WALLABY Pilot Survey: A gas-rich diffuse dwarf on the baryonic Tully Fisher relation

TL;DR

This study analyzes the gas-rich diffuse dwarf KK176 using high-resolution HI data from the WALLABY survey to build a self-consistent kinematic and stellar model anchored to a WISE center. Through a robust 3D tilted-ring analysis plus a stellar mass assessment, the authors perform a mass decomposition that shows KK176 is dark matter dominated at all radii, while placing it on the baryonic Tully-Fisher relation consistent with low-mass dwarfs. The results challenge claims of dark-matter-deficient diffuse systems by illustrating a DM-consistent case among gas-rich dwarfs and UDGs, and underscore the importance of accurate distance, geometry, and uncertainty propagation. The work demonstrates the potential of WALLABY and follow-up surveys to expand the census of diffuse dwarfs and to inform DM and galaxy formation theories across the low-mass regime.

Abstract

Diffuse dwarf galaxies, and particularly ultra diffuse galaxies (UDGs), challenge our understanding of galaxy formation and the role of dark matter due to their large sizes, low surface brightness, and varying dark matter content. In this work, we investigate the gas-rich diffuse dwarf galaxy WALLABY J125956-192430 (aka. KK176) using high-resolution HI data from the WALLABY survey. We produce the most reliable kinematic model for KK176 to date. Using this model, the derived mass decomposition shows that KK176 is dark matter dominated. We also place KK176 on the baryonic Tully-Fisher relation (bTFR), finding that it is consistent with low-mass dwarf galaxies but distinctly different from reported dark matter-deficient UDGs.

Figures (6)

• Figure 1: KK 176 as seen at various wavelengths. Panel A shows the H i component overlayed on a $grz$ composite optical image from DECaLS. The H i color is mapped to the velocity (blue approaching, red receding relative to the estimated $V_{\rm{sys}}$) with the brightness mapped to the intensity. Panels B and C show the integrated WALLABY H i intensity and line-of-sight velocity maps, respectively. Panel D shows the optical image, using a different stretch from Panel A to highlight low surface brightness features. Panel E shows the WISE W1 image. In Panels B-E the magenta circle shows the WALLABY beam (which is the same as the scale line in Panel A) and the yellow contour shows the $\Sigma_{\rm{HI}}=1~\textrm{M}_{\odot}~\rm{pc^{-2}}$ H i surface density contour level of the integrated intensity map.
• Figure 2: Our preferred kinematic model of KK 176 (red curves), in comparison to the pipeline-generated model presented by Deg2024 (blue and green curves). Shaded regions represent the uncertainties on the curve of the corresponding color. Panel A shows rotation curves, and panel B shows surface density profiles. The horizontal dotted line in panel B shows an H i surface density of $\Sigma_{\rm{HI}}=1~\textrm{M}_{\odot}/\rm{pc}^2$, from which we define $R_{\rm{HI}}$, represented by the vertical dotted line line (see Sec. \ref{['sec:modelling']}). The horizontal dotted line in panel A shows $V_{\rm{HI}}$. The green curve (= DC) is the D24 model corrected to KK 176's TRGB distance $d$. The blue curve (=DIC) is the same initial model when corrected for both $d$ and $i$ from our preferred model (see Table \ref{['Fluffy_Table']}).
• Figure 3: Mass decomposition of KK 176, showing the contributions to the circular velocity $V_{\rm{circ}}$ (blue curve) from different components (see Sec. \ref{['ssec:massmodel']}). The contributions from the stars and the gas are shown in green and orange, respectively. The black curve shows the implied DM contribution (see Eq. \ref{['eq:massmodel']}). The shaded regions surrounding each curve show standard deviations of the model components, estimated from bootstrap resampling (see Sec. \ref{['ssec:KinModel']}). The vertical and horizontal dotted red lines show $R_{\text{H\,\sc{i}}}$ and $V_{\rm{circ}}(R_{\text{H\,\sc{i}}})$ respectively.
• Figure 4: KK 176's position on the bTFR, along with other gas-rich galaxy samples from the literature. Both panels show the same data and relation, using either a logarthmic (panel A) or linear (panel B) stretch. Our measurements of KK 176 are shown as a blue star. The green downward triangle shows the previous estimate from the pipeline-generated model in Deg2024 with a cosmic flow distance estimate instead of the TRGB distance $d$ adopted here. The SPARC data points from Lelli2019, the WALLABY points from Deg2024, and the UDGs from ManceraPina2019 (and the updated measurement for AGC 114905 from Mancera_Pina_2024) are all constructed using kinematic models. By contrast the low mass McQuinn2022 measurements are from position-velocity (PV) diagram modelling (including an asymmetric drift correction). Similarly the rotation velocity for Leo P is also determined from the PV diagram (via a slightly different technique), but does not include an asymmetric drift correction Giovanelli2013. The SPARC and WALLABY points only show galaxies with $i\ge40^{\circ}$ (in accordance with (Deg+ 2025)). Due to their lower sample sizes, we show the full ManceraPina2019 and McQuinn2022 samples, but their galaxies with $i<40^{\circ}$ are shown as open symbols. We show the bTFR scaling relation fit and associated scatter from Deg2024 as the dashed grey line and shaded region, and its extrapolation to the low mass regime as the dotted red line and shaded region. Note that this fit is based solely on the grey WALLABY data points. KK 176, and the majority of the other systems at the low-mass end of the bTFR, are consistent with the Deg2024 estimate within its scatter.
• Figure 5: Pairs of H i maps (left panels) and data -- model residuals for our preferred kinematic model (right panels) of Fluffy, labeled by each channel's heliocentric H i recessional velocity. In all panels, the green cross shows our preferred model center measured from the WISE data (see Sec. \ref{['sec:modelling']}). In the left panels, the cyan circle shows the WALLABY beam, and the magenta and orange lines show isodensity contours of our preferred model at $2\sigma = 5.2\,$mJy/beam and $5\sigma = 13\,$mJy/beam, respectively, where $\sigma$ is the RMS map noise. In the right panels, the solid and dashed lines show data -- model residuals at the $+2\sigma$ and $-2\sigma$ levels, respectively.