How probable is the Ly- α \alpha damping wing in the spectrum of the redshift z = 5.9896 quasar ULAS J0148+0600? Fiona Sawyer, James S. Bolton, George D. Becker, Luke Conaboy, Martin G. Haehnelt, Laura Keating, Girish Kulkarni, Ewald Puchwein Abstract The shape of the Ly- α \alpha transmission in the near zone of the redshift z = 5.9896 z=5.9896 quasar ULAS J0148 + + 0600 (hereafter J0148) is consistent with a damping wing arising from an extended neutral hydrogen island in the diffuse intergalactic medium (IGM). Here we use simulations of late-ending reionisation from Sherwood-Relics to assess the expected incidence of quasars with Ly- α \alpha and Ly- β \beta absorption similar to the observed J0148 spectrum. We find a late end to reionisation at z = 5.3 z=5.3 is a necessary requirement for reproducing a Ly- α \alpha damping wing consistent with J0148. This occurs in ∼ 3 \sim 3 per cent of our simulated spectra for an IGM neutral fraction ⟨ x HI ⟩ = 0.14 \langle x_{\rm HI}\rangle=0.14 at z = 6 z=6 . However, using standard assumptions for the ionising photon output of J0148, the a priori probability of drawing a simulated quasar spectrum with a Ly- α \alpha damping wing profile and Ly- α \alpha near zone size that simultaneously match J0148 is low, p < 10 − 2 p<10^{-2} . This may indicate that the ionising emission from J0148 is variable on timescales t < 10 5 ​ yr t<10^{5}\rm\,yr , or alternatively that the Ly- α \alpha transmission in the J0148 near zone is impacted by the transverse proximity effect from nearby star-forming galaxies or undetected quasars. We also predict the IGM temperature should be T ∼ 4 × 10 4 ​ K T\sim 4\times 10^{4}\rm\,K within a few proper Mpc of the Ly- α \alpha near zone edge due to recent H I \,\rm\scriptstyle I and He II \,\rm\scriptstyle II photo-heating. Evidence for enhanced thermal broadening in the Ly- α \alpha absorption near the damping wing edge would provide further evidence that the final stages of reionisation are occurring at z < 6 z<6 . 2025 How probable is the Ly- \alpha damping wing in the spectrum of the redshift z = 5.9896 quasar ULAS J0148+0600? – C 1 Introduction Recent high redshift observations from JWST are challenging our understanding of galaxy formation and the early stages of reionisation ( Adamo et al., 2024 , see e.g.,) . The discovery of numerous high redshift galaxies – now spectroscopically confirmed up to z\sim 14 ( Adams et al., 2023 ; Carniani et al., 2024 ) – that efficiently produce ionising photons has led to a reassessment of the photon budget for reionisation ( Atek et al., 2024 ; Simmonds et al., 2024 ; Begley et al., 2025 ; Muñoz et al., 2024 ) . In addition, a large number of intrinsically faint AGN at z>5 ( Harikane et al., 2023 ; Maiolino et al., 2024 ) along with a newly discovered population of ’Little Red Dots’ ( Labbé et al., 2023 ; Matthee et al., 2024 ) has led to renewed interest in the role that black hole accretion plays in setting the reionisation photon budget ( Dayal et al., 2024 ; Madau et al., 2024 ; Asthana et al., 2024 ) . At the same time, much recent progress has been made in our understanding of the final stages of reionisation. While the first observations of intergalactic Ly- \alpha transmission and the Gunn & Peterson (1965) trough in z\simeq 6 quasars implied that reionisation was largely complete by z=6 ( Fan et al., 2006 , e.g.,) , more recent observations – including notably the XQR-30 survey ( D’Odorico et al., 2023 ) – have delivered high quality data that is now firmly challenging this view. This includes confirmation that a fully-ionised intergalactic medium appears to be inconsistent with observed fluctuations in the Ly- \alpha forest transmission at z<6 ( Becker et al., 2015 ; Eilers et al., 2018 ; Bosman et al., 2018 ; Bosman et al., 2022 ) , the rapid evolution of the mean free path of ionising photons at 5\leq z\leq 6 ( Becker et al., 2021 ; Bosman et al., 2021 ; Zhu et al., 2023 ; Davies et al., 2024 ) , and an incidence of rare dark gaps in the Ly- \alpha and Ly- \beta forest at z<6 that is consistent with at least some fully neutral hydrogen remaining in the diffuse IGM ( Becker et al., 2015 ; Zhu et al., 2022 ; Jin et al., 2023 ) . As speculated by Lidz et al. (2007) and Mesinger (2010) , these data appear to be better reproduced by theoretical models where reionisation ends late ( z\sim 5.3 ) and neutral hydrogen islands persist in the diffuse IGM at z<6 ( Kulkarni et al., 2019 ; Keating et al., 2020a ; Keating et al., 2020b ; Nasir & D’Aloisio, 2020 ) . Only very recently, however, has the first possible direct evidence for neutral islands in the IGM at z<6 been presented. It was already known that the spectrum of the z=5.9896 quasar ULAS J0148 + 0600 (hereafter J0148) contained a remarkably long and dark Ly \alpha trough, with an extent of \sim 110h^{-1} cMpc ( Becker et al., 2015 ) . However, the physical origin of this trough was uncertain; Ly \alpha absorption saturates for neutral hydrogen fractions as small as x_{\rm HI}\sim 10^{-4} . Although some coeval Ly- \beta transmission spikes were present in the J0148 spectrum presented by Becker et al. (2015) – implying at least some of the long Ly- \alpha trough remains highly ionised – the extended regions with no Ly- \alpha and Ly- \beta transmission could be due to neutral islands or an ionised IGM. In this context, Becker et al. (2024) recently reexamined the Ly \alpha transmission at the red edge of the J0148 trough – corresponding to the edge of the J0148 Ly- \alpha near zone – using an improved method for removing the quasar continuum. They determined that the shape of the transmission is consistent with the damping wing profile expected from either an extended neutral island or a compact absorber (i.e., a damped Ly- \alpha absorption system, or DLA) at the blue edge of the Ly- \alpha near zone. The DLA interpretation was disfavoured by Becker et al. (2024) , however, due to the lack of any corresponding metal lines and/or an associated galaxy. Subsequently, further weight has been added to the damping wing interpretation by the detection of damping wing-like features in stacked dark gaps at 5.5\leq z\leq 6 ( Zhu et al., 2024 ; Spina et al., 2024 ) , following a method originally proposed by Malloy & Lidz (2015) . These findings therefore raise the important question of how common the damping wing presented in Becker et al. (2024) is within late reionisation models. While it has only been observed for an individual line of sight in J0148, detailed simulations of inhomogeneous reionisation that resolve the Ly- \alpha forest can provide a far larger set of spectra for analysis under a variety of different model assumptions. In this work we therefore use the Sherwood-Relics simulations of inhomogeneous reionisation ( Puchwein et al., 2023 ) coupled with a 1D radiative transfer code that follows the ionising emission from quasars ( Bolton & Haehnelt, 2007 ) to investigate the likelihood of observing J0148 for a range of different reionisation histories and quasar emission models. Our approach extends the earlier modelling performed by Becker et al. (2024) by using a self-consistent model for patchy reionisation and the local photo-ionisation and heating of the IGM by the quasar. We also examine the properties of neutral islands in the models that are consistent with the J0148 spectrum, and test the efficacy of the summary statistic used by Becker et al. (2024) for identifying the damping wing signature. This paper is organised as follows. In Section 2 , we describe the numerical methods used throughout the paper to create our simulated data set. In Section 3 we discuss the probability that a spectrum with a damping wing signature like J0148 is found within the Sherwood-Relics simulations. We then examine the physical properties of the IGM around the edge of the quasar near zone in Section 4 , and assess the Becker et al. (2024) summary statistic in Section 5 . Finally, our conclusions are outlined in Section 6 . 2 Numerical modelling 2.1 Hydrodynamical simulations of inhomogeneous reionisation The cosmological hydrodynamical simulations used in this work are from the Sherwood-Relics suite, described in detail in Puchwein et al. (2023) . The Sherwood-Relics simulations model the IGM in high resolution during and after hydrogen reionisation using a modified version of the cosmological smoothed particle hydrodynamics code P-Gadget-3 ( Springel, 2005 ) . A novel hybrid radiative transfer (RT) scheme is also used to capture the hydrodynamical effects of inhomogeneous reionisation. The radiative transfer is followed using aton ( Aubert & Teyssier, 2008 ) , where the luminosity of the ionising sources are proportional to halo mass, and the minimum mass of ionising sources is M_{\rm h}>10^{9}h^{-1}\rm{M_{\odot}} . The ionising photons have mean energy 18.6~{\rm eV} , corresponding to a blackbody spectrum with temperature T=4\times 10^{4}~{\rm K} . Further technical details may be found in Puchwein et al. (2023) . Table 1 : Table detailing the properties of each of the simulations used in this work. From left to right, the table lists: the simulation name used in this work, the box size, the number of particles, the dark matter particle mass, the gas particle mass, the end redshift of reionisation (defined as the redshift where the volume averaged neutral hydrogen fraction first falls below \langle x_{\rm HI}\rangle=10^{-3} ), the mid-point of reionisation when \langle x_{\rm HI}\rangle=0.5 , and finally \langle x_{\rm HI}\rangle and the volume averaged gas temperature, \langle T\rangle , at z=6.0 . Name Box size N_{\rm part} \rm{M_{dm}} \rm{M_{gas}} z_{\rm r} z_{\rm mid} \langle x_{\rm HI}(z=6)\rangle \langle T(z=6)\rangle [ h^{-1} cMpc] [ h^{-1}\rm{M_{\odot}} ] [ h^{-1}\rm{M_{\odot}} ] [ 10^{4} K] 160-2048-zr53 160 2\times 2048^{3} 3.44 \times 10^{7} 6.38 \times 10^{6} 5.3 7.2 0.140 10.8 40-2048-zr53 40 2\times 2048^{3} 5.37 \times 10^{5} 9.97 \times 10^{4} 5.3 7.2 0.142 10.2 40-2048-zr57 40 2\times 2048^{3} 5.37 \times 10^{5} 9.97 \times 10^{4} 5.7 7.5 0.053 11.2 Here we use the simulations with box sizes 40 h^{-1}\,\rm cMpc and 160 h^{-1}\,\rm cMpc , with 2\times 2048^{3} dark matter and gas particles. The models used are summarised in Table 1 . Our fiducial model uses the 160 h^{-1}\,\rm cMpc box as this is better able to capture the large scale structure of the IGM and the sizes of neutral islands / ionised bubbles. In this model reionisation completes at z_{\rm r}=5.3 . This late-ending reionisation history was chosen as it provides good agreement with the distribution of the Ly- \alpha forest effective optical depth at 50.4 the value of R_{\rm wing} becomes independent of \langle\Delta F\rangle . Interestingly, the quasar models from which we are most likely to observe a J0148-like spectrum now occur for a higher luminosity / absolute magnitude M_{1450} (4 L_{0} ), or a harder quasar spectrum ( \alpha_{s}=0.5 ), or for a longer optically/UV bright lifetime ( t_{\rm q}=10^{7.6}\rm\,yr ). The higher luminosity or harder spectrum both increase the \dot{N} of the quasar compared to the fiducial model, and so are capable of generating larger R_{\rm wing} values. Similarly, a longer optically/UV bright lifetime can extend the Ly- \alpha near zone around the quasar. However, it seems unlikely that the EUV spectral slope of J0148 will be significantly harder than the composite quasar continuum used in Becker et al. (2024) , or that the UV absolute magnitude is significantly in error. Morey et al. (2021) find a typical optically/UV bright lifetime of t_{\rm q}\sim 10^{6}\rm\,yr in their analysis of z\sim 6 quasar near zone sizes, which is over an order of magnitude shorter than the model that reproduces the J0148 spectrum most readily. On the other hand, for 22 quasars from the XQR-30 survey Satyavolu et al. (2023b) reported a broad range of near-zone sizes at z\sim 6 that are consistent with optically/UV bright lifetimes of t_{\rm q}\sim 10^{4} – 10^{8}\rm\,yr . A quasar lifetime of t_{\rm q}=10^{7.6} yr could therefore partially explain the J0148 spectrum, although the probability of producing a damping wing similar to J0148 is still small, p<10^{-2} . This suggests that either (i) quasars with Ly- \alpha near zone absorption profiles and giant Ly- \alpha absorption troughs similar to J0148 at z=6 are rare (with a priori probabilities p<10^{-2} ), (ii) the damping wing in J0148 is instead due to an extremely low metallicity DLA 5 5 5 Note that we do not model compact absorbers with damping wings in our simulations, and are therefore unable to directly assess how likely this is. , although as mentioned earlier this was already argued against by Becker et al. (2024) or (iii) some additional physical effect is missing from our models. In the latter case, one possibility is that J0148 may have a variable light curve ( Davies et al., 2020 ; Satyavolu et al., 2023a ; Šoltinský et al., 2023 , see e.g.) and until recently it was much brighter. This would require variability on a timescale comparable to or shorter than the equilibriation timescale, t_{\rm eq} , for the H \,\rm\scriptstyle I in the Ly- \alpha near zone, where t_{\rm eq}=\frac{x_{\rm HI,\rm\,eq}}{n_{\rm e}\alpha_{A}(T)}\simeq\frac{10^{5.0}\rm\,yr}{\Delta}\left(\frac{x_{\rm HI,\rm\,eq}}{10^{-4}}\right)\left(\frac{T}{10^{4}\rm\,K}\right)^{0.72}\left(\frac{1+z}{7}\right)^{-3}. (4) Here x_{\rm HI,\,eq} is the H \,\rm\scriptstyle I fraction in ionisation equilibrium, \alpha_{\rm A}=4.06\times 10^{-13}\mathrm{\,cm^{3}\,s^{-1}}(T/10^{4}\rm\,K)^{-0.72} is the case-A recombination coefficient and n_{\rm e}=1.158n_{\rm H} for fully ionised hydrogen and helium. Alternatively, as already noted by Becker et al. (2024) , another possibility is the transverse proximity effect. Enhanced ionisation may occur if there are additional ionising sources near the location of the blue edge of the J0148 Ly- \alpha near zone (see e.g. Eilers et al., 2024 , for possible candidates at z\simeq 5.92 ) . The sources could also potentially be undetected quasars, which are expected to cluster on the relevant scales ( Shen et al., 2007 ; Eilers et al., 2024 ) but may not be detectable due to the anisotropic nature of quasar emission, particularly if the quasar is intrinsically faint. Quasar variability and/or duty cycle could also play a role in limiting detectability. 4 Physical properties of gas associated with the damping wing Figure 5 : The median line of sight gas density, \Delta=\rho/\langle\rho\rangle , H \,\rm\scriptstyle I fraction, He \,\rm\scriptstyle II fraction, He \,\rm\scriptstyle III fraction and gas temperature, for the sub-set of sight-lines with \langle\Delta F\rangle=0.15\pm 0.05 drawn from a total of 6000 simulated spectra. The left column shows the fiducial model (averaged over 296 sight-lines), while the right column shows the M_{1450}=-28.9 model (averaged over 493 sight-lines). The distances on the horizontal axes have been renormalised such that R=R_{\rm wing} is zero for all sight-lines, where R is the distance from the quasar and R_{\rm wing} is the distance at which the damping wing template first exceeds F=0.01 . Medians are plotted as the solid lines, and the 68 per cent interval around the median is displayed as the shaded region. The quasar model corresponds to the blue curves, while the red curves show the same quantities before the quasar has turned on (note again that x_{\rm HeIII}=0 initially). Neutral hydrogen islands are evident at R-R_{\rm wing}<1\rm\,pMpc , and elevated gas temperatures due to H \,\rm\scriptstyle I and He \,\rm\scriptstyle II photo-heating by the hard quasar spectrum are present at R-R_{\rm wing}<0\rm\,pMpc (i.e., behind the H \,\rm\scriptstyle II and He \,\rm\scriptstyle III ionisation fronts). Table 3 : Median island properties for islands with \langle\Delta F\rangle=0.15\pm 0.05 . As in Table 2 , the model and quasar properties are listed in columns 1-5. The fiducial model is in bold. Column 6 shows the number of lines of sight with \langle\Delta F\rangle=0.15\pm 0.05 , and these are used to construct the median profiles in Figure 5 . Columns 7 and 8 give the properties of the median neutral island at the edge of the H \,\rm\scriptstyle II ionisation front in the stacked lines of sight. These are the size of the neutral island in pMpc, where the neutral island is defined by the length of the continuous region with x_{\rm HI}>0.5 , and the logarithm of the H \,\rm\scriptstyle I column density for the neutral island, calculated over the distance specified in column 7. Hydrodynamical \langle x_{\rm HI}\rangle \rm{M_{1450}} \rm{\alpha_{s}} \rm{t_{\rm q}} [yr] # candidates Island size \rm{log_{10}(N_{HI}/\rm cm^{-2})} simulation name [pMpc] 160-2048-zr53 0.140 -27.4 1.5 10^{7} 296 0.975 20.01 160-2048-zr53 0.140 -27.4 0.5 10^{7} 332 1.263 20.08 160-2048-zr53 0.140 -27.4 2.5 10^{7} 210 0.782 19.92 160-2048-zr53 0.140 -27.4 1.5 10^{5} 485 1.276 20.11 160-2048-zr53 0.140 -27.4 1.5 10^{6} 450 1.049 20.03 160-2048-zr53 0.140 -27.4 1.5 10^{7.6} 77 0.387 19.56 160-2048-zr53 0.140 -27.4 1.5 10^{8} 5 0.539 19.88 160-2048-zr53 0.140 -26.65 1.5 10^{7} 230 0.864 19.97 160-2048-zr53 0.140 -28.15 1.5 10^{7} 412 1.058 20.05 160-2048-zr53 0.140 -28.9 1.5 10^{7} 493 1.276 20.11 40-2048-zr53 0.142 -27.4 1.5 10^{7} 422 0.243 19.28 40-2048-zr57 0.053 -27.4 1.5 10^{7} 30 1.539 20.12 We now investigate the underlying physical properties of the simulated spectra that return \langle\Delta F\rangle values consistent with J0148. In Figure 5 , for our fiducial model (left column) and bright quasar model ( M_{1450}=-28.9 , right column), following Becker et al. (2024) we take simulated lines of sight with \langle\Delta F\rangle =0.15\pm 0.05 and stack them to find the median and 68 per cent scatter of the underlying physical properties. These are shown in Figure 5 as solid lines and shaded regions respectively. Note the corresponding Ly- \alpha and Ly- \beta transmission profiles are already displayed in Figure 2 . From top to bottom, the panels in Figure 5 show the gas density \Delta=\rho/\langle\rho\rangle , the H \,\rm\scriptstyle I fraction, the He \,\rm\scriptstyle II fraction, the He \,\rm\scriptstyle III fraction and the gas temperature. The gas temperature and ionised fractions are furthermore shown just before the quasar emission was turned on (red curves), and for an optically/UV bright lifetime of t_{\rm q}=10^{7}\rm\,yr (blue curves). Both quasar emission models give similar results. Before the quasar turns on, the lines of sight producing damping wing-like features typically have a large region of neutral hydrogen (initially \sim\,2 pMpc) situated in a region that is slightly more underdense than the surrounding IGM. A very similar result was reported for the simple model used in Becker et al. (2024) (see their figure 13), consistent with their suggestion that the J0148 Ly- \alpha near zone may terminate at a void. The neutral region is also significantly cooler than its partially ionised surroundings because it is yet to be reionised and photo-heated. After the quasar turns on the IGM becomes more highly ionised, and the pre-existing H \,\rm\scriptstyle II region around the quasar host halo – originally created by the ionising emission from clustered galaxies – expands in size. The quasar ionisation front pushes into the neutral region to ionise the hydrogen and doubly ionise the helium. Note also that prior to the quasar turning on, the regions containing ionised hydrogen contained almost all of their helium in singly ionised form, but for t_{\rm q}=10^{7}\rm\,yr the hard photons from the quasar doubly ionise the helium for R0.5 over a region of length \geq 0.5 pMpc to return a positive identification, while the blue curves require a shorter length of \geq 0.1 pMpc. The purity for neutral islands with \geq 0.1 pMpc and \langle\Delta F\rangle \leq 0.15 is 0.97 , although the completeness shows these represent only small fraction of the total neutral islands in the simulated quasar spectra. We lastly examine the effectiveness of the \langle\Delta F\rangle statistic introduced by Becker et al. (2024) at correctly identifying diffuse neutral islands in our simulated quasar spectra. It is expected that the presence of a neutral island close to the edge of the quasar near zone at R_{\rm wing} will produce a characteristic damping wing shape in the spectrum, and hence return a small value of \langle\Delta F\rangle . However, a similar signal could be produced by chance without the presence of a neutral island. To assess the presence of neutral islands in the diffuse IGM near R_{\rm wing} we must first adopt an operational definition for a neutral island. We opt to select regions where x_{\rm{HI}}\geq 0.5 continuously over a distance of either 0.1 or 0.5\rm\,pMpc , where addditionally the red edge of the neutral island must fall between R_{\rm wing} and R_{\rm wing}+1\rm\,pMpc . Multiple choices for the position of the red edge of the neutral island were tested, but this choice was not found to affect our results significantly; most neutral islands that produce a damping wing have a red edge very close to R_{\rm wing} . However, the distance over which we require a neutral region to have x_{\rm HI}\geq 0.5 has a larger effect. In our analysis we therefore use both 0.1 or 0.5\rm\,pMpc . Adopting the larger value of 0.5\rm\,pMpc excludes sight-lines that exhibit multiple small neutral islands that are punctuated by narrow regions of highly ionised IGM. Next, using our 6000 simulated lines of sight drawn from the fiducial model, we may calculate both the purity and completeness of the neutral island sample recovered from the simulated spectra as a function of the \langle\Delta F\rangle statistic. We define the purity as {\rm Purity}=\frac{N({\rm neutral\,island}\land x\leq\langle\Delta F\rangle0.4 , demonstrating there are many lines of sight containing neutral islands that do not exhibit a Ly- \alpha damping wing profile as obvious as J0148 (an example is shown in the right column of Figure 1 ). In summary – assuming that damping wings from compact absorbers can be ruled out – we confirm the method introduced by Becker et al. (2024) should reliably indicate the presence of a neutral island of extent >0.1\rm\,pMpc in the spectrum of J0148 if \langle\Delta F\rangle \leq 0.15 (i.e. if there is a reasonably obvious damping wing profile). However, we also expect this criterion will only catch at most a few per cent of the total population of neutral islands in the vicinity of quasar near-zones. The detection of damping wings from neutral islands in stacked spectra ( Malloy & Lidz, 2015 ; Zhu et al., 2024 ; Spina et al., 2024 , e.g.) may be a more effective approach in these cases. Alternatively, the detection of 21\rm\,cm absorption from the diffuse IGM at z\simeq 6 could also provide evidence for the presence of neutral islands, although the detectability will depend on the strength of the heating in the neutral IGM by the X-ray background ( Šoltinský et al., 2021 ; Šoltinský et al., 2025 ) . 6 Conclusions Becker et al. (2024) recently demonstrated the shape of the Ly- \alpha transmission in the near zone of the z=5.9896 quasar ULAS J0148 + 0600 (hereafter J0148) is consistent with the damping wing profile expected from an extended neutral hydrogen island in the diffuse intergalactic medium (IGM). This represents some of the first direct evidence for an extended region of neutral hydrogen in the diffuse, low density IGM at z<6 ( Zhu et al., 2024 ; Spina et al., 2024 , see also) . In this work, we therefore use simulations of late ending reionisation from Sherwood-Relics ( Puchwein et al., 2023 ) , combined with 1D radiative transfer simulations of the ionising radiation from the quasar ( Bolton & Haehnelt, 2007 ; Šoltinský et al., 2023 ) , to better quantify the expected incidence rate of quasar sight-lines with Ly- \alpha and Ly- \beta absorption that quantitatively match the J0148 spectrum. We also use our simulations to investigate the underlying physical properties of the neutral island implied by J0148 damping wing, and test the efficacy of the summary statistic, \langle\Delta F\rangle introduced by Becker et al. (2024) for identifying diffuse neutral hydrogen in the IGM. We confirm that simulations with reionisation ending late at z=5.3 ( Kulkarni et al., 2019 ; Keating et al., 2020a ; Keating et al., 2020b ; Bosman et al., 2022 , see also) are necessary for reproducing quasar spectra at z\leq 6 that exhibit a damping wing profile at the blue edge of the quasar Ly- \alpha near zone consistent with J0148. Although quasar spectra with \langle\Delta F\rangle matching J0148 are rare ( p(\langle\Delta F\rangle\leq 0.15)\sim 0.03 ) in our late reionisation model, almost all ( 97 per cent) of our simulated quasar spectra with \langle\Delta F\rangle\leq 0.15 contain a neutral island of extent >0.1\rm\,pMpc blueward of the Ly- \alpha near zone. Hence, assuming the damping wing in the spectrum of J0148 does not originate from a compact absorber (see Becker et al., 2024 , for a detailed discussion of this point) , it appears highly likely the damping wing reported by Becker et al. (2024) with \langle\Delta F\rangle=0.15 arises from a neutral hydrogen island in the underdense IGM with approximate extent of \sim 1\,\rm pMpc and total H \,\rm\scriptstyle I column density N_{\rm HI}\sim 10^{20}\rm\,cm^{-2} . Our modelling furthermore predicts the Ly- \alpha transmission at the edge of the Ly- \alpha near zone should consist of recently reionised hydrogen and helium with temperature T\sim 4\times 10^{4}\rm\,K , as suggested by Becker et al. (2024) . Intriguingly, however, we also find that the a priori probability of drawing a quasar line of sight with both a damping wing profile and Ly- \alpha near zone size (defined here as R_{\rm wing}=7\rm\,pMpc ) that matches J0148 is low, p<10^{-2} , even when allowing for an optically/UV bright quasar lifetime of up to t_{\rm q}=10^{8}\rm\,yr . The main difficulty is reproducing a region of sufficiently high transmission at the edge of the Ly- \alpha near zone, given the photo-ionisation rate from the quasar will decline at least as rapidly as \Gamma_{\rm HI}\propto R^{-2} . Matching the J0148 spectrum requires highly underdense gas and significantly elevated gas temperatures at the near-zone edge, and this combination occurs relatively infrequently in our models. The probability of recovering a J0148-like spectrum from our models can be modestly increased if we assume J0148 produces significantly more ionising photons than expected for standard assumptions (e.g., for a harder extreme UV spectrum, \alpha_{\rm s}<1.5 or a brighter UV absolute magnitude, M_{1450}<-27.4 ). However, this would require the composite quasar continuum for J0148 and/or the quasar UV photometry to be significantly in error. It remains possible that J0148 is simply an unusual line of sight; it is the only object to date at z<6 which exhibits evidence for both a damping wing and a giant Ly- \alpha absorption trough. We nevertheless speculate that there could be some additional physical effects missing from our models. One possibility is quasar variability. If J0148 has very recently dimmed on a timescale of t1200\,\rm{km\,s^{-1}} , \langle\Delta F\rangle_{\rm J0148} increases as a consequence of this saturated region. One might therefore ask whether adopting a larger velocity window that includes this absorption, say \sim 1500\rm\,km\,s^{-1} , will change our conclusions regarding the rarity of the J0148 configuration. This is addressed in the right panel of Figure 7 , which shows the fraction of simulated sight-lines, f(\langle\Delta F\rangle0.2 are much larger. This is because the lower resolution models produce smoother transmission, and so the requirement that the simulated spectrum does not exceed the ideal wing template over our calculation region is more likely to be satisfied. This also imples that if the mass resolution were increased further, J0148-like spectra would be even rarer in our models.