. We also discuss improved prospects for the detection of the stochastic fluctuations of ultra-light dark matter, which may analogously perturb the binary orbits.

arXiv:2504.14682v1 Announce Type: new Abstract: We have discovered a substantial sodium doublet (Na D $\lambda\lambda$5890, 5896\AA)-traced neutral outflow in a quenching galaxy JADES-GS-206183 at $z=1.317$ in GOODS-S field. Its JWST NIRSpec/MSA spectrum shows a significantly blueshifted and deep Na D absorption, revealing a neutral outflow with a velocity of $v_{\rm out}=828^{+79}_{-49}\,\mathrm{km\,s^{-1}}$ and a mass outflow rate of $\log(\dot{M}_{\rm out}/\mathrm{M_{\odot}\,yr^{-1}})=2.40^{+0.11}_{-0.16}$. The mass outflow rate of this outflow is higher than any of the neutral outflows identified previously beyond $z\sim1$ by the same line diagnostic and is comparable with those in local galaxies with extremely strong star formation activities or luminous AGN. Nonetheless, the best-fit SED modeling of JADES-GS-206183, based on its multi-band photometry from HST/ACS to JWST/NIRCam, suggests that the host galaxy now is quenched, and the Paschen $\alpha$ (Pa$\alpha$) emission in the FRESCO NIRCam grism spectrum confirms its current low star formation rate ($10.78\pm 0.55\,\mathrm{M_{\odot}\,yr^{-1}}$). More surprisingly, optical line ratio diagnostics indicate that the current AGN activity of JADES-GS-206183, if present, is weak. Even though we tentatively detect a broad component of the H$\alpha$ line, it is more likely tracing the ionized outflow than an AGN. The results demonstrate that the Na D outflow in JADES-GS-206183 is highly unlikely to be driven by current star formation or nuclear activity. Instead, we propose that the outflow that we are witnessing in JADES-GS-206183 may be a long-lasting fossil outflow, powered by previous AGN activity that has recently shut down.

arXiv:2504.14009v1 Announce Type: new Abstract: Supernova (SN) 2014C is a rare transitional event that exploded as a hydrogen-poor, helium-rich Type Ib SN and subsequently interacted with a hydrogen-rich circumstellar medium (CSM) a few months post explosion. This unique interacting object provides an opportunity to probe the mass-loss history of a stripped-envelope SN progenitor. Using the James Webb Space Telescope (JWST), we observed SN 2014C with the Mid-InfraRed Instrument Medium Resolution Spectrometer at 3477 days post explosion (rest frame), and the Near-InfraRed Spectrograph Integral Field Unit at 3568 days post explosion, covering 1.7 to 25 $\mu$m. The bolometric luminosity indicates that the SN is still interacting with the same CSM that was observed with the Spitzer Space Telescope 40--1920 days post explosion. JWST spectra and near-contemporaneous optical and near-infrared spectra show strong [Ne II] 12.831 $\mu$m, He 1.083 $\mu$m, H$\alpha$, and forbidden oxygen ([O I] $\lambda$$\lambda$6300, 6364, [O II] $\lambda$$\lambda$7319, 7330, and [O III] $\lambda$$\lambda$4959, 5007) emission lines with asymmetric profiles, suggesting a highly asymmetric CSM. The mid-IR continuum can be explained by ~0.036 $M_\odot$ of carbonaceous dust at ~300 K and ~0.043 $M_\odot$ of silicate dust at $\sim$200 K. The observed dust mass has increased tenfold since the last Spitzer observation 4 yr ago, with evidence suggesting that new grains have condensed in the cold dense shell between the forward and reverse shocks. This dust mass places SN 2014C among the dustiest SNe in the mid-IR and supports the emerging observational trend that SN explosions produce enough dust to explain the observed dust mass at high redshifts.

arXiv:2502.02983v2 Announce Type: replace Abstract: Recent work has suggested that, during reionisation, spatial variations in the ionising radiation field should produce enhanced Ly ${\alpha}$ forest transmission at distances of tens of comoving Mpc from high-redshift galaxies. We demonstrate that the Sherwood-Relics suite of hybrid radiation-hydrodynamical simulations are qualitatively consistent with this interpretation. The shape of the galaxy--Ly ${\alpha}$ transmission cross-correlation is sensitive to both the mass of the haloes hosting the galaxies and the volume averaged fraction of neutral hydrogen in the IGM, $\bar{x}_{\rm HI}$. The reported excess Ly ${\alpha}$ forest transmission on scales r ~ 10 cMpc at $\langle z \rangle \approx 5.2$ -- as measured using C IV absorbers as proxies for high-redshift galaxies -- is quantitatively reproduced by Sherwood-Relics at z = 6 if we assume the galaxies that produce ionising photons are hosted in haloes with mass $M_{\rm h}\geq 10^{10}~h^{-1}\,{\rm M}_\odot$. However, this redshift mismatch is equivalent to requiring $\bar{x}_{\rm HI}\sim 0.1$ at $z\simeq 5.2$, which is inconsistent with the observed Ly ${\alpha}$ forest effective optical depth distribution. We suggest this tension may be partly resolved if the minimum C IV absorber host halo mass at z > 5 is larger than $M_{\rm h}=10^{10}~h^{-1}\,{\rm M}_\odot$. After reionisation completes, relic IGM temperature fluctuations will continue to influence the shape of the cross-correlation on scales of a few comoving Mpc at $4 \leq z \leq 5$. Constraining the redshift evolution of the cross-correlation over this period may therefore provide further insight into the timing of reionisation.

arXiv:2410.24134v3 Announce Type: replace Abstract: We use the projected clustering of quasars in the Gaia-unWISE quasar catalog, Quaia, and its cross-correlation with CMB lensing data from Planck, to measure the large-scale turnover of the matter power spectrum, associated with the size of the horizon at the epoch of matter-radiation equality. The turnover is detected with a significance of between $2.3$ and $3.1\sigma$, depending on the method used to quantify it. From this measurement, the equality scale is determined at the $\sim20\%$ level. Using the turnover scale as a standard ruler alone (suppressing information from the large-scale curvature of the power spectrum), in combination with supernova data through an inverse distance ladder approach, we measure the current expansion rate to be $H_0=62.7\pm17.2\,{\rm km}\,{\rm s}^{-1}\,{\rm Mpc}^{-1}$. The addition of information coming from the power spectrum curvature approximately halves the standard ruler uncertainty. Our measurement in combination with calibrated supernovae from Pantheon$+$ and SH0ES constrains the CMB temperature to be $T_{\rm CMB}=3.10^{+0.48}_{-0.36}\,{\rm K}$, independently of CMB data. Alternatively, assuming the value of $T_{\rm CMB}$ from COBE-FIRAS, we can constrain the effective number of relativistic species in the early Universe to be $N_{\rm eff}=3.0^{+5.8}_{-2.9}$.

arXiv:2407.04660v2 Announce Type: replace Abstract: We introduce COBRA (Cosmology with Optimally factorized Bases of Radial Approximants), a novel framework for rapid computation of large-scale structure observables. COBRA separates scale dependence from cosmological parameters in the linear matter power spectrum while also minimising the number of necessary basis terms $N_b$, thus enabling direct and efficient computation of derived and nonlinear observables. Moreover, the dependence on cosmological parameters is efficiently approximated using radial basis function interpolation. We apply our framework to decompose the linear matter power spectrum in the standard $\Lambda$CDM scenario, as well as by adding curvature, dynamical dark energy and massive neutrinos, covering all redshifts relevant for Stage IV surveys. With only a dozen basis terms $N_b$, COBRA reproduces exact Boltzmann solver calculations to $\sim 0.1\%$ precision, which improves further to $0.02\%$ in the pure $\Lambda$CDM scenario. Using our decomposition, we recast the one-loop redshift space galaxy power spectrum in a separable minimal-basis form, enabling $\sim 4000$ model evaluations per second at $0.02\%$ precision on a single thread. This constitutes a considerable improvement over previously existing methods (e.g., FFTLog) opening a window for efficient computations of higher loop and higher order correlators involving multiple powers of the linear matter power spectra. The resulting factorisation can also be utilised in clustering, weak lensing and CMB analyses. Our implementation will be made public upon publication.

arXiv:2307.13768v2 Announce Type: replace Abstract: We revisit the flat-sky approximation for evaluating the angular power spectra of projected random fields by retaining information about the correlations along the line of sight. With broad, overlapping radial window functions, these line-of-sight correlations are suppressed and are ignored in the Limber approximation. However, retaining the correlations is important for narrow window functions or unequal-time spectra but introduces significant computational difficulties due to the highly oscillatory nature of the integrands involved. We deal with the integral over line-of-sight wave-modes in the flat-sky approximation analytically, using the FFTlog expansion of the 3D power spectrum. This results in an efficient computational method, which is a substantial improvement compared to any full-sky approaches. We apply our results to galaxy clustering (with and without redshift-space distortions), CMB lensing and galaxy lensing observables. For clustering, we find excellent agreement with the full-sky results on large (percent-level agreement) and intermediate or small (subpercent agreement) scales, dramatically out-performing the Limber approximation for both wide and narrow window functions, and in equal- and unequal-time cases. In the case of lensing, we show on the full sky that the angular power spectrum of the convergence can be very well approximated by projecting the 3D Laplacian (rather than the correct angular Laplacian) of the gravitational potential, even on large scales. Combining this approximation with our flat-sky techniques provides an efficient and accurate evaluation of the CMB lensing angular power spectrum on all scales.

arXiv:2503.21728v2 Announce Type: replace Abstract: Radio-frequency interference (RFI) is a major systematic limitation in radio astronomy, particularly for science cases requiring high sensitivity, such as 21 cm cosmology. Traditionally, RFI is dealt with by identifying its signature in the dynamic spectra of visibility data and flagging strongly affected regions. However, for RFI sources that do not occupy narrow regions in the time-frequency space, such as persistent local RFI, modeling these sources could be essential to mitigating their impact. This paper introduces two methods for detecting and characterizing local RFI sources from radio interferometric visibilities: matched filtering and maximum a posteriori (MAP) imaging. These algorithms use the spherical wave equation to construct three-dimensional near-field image cubes of RFI intensity from the visibilities. The matched filter algorithm can generate normalized maps by cross-correlating the expected contributions from RFI sources with the observed visibilities, while the MAP method performs a regularized inversion of the visibility equation in the near field. We developed a full polarization simulation framework for RFI and demonstrated the methods on simulated observations of local RFI sources. The stability, speed, and errors introduced by these algorithms were investigated, and, as a demonstration, the algorithms were applied to a subset of NenuFAR observations to perform spatial, spectral, and temporal characterization of two local RFI sources. We used simulations to assess the impact of local RFI on images, the uv plane, and cylindrical power spectra, and to quantify the level of bias introduced by the algorithms in order to understand their implications for the estimated 21 cm power spectrum with radio interferometers. The near-field imaging and simulation codes are publicly available in the Python library nfis.

arXiv:2504.08081v1 Announce Type: new Abstract: JAX-bandflux is a JAX implementation of critical supernova modelling functionality for cosmological analysis. The codebase implements key components of the established library SNCosmo in a differentiable framework, offering efficient parallelisation and gradient-based optimisation capabilities through GPU acceleration. The package facilitates differentiable computation of supernova light curve measurements, supporting the inference of SALT parameters necessary for cosmological analysis.

arXiv:2504.08041v1 Announce Type: new Abstract: Feedback from active galactic nuclei (AGN) is crucial for regulating galaxy evolution. Motivated by observations of broad absorption line winds from rapidly accreting supermassive black holes (SMBHs), we introduce the Mistral AGN feedback model, implemented in the Arepo code. Mistral comes in two versions: continuous radial (Mistral-continuous) and stochastic bipolar momentum deposition (Mistral-stochastic). Using the framework of the IllustrisTNG simulations, we explore the effect of Mistral on BH and galaxy properties, through an idealized Milky Way-mass galaxy and cosmological zoom simulations run down to $z=2$. Unlike standard thermal AGN feedback prescriptions, Mistral generates galaxy-scale winds that mimic outflows driven by BH accretion. Mistral-continuous produces short-lived galactic fountains, and is inefficient at regulating the growth of massive galaxies at $z=2$. In contrast, Mistral-stochastic efficiently suppresses star formation in massive galaxies, and reproduces the empirical stellar-to-halo mass and ($z=0$) BH-stellar mass relations. By supporting large-scale ($>50\,\rm kpc$) outflows while simultaneously preventing gas inflows, Mistral-stochastic additionally regulates the cold and hot gas fractions at both galaxy and halo scales. Mistral-stochastic therefore works self-consistently across the halo mass range explored $\left(10^{12}-3\times10^{13}\,\rm M_\odot\right)$, without adopting a SMBH-mass dependent AGN feedback scheme such as the one used in IllustrisTNG. Our model is a promising tool for predicting the impact of radiatively efficient AGN winds on galaxy evolution, and interpreting the growing population of high-redshift galaxies and quasars observed by JWST. This work is part of the "Learning the Universe" collaboration, which aims to infer the physical processes governing the evolution of the Universe.