Makani Galaxy: Extreme Starburst Winds

• Makani Galaxy is a compact, massive post-starburst system exhibiting a 100 kpc-scale, starburst-driven multiphase wind.
• It features two distinct outflow episodes: an older, slower outer wind and a recent, fast inner wind reaching speeds up to 2000 km/s.
• Multiwavelength studies reveal shock-dominated ionization, active mass and momentum transport, and PAH-bearing dust survival in the CGM.

Makani Galaxy, SDSS J211824.06+001729.4, is a compact, massive galaxy at $z = 0.459$ whose defining property is a starburst-driven galactic wind extending from the interstellar medium into the circumgalactic medium on $\sim 100$ kpc scales. It is described as a compact, massive post-starburst galaxy and as a compact merger remnant undergoing an extreme starburst, with stellar mass $M_* \approx 10^{11.1}\,M_\odot$, an effective radius reported as $r_e = 2.3$ kpc or $r_e \approx 2.5$ kpc, and a central starburst of radius $\approx 400$ pc that is consistent with being Eddington-limited. Makani is a benchmark system because its outflow can be decomposed into two star-formation and wind episodes separated in space and time, and because warm ionized gas, molecular gas, neutral gas, O VI/Ly$\alpha$-emitting coronal interfaces, and PAH-bearing warm dust have all been traced across the inner halo and CGM (Rupke et al., 2023, Ha et al., 25 Mar 2025, Veilleux et al., 10 Jul 2025).

1. Host galaxy and star-formation history

Makani lies at a physical scale of $6.02$ kpc per arcsecond for a flat $\Lambda$CDM cosmology with $\Omega_m = 0.315$ and $\sim 100$0. It is massive, compact, and centrally concentrated. The central starburst is extremely compact, intense, and consistent with being Eddington-limited, while the broader system is identified as post-starburst and merger-remnant-like. These descriptions frame Makani as a transition object in which a recent or ongoing starburst coexists with evidence for rapid structural and gaseous transformation (Rupke et al., 2023, Veilleux et al., 10 Jul 2025, Ha et al., 25 Mar 2025).

The star-formation rate depends strongly on tracer. Radio and infrared indicators give $\sim 100$1–$\sim 100$2, whereas H$\sim 100$3 from the nuclear star-forming component gives a much smaller $\sim 100$4. The optical study explicitly interprets this discrepancy as suggesting very recent quenching and/or LyC leakage and/or heavy obscuration. A plausible implication is that Makani is being observed during a short-lived stage in which feedback has already altered the observable star-forming signatures without erasing the energetic imprint of the recent burst (Rupke et al., 2023).

The temporal structure inferred for the central activity consists of two episodes: Episode I, approximately $\sim 100$5 Gyr ago, and Episode II, approximately $\sim 100$6 Myr ago. The wind properties strongly track this history, so the star-formation chronology is not merely a stellar-population inference but the organizing principle for the gaseous phenomenology seen from the nucleus to the halo (Veilleux et al., 10 Jul 2025).

2. Two-episode outflow architecture and phase structure

Makani’s outflow is explicitly resolved into two episodes. Episode I is the outer, older component, with age $\sim 100$7 Myr and radial range $\sim 100$8–$\sim 100$9 kpc. It has slowed substantially: projected speeds are $M_* \approx 10^{11.1}\,M_\odot$0, the linewidth is $M_* \approx 10^{11.1}\,M_\odot$1, and the mean bulk motion is consistent with ballistic travel over $M_* \approx 10^{11.1}\,M_\odot$2 Myr, $M_* \approx 10^{11.1}\,M_\odot$3. Episode II is the inner, recent component, with age $M_* \approx 10^{11.1}\,M_\odot$4 Myr and radial range $M_* \approx 10^{11.1}\,M_\odot$5–$M_* \approx 10^{11.1}\,M_\odot$6 kpc, although faint [O II] reaches $M_* \approx 10^{11.1}\,M_\odot$7 kpc; its maximum speeds exceed $M_* \approx 10^{11.1}\,M_\odot$8, and $M_* \approx 10^{11.1}\,M_\odot$9 (Rupke et al., 2023).

The phase structure is strongly stratified with radius. Warm ionized gas is detected to $r_e = 2.3$0–$r_e = 2.3$1 kpc in Balmer lines and to $r_e = 2.3$2 kpc in [O II]. Neutral and molecular gas, together with dust traced by Balmer decrements, Na I D, Mg II, and CO, are prominent within $r_e = 2.3$3–$r_e = 2.3$4 kpc and are not detected beyond $r_e = 2.3$5 kpc in the optical/near-UV studies. The wind-to-galaxy size ratio $r_e = 2.3$6 proves that the outflow extends into the CGM, not merely the stellar body or a local superbubble environment (Rupke et al., 2023).

Subsequent ultraviolet and infrared observations broadened this phase inventory. Deep HST imaging detects O VI and Ly$r_e = 2.3$7 emission across the [O II] nebula with similar morphology and extent, out to $r_e = 2.3$8 kpc, while JWST detects PAH-bearing warm dust to $r_e = 2.3$9 kpc. These later data show that the outer halo is not only ionized but also hosts coronal cooling interfaces and dust-bearing material, even where earlier slit-based reddening estimates suggested little or no extinction (Ha et al., 25 Mar 2025, Veilleux et al., 10 Jul 2025).

The resulting picture is a temporally resolved, multiphase, CGM-scale wind. Episode II is fast, dusty, neutral, molecular, and strongly multiphase in the inner halo. Episode I is older, slower, and dominated observationally by warm ionized gas and coronal/interface tracers at larger radii. This suggests radial and temporal evolution in which a recently launched compact-starburst wind transitions into a more diffuse and CGM-coupled structure.

3. Optical spectroscopy, ionization state, and shocks

The foundational spectroscopic analysis used Keck II/ESI echellette observations with a $r_e \approx 2.5$0 slit, $r_e \approx 2.5$1, and simultaneous rest-frame optical coverage from [O II] $r_e \approx 2.5$2 through [S II] $r_e \approx 2.5$3. Balmer lines were detected across the nebula, together with numerous collisionally excited lines including [O III] and its auroral $r_e \approx 2.5$4 Å line, [O I] $r_e \approx 2.5$5 Å, [N II] $r_e \approx 2.5$6 Å, [S II], [Ne III], and [Ne V]. These line detections enabled extinction, density, temperature, and excitation diagnostics across both wind episodes (Rupke et al., 2023).

The extinction profile is radially structured. Assuming Case B and $r_e \approx 2.5$7, $r_e \approx 2.5$8 peaks at $r_e \approx 2.5$9 around $\approx 400$0 kpc and declines to $\approx 400$1 by $\approx 400$2 kpc. Electron densities derived from [S II] $\approx 400$3 and [O II] $\approx 400$4 vary strongly with position: in the inner fast wind, $\approx 400$5 with a $\approx 400$6 range of $\approx 400$7–$\approx 400$8, whereas outer apertures have $\approx 400$9, weakly constrained below $\alpha$0. Auroral-line ratios, specifically [O III] $\alpha$1 dex and [N II] $\alpha$2 to $\alpha$3 dex, imply $\alpha$4 K in the shocked gas (Rupke et al., 2023).

Line ratios were interpreted with MAPPINGS shock models, using Solar-metallicity fast-shock grids with $\alpha$5 for the outer wind and $\alpha$6 for the inner wind, at low magnetic parameter $\alpha$7. Both shock-only and shock+precursor cases were considered, with precursor pre-ionization especially important in the inner wind. In diagnostic diagrams outside the nucleus, apertures lie in composite/LINER regions characterized by high [O I]/H$\alpha$8 and [S II]/H$\alpha$9, modest [O III]/H$6.02$0, and low [O III]/[O II]. The broad nuclear component falls in the AGN region, but Makani shows no multiwavelength AGN signatures; in this system, the AGN-like optical ratios and [Ne V] are explained by fast shocks rather than by a luminous AGN (Rupke et al., 2023).

The shock speeds inferred from the optical line ratios track the kinematics. Episode II requires $6.02$1–$6.02$2, consistent with $6.02$3, together with a hard radiation field and shock+precursor emission. Episode I requires $6.02$4, consistent with $6.02$5, with lower ionization parameter and weaker precursor contribution. The empirical relation $6.02$6 in both episodes is one of the clearest arguments that shocks, rather than stellar photoionization, dominate the ionization of the extended nebula (Rupke et al., 2023).

4. Mass, momentum, energy, and wind driving

For case B recombination and uniform density, the ionized-gas mass was estimated from the H$6.02$7 luminosity as

$6.02$8

In Makani, the H$6.02$9 luminosity was apportioned between shock precursor and post-shock regions using MAPPINGS guidance, with representative densities of $\Lambda$0 and $\Lambda$1 for Episode I and $\Lambda$2 and $\Lambda$3 for Episode II. This procedure weights the mass toward the lower-density precursor component. The outflow rate was then bracketed with $\Lambda$4 and $\Lambda$5, using deprojected velocities, characteristic radii $\Lambda$6 kpc and $\Lambda$7 kpc, and both $\Lambda$8 and $\Lambda$9 to span the dynamics (Rupke et al., 2023).

The total nebular luminosity is $\Omega_m = 0.315$0, bootstrapped from ESI H$\Omega_m = 0.315$1 and KCWI [O II]. The star-forming component accounts for $\Omega_m = 0.315$2–$\Omega_m = 0.315$3, with an adopted value of $\Omega_m = 0.315$4. For the inner fast wind, the preferred estimates give $\Omega_m = 0.315$5–$\Omega_m = 0.315$6 and $\Omega_m = 0.315$7–$\Omega_m = 0.315$8, comparable to the molecular phase with $\Omega_m = 0.315$9 and $\sim 100$00. Representative kinematics are $\sim 100$01–$\sim 100$02; for the ionized component alone, $\sim 100$03 and $\sim 100$04 (Rupke et al., 2023).

For the outer slow wind, the preferred estimates give $\sim 100$05 and $\sim 100$06–$\sim 100$07, with representative kinematics $\sim 100$08–$\sim 100$09. The corresponding rates are much smaller than in Episode II: $\sim 100$10 and $\sim 100$11. Makani therefore combines a very massive outer ionized reservoir with a relatively small present-day outer-wind mass flux (Rupke et al., 2023).

The shock-radiated power was estimated from

$\sim 100$12

with $\sim 100$13. For Episode II, $\sim 100$14 implies $\sim 100$15, while the mechanical power in the ionized+molecular wind, $\sim 100$16, is sufficient to power the observed line emission. For Episode I, $\sim 100$17 implies $\sim 100$18, whereas $\sim 100$19. The optical analysis therefore concludes that much of the energy driving Episode I shocks must reside in an unseen hotter phase or has propagated to larger CGM radii (Rupke et al., 2023).

The driving mechanism is cast in terms of the momentum boost,

$\sim 100$20

Using $\sim 100$21, so that $\sim 100$22, the inner outflow has a momentum boost of $\sim 100$23. The interpretation advanced for Makani is a momentum-driven flow supplied jointly by hot ejecta and radiation pressure from the compact Eddington-limited starburst, without requiring a clear AGN (Rupke et al., 2023).

5. Ultraviolet line emission and coronal cooling in the CGM

Deep HST/ACS-SBC imaging with the F150LP and F165LP long-pass filters provided a differential narrow-band method for separating O VI from Ly$\sim 100$24. At Makani’s redshift, O VI $\sim 100$25 shifts to $\sim 100$26 and $\sim 100$27 Å and is included in F150LP but excluded from F165LP, whereas Ly$\sim 100$28 at $\sim 100$29 Å lies in both filters. The critical practical result is that F150LP traces O VI+Ly$\sim 100$30, F165LP traces Ly$\sim 100$31, and the filter difference isolates O VI with minimal Ly$\sim 100$32 leakage. After dark-current control, background modeling, drizzling to the KCWI scale, and Voronoi binning, O VI and Ly$\sim 100$33 emission were detected across the [O II] nebula with similar morphology and extent, out to $\sim 100$34 kpc (Ha et al., 25 Mar 2025).

The morphology is highly constraining. The F150LP emission reproduces the [O II] hourglass, including four lobes and the northern cavity, to $\sim 100$35–$\sim 100$36, or $\sim 100$37–$\sim 100$38 kpc. The inner $\sim 100$39–$\sim 100$40 region is dominated by continuum and/or Ly$\sim 100$41 from the compact starburst, but beyond $\sim 100$42 the emission is line-dominated. In the extended nebula the measured count-rate ratio is $\sim 100$43 at $\sim 100$44–$\sim 100$45 and $\sim 100$46 at $\sim 100$47–$\sim 100$48, implying that O VI contributes $\sim 100$49–$\sim 100$50 of F150LP counts in the inner extended zone and $\sim 100$51–$\sim 100$52 in the outer zone (Ha et al., 25 Mar 2025).

The adopted line-separation relations are

$\sim 100$53

with

$\sim 100$54

$\sim 100$55

and

$\sim 100$56

After correction for Milky Way extinction with $\sim 100$57, the extended-nebula ratios are $\sim 100$58 at $\sim 100$59–$\sim 100$60 and $\sim 100$61 at $\sim 100$62–$\sim 100$63. Summing the detected bins gives $\sim 100$64, while model extrapolation gives $\sim 100$65 integrated to $\sim 100$66 kpc and $\sim 100$67 integrated to infinity; the adopted value is $\sim 100$68. This is comparable to $\sim 100$69 (Ha et al., 25 Mar 2025).

The radial surface brightness of O VI is well described by

$\sim 100$70

with $\sim 100$71 (fixed), $\sim 100$72 kpc, $\sim 100$73, and $\sim 100$74 kpc. The O VI half-light radius is $\sim 100$75–$\sim 100$76 kpc, similar to [O II] at $\sim 100$77 kpc. Only $\sim 100$78 of the total O VI flux lies within the compact central $\sim 100$79 aperture, while the remaining $\sim 100$80 is extended, which the UV study takes as evidence for in-situ O VI excitation rather than dominant resonant scattering of nuclear light (Ha et al., 25 Mar 2025).

Physically, O VI is interpreted as tracing radiative cooling at $\sim 100$81 K in hot-cold interfaces, where the $\sim 100$82 K CGM or hot wind exchanges mass with $\sim 100$83 K clouds. Using a metal-line cooling coefficient $\sim 100$84 and

$\sim 100$85

the paper notes that $\sim 100$86 and $\sim 100$87 imply $\sim 100$88 if the O VI mass equals the ensemble halo O VI mass of $\sim 100$89. With $\sim 100$90 Myr for O VI-bearing coronal gas, the resulting oxygen-phase cooling rate is $\sim 100$91, and scaling to total gas gives $\sim 100$92 for $\sim 100$93. This suggests very strong mass exchange and cloud growth in the outer wind (Ha et al., 25 Mar 2025).

6. Warm dust, PAHs, and dust survival to tens of kiloparsecs

JWST NIRCam and MIRI imaging exploited a coincidental redshift match between Makani’s PAH features and standard imaging filters. At $\sim 100$94, PAH $\sim 100$95m falls in NIRCam F480M, PAH $\sim 100$96m in MIRI F1130W, PAH $\sim 100$97m in MIRI F1800W, H$\sim 100$98 $\sim 100$99–$M_* \approx 10^{11.1}\,M_\odot$00 S(1) $M_* \approx 10^{11.1}\,M_\odot$01m in MIRI F2550W, and Pa$M_* \approx 10^{11.1}\,M_\odot$02 in NIRCam F187N. “Off-band” continuum windows were provided by MIRI F770W and F2100W. After PSF subtraction with STPSF and photutils PSFPhotometry, PAH $M_* \approx 10^{11.1}\,M_\odot$03m was detected to $M_* \approx 10^{11.1}\,M_\odot$04 kpc, PAH $M_* \approx 10^{11.1}\,M_\odot$05m to $M_* \approx 10^{11.1}\,M_\odot$06 kpc, and PAH $M_* \approx 10^{11.1}\,M_\odot$07m to $M_* \approx 10^{11.1}\,M_\odot$08 kpc (Veilleux et al., 10 Jul 2025).

The spatial relation to other phases is important. Warm dust extends well beyond the inner CO(2–1) and Mg II emission at $M_* \approx 10^{11.1}\,M_\odot$09 kpc, but does not reach as far as the outer [O II] nebula at up to $M_* \approx 10^{11.1}\,M_\odot$10 kpc. Within $M_* \approx 10^{11.1}\,M_\odot$11 kpc, the flux ratios F1130W/CO(2–1), F1800W/CO(2–1), F1130W/Mg II, and F1800W/Mg II remain roughly constant, suggesting co-spatial dust with cool gas in the inner wind at similar relative strengths. Beyond $M_* \approx 10^{11.1}\,M_\odot$12 kpc, PAHs persist where CO and Mg II are not detected in the current ALMA/KCWI data. A NW halo cloud complex is evident in F1130W and F1800W out to $M_* \approx 10^{11.1}\,M_\odot$13–$M_* \approx 10^{11.1}\,M_\odot$14 kpc, and the extended emission is clumpy and asymmetric (Veilleux et al., 10 Jul 2025).

The primary quantitative dust diagnostic is

$M_* \approx 10^{11.1}\,M_\odot$15

Measured values are $M_* \approx 10^{11.1}\,M_\odot$16 in the nucleus, $M_* \approx 10^{11.1}\,M_\odot$17 in the inner halo at $M_* \approx 10^{11.1}\,M_\odot$18–$M_* \approx 10^{11.1}\,M_\odot$19 kpc, $M_* \approx 10^{11.1}\,M_\odot$20 in the outer halo cloud CGM-E, and $M_* \approx 10^{11.1}\,M_\odot$21 in CGM-W. Because the $M_* \approx 10^{11.1}\,M_\odot$22 and $M_* \approx 10^{11.1}\,M_\odot$23m bands are stronger in neutral and larger PAHs, whereas $M_* \approx 10^{11.1}\,M_\odot$24m strengthens in ionized or smaller PAHs under harder radiation fields, the observed radial decline in $M_* \approx 10^{11.1}\,M_\odot$25 was modeled as indicating decreasing starlight intensity, decreasing PAH sizes, and increasing PAH ionization fractions with increasing distance from the nucleus (Veilleux et al., 10 Jul 2025).

The modeling compared measured filter ratios to the PAH+dust spectral library of Draine (2021), redshifted to $M_* \approx 10^{11.1}\,M_\odot$26 and convolved with the exact JWST throughputs, across starlight intensity $M_* \approx 10^{11.1}\,M_\odot$27, PAH size distribution $M_* \approx 10^{11.1}\,M_\odot$28, and PAH ionization fraction $M_* \approx 10^{11.1}\,M_\odot$29. In the nucleus, high $M_* \approx 10^{11.1}\,M_\odot$30 and elevated continuum ratios are consistent with large $M_* \approx 10^{11.1}\,M_\odot$31, larger grains, and lower PAH ionization fractions. In the inner and outer halo, lower $M_* \approx 10^{11.1}\,M_\odot$32 values require lower $M_* \approx 10^{11.1}\,M_\odot$33 with $M_* \approx 10^{11.1}\,M_\odot$34, smaller average PAH sizes than the standard distribution, and elevated PAH ionization fractions relative to the nucleus. No dust temperatures or masses were reported (Veilleux et al., 10 Jul 2025).

The dust-survival problem is acute. For representative values $M_* \approx 10^{11.1}\,M_\odot$35 kpc and $M_* \approx 10^{11.1}\,M_\odot$36, the travel time is $M_* \approx 10^{11.1}\,M_\odot$37 Gyr, consistent with $M_* \approx 10^{11.1}\,M_\odot$38 yr. By contrast, the adopted thermal sputtering time,

$M_* \approx 10^{11.1}\,M_\odot$39

with $M_* \approx 10^{11.1}\,M_\odot$40, gives $M_* \approx 10^{11.1}\,M_\odot$41–$M_* \approx 10^{11.1}\,M_\odot$42 yr for PAH-sized grains with $M_* \approx 10^{11.1}\,M_\odot$43–$M_* \approx 10^{11.1}\,M_\odot$44m under the densities and temperatures adopted for the hot phase, far shorter than the travel time. The observed radial decline in $M_* \approx 10^{11.1}\,M_\odot$45, reduced F480M detections, and low halo continuum are therefore interpreted as evidence that PAHs survive to $M_* \approx 10^{11.1}\,M_\odot$46–$M_* \approx 10^{11.1}\,M_\odot$47 kpc but are eroded and processed during transport. A possible survival mechanism proposed in the JWST study is shielding in cloud-wind mixing layers, where dust can survive for $M_* \approx 10^{11.1}\,M_\odot$48 if condensation and cooling are efficient (Veilleux et al., 10 Jul 2025).

The JWST detections also revise the earlier radial dust picture. Optical slit data had found $M_* \approx 10^{11.1}\,M_\odot$49 by $M_* \approx 10^{11.1}\,M_\odot$50 kpc along the slit, but the PAH maps directly reveal dust in the outer warm-ionized wind and note that the slit may have missed parts of the NW PAH clouds. This is not a contradiction in the strict sense; it is a demonstration that slit-based extinction measurements did not fully sample the three-dimensional dust distribution (Veilleux et al., 10 Jul 2025).

7. Uncertainties, interpretive limits, and broader significance

Each observational window carries substantial systematics. In the optical, uncertainties arise from extinction corrections, especially at large radius, from the $M_* \approx 10^{11.1}\,M_\odot$51 dex RMS scatter in the H$M_* \approx 10^{11.1}\,M_\odot$52–[O II] bootstrap, from difficult low-S/N density measurements in the outer wind, from the adopted two-phase precursor/post-shock density model, from shock-model degeneracies involving precursor fractions and magnetic parameter, and from the geometry and deprojection assumptions used to derive $M_* \approx 10^{11.1}\,M_\odot$53, $M_* \approx 10^{11.1}\,M_\odot$54, and $M_* \approx 10^{11.1}\,M_\odot$55. Flux calibration differences between ESI and SDSS/KCWI and the assumed $M_* \approx 10^{11.1}\,M_\odot$56 subtraction of star-forming H$M_* \approx 10^{11.1}\,M_\odot$57 in the inner nebula are additional systematic terms (Rupke et al., 2023).

In the ultraviolet, the main limitations are internal dust attenuation, which is poorly constrained spatially in the UV; Ly$M_* \approx 10^{11.1}\,M_\odot$58 radiative transfer, which can depress intrinsic O VI/Ly$M_* \approx 10^{11.1}\,M_\odot$59 ratios by adding scattered Ly$M_* \approx 10^{11.1}\,M_\odot$60; the noise-dominated nature of the pure F150LP–F165LP difference image beyond $M_* \approx 10^{11.1}\,M_\odot$61; residual dark-current gradients in ACS-SBC at extremely low surface brightness; and the possibility that up to $M_* \approx 10^{11.1}\,M_\odot$62 of the total O VI light could arise from dust-scattered UV continuum if the entire inner exponential were due to scattering. The UV analysis therefore treats shock ionization as plausible rather than uniquely established, with the favored shock parameters depending sensitively on extinction and Ly$M_* \approx 10^{11.1}\,M_\odot$63 transport (Ha et al., 25 Mar 2025).

In the infrared, the dominant issues are PSF-subtraction residuals, especially at long wavelength; lower-than-expected F2550W sensitivity; the fact that filters integrate across broad PAH features and were analyzed without explicit continuum subtraction; and the possibility that current ALMA/KCWI limits are too shallow to exclude co-spatial cool gas beyond $M_* \approx 10^{11.1}\,M_\odot$64 kpc. The JWST study also notes that radial $M_* \approx 10^{11.1}\,M_\odot$65 trends, while favoring erosion and increased ionization, could be influenced by varying radiation hardness or local shock conditions (Veilleux et al., 10 Jul 2025).

Despite these caveats, Makani occupies an unusual position in feedback studies. The optical work describes it as the poster child of a galactic wind on scales of the circumgalactic medium; the UV work presents the first imaging detection of spatially extended O VI emission coextensive with a $M_* \approx 10^{11.1}\,M_\odot$66 kpc wind in a massive galaxy’s CGM and only the second resolved O VI emission halo imaged for any galaxy; and the JWST work reports direct PAH detections to $M_* \approx 10^{11.1}\,M_\odot$67 kpc, providing strong evidence that ejected dust can survive to CGM scales while being processed in transit (Rupke et al., 2023, Ha et al., 25 Mar 2025, Veilleux et al., 10 Jul 2025).

Taken together, the Makani data define a physically coherent but still incomplete feedback case study. Episode II is a fast, massive, dusty, multiphase inner-halo wind whose warm ionized mass and outflow rate are comparable to those of the molecular phase and whose momentum budget is consistent with a momentum-driven flow from a compact Eddington-limited starburst. Episode I is an older, slower CGM-scale outflow containing a large ionized reservoir, coronal O VI interfaces, and PAH-bearing dust, but with evidence that much of its energy and momentum reside in a hotter phase or at still larger radii. This suggests that Makani is best understood not as a single outflow snapshot but as a resolved feedback sequence linking compact-starburst launch physics, shock-powered ionization, multiphase mass exchange, CGM cooling, and dust transport over $M_* \approx 10^{11.1}\,M_\odot$68 yr.