Dual Role of Accretion Disk Winds as X-Ray Obscurers and UV Line Absorbers in AGN

X-ray obscuration of active galactic nuclei (AGNs) is considered in the context of ionized winds of stratified structure launched from accretion disks. We argue that a Compton-thick layer of a large-scale disk wind can obscure continuum X-rays and also lead to broad UV absorption, such as in the blue wing of C iv; the former originates from the inner wind and the latter from the outer wind, as a dual role. Motivated by a number of lines of observational evidence showing strong AGN obscuration phenomena in Seyfert 1 AGNs such as NGC 5548, we demonstrate in this work, by utilizing a physically motivated wind model coupled to post-process radiative transfer calculations, that an extended disk wind under certain physical conditions (e.g., morphology and density) could naturally cause a sufficient obscuration qualitatively consistent with UV/X-ray observations. Predicted UV/X-ray correlation is also presented as a consequence of variable spatial size of the wind in this scenario.

Outflows are among the most fundamental features almost ubiquitously seen in active galactic nuclei (AGNs) being manifested in UV/X-ray spectra in the form of blueshifted absorption lines (e.g.Crenshaw, Kraemer & George 2003;Tombesi et al. 2010;Arav et al. 2015).Galactic outflows beyond parsec-scale in the host galaxies have also been seen in the form of molecular winds, most likely having a physical origin conncted to the hotter and faster innermost AGN winds (e.g.Faucher-Giguère & Quataert 2012;Wagner et al. 2013;Tombesi et al. 2015).In particular, ionized outflows seen in X-ray, conventionally termed as warm absorbers (e.g.Blustin et al. 2005;Steenbrugge et al. 2005;McKernan et al. 2007;Laha et al. 2014Laha et al. , 2016)), may be related to accretion disks, narrow line region or torus around supermassive black holes (BHs), although a definitive identification of their launching mechanisms remains elusive to date (i.e.thermal, radiation and magnetic).These accretion-disk winds of a line-of-sight velocity v consist of a variety of chemical elements at different ionization states.Particularly in X-rays, most notable is hydrogen/helium-like absorbers through resonance transition (e.g.O vii/O viii, Ne x and Fe xxv/Fe xxvi) independently characterized by hydrogen-equivalent column density N H and ionization parameter ξ.
In addition to the observed ionized outflows, a number of AGNs are further known to undergo significant spectral variations through UV absorption and soft X-ray obscuration by variable opticallythick materials.In some case, these are suggestive of localized obscuration by Compton-thick media conceived in the innermost circumnuclear region for Seyfert 2 AGNs (e.g.Marchesi et al. 2018Marchesi et al. , 2022)).While not clear how and where the implied obscurers are produced and driven, their origin appears to be closely related to large-scale disk winds perhaps extending all the way out to broad line region (BLR; e.g.Kaastra et al. 2014;Mehdipour et al. 2022, hereafter, M22;Kara et al. 2021;Homayouni et al. 2023).These X-ray obscurers have been extensively studied and analyzed in conjunction with the response of their UV counterpart found in an increasing fraction of Seyfert 1 AGNs to date; e.g.NGC 3783 (e.g.Mehdipour et al. 2017;Kriss et al. 2019), NGC 5548 (e.g.Kaastra et al. 2014;Mehdipour et al. 2015;Dehghanian et al. 2021;M22), Mrk 817 (e.g.Cackett et al. 2023;Partington et al. 2023;Homayouni et al. 2023) and NGC 985 (e.g.Ebrero et al. 2016), among others, based primarily on synergistic monitoring programs such as AGN STORM collaboration (Dalla Bontà et al. 2016).
For example, the 2013 campaign on NGC 5548 (Kaastra et al. 2014) showed that the X-ray obscurer in NGC 5548 consists of two different ionized components.The 1st component with N H ∼ 1.2 × 10 22 cm −2 covers about 90% of the X-ray source, while the 2nd component has higher column density (N H ∼ 10 23 cm −2 ) and lower covering fraction (30%).This 2nd component with lower ionization parameter is thought to be denser clumps embedded in the medium of the 1st component.Further photoionization modeling studies of the obscurer in NGC 5548 (Kriss et al. 2019) found that it is very challenging to explain both the X-ray obscuration and the broad UV absorption lines together consistently.Most likely because of the complex and multi-component nature of the obscuring wind, finding a unique photoionization solution is not feasible, and in particular the ionization parameter of the obscurer remains uncertain.This complexity makes it difficult to derive other physical parameters of the obscurer.Nonetheless, the broad UV absorption lines (such as C iv, N v, Si iv, and Lyα) show the outflow velocity of the obscurer is dominant at around 2,000 km/s , reaching up to 5,000 km/s (Kaastra et al. 2014).The previous X-ray and UV studies have found that the obscurer varies from short timescales (ks/days; Kaastra et al. 2014, Di Gesu et al. 2015, Cappi et al. 2016) to long timescales (month/years; M22).Based on the broad UV absorption lines, which partially cover the broad emission lines, Kaastra et al. (2014) inferred the obscurer is at a distance of 2-7 light days from the nucleus.We refer to these previous studies of NGC 5548 for more details on the X-ray and UV observational characteristics of the obscurer.Any reliable time lag for the obscurer itself is yet to be known to date.
These studies have clearly indicated that X-ray obscuration and its UV signatures are physically correlated through variability in terms of certain quantities such as UV/X-ray covering fractions, the strength of C iv absorption line and X-ray hardening, for example.Peculiar time-lags are seen in the observed C iv light curves based on an intensive reverberation mapping campaign with HST for Mrk 817 indicative of an obscuring cloud dynamically affecting ionization balance in the gas (e.g.Homayouni et al. 2023).
As a canonical Seyfert 1 AGN, NGC 5548 (z = 0.01717) has been extensively studied with most of the space-based observatories such as Chandra, XMM-Newton, NuSTAR, Swift, and HST for exhibiting a plethora of rich UV/X-ray features.Among others, M22 show that this AGN has gone through an evolutionary phase of X-ray obscuration since 2012, accompanied by broad (e.g.C iv) and narrow (e.g.C ii) UV absorption features (∼ 1, 000 − 5, 000 km s −1 ; Kaastra et al. 2014).While the column density of the observed X-ray warm absorbers is found to have a typical value of N H ∼ 10 22 cm −2 during the unobscured state, it is found that the observed spectra during a series of obscured epochs (i.e.2012 onwards) require an additional higher column (i.e.N H ∼ 10 23 cm −2 ).Detailed spectral analyses suggest that the observed absorption in UV and X-ray obscuration can be systematically accounted for by high column with variable covering fractions in UV (C UV ∼ 0 − 0.3) and X-ray (C X ∼ 0.5 − 0.95), implying an intriguing scenario in which the UV absorber nearly vanishes while X-ray obscuration is still significant.On another interesting note, the observed C iv absorption appears to be imprinted almost exclusively in the blue tail of the observed C iv emission line, while very weakly (if not none) in the red tail.
Another similar observational signature showing a coexistence of UV absorbers and X-ray obscurers has been reported in NGC 3783 (z = 0.00973) with the observed absorption, emission and extinction effects being simultaneously analyzed, resulting in nine distinct X-ray warm absorber components (e.g.Mao et al. 2019).Multi-epoch synergistic observations of NGC 3783 have further revealed the presence of X-ray-obscuring column (N H ∼ 10 23 cm −2 ) in XMM-Newton spectra simultaneous with the UV absorbers (e.g.C iv line up to ∼ 6, 000 km s −1 ) in the observed HST/COS data (Kriss et al. 2019).
These unique case studies clearly point to the importance of a fundamental coupling that underlies the separate phenomena seen in UV and X-ray.Motivated by these observational facts, it is suggested that fragmentation of disk-winds into clumpy clouds due to thermal instabilities (e.g.Waters et al. 2022) might be among plausible physical configurations consistent with multi-wavelength data (e.g.M22; Homayouni et al. 2023).Nonetheless, a definitive characterization of the implied absorbing gas and obscuring materials has yet to be fully investigated from a theoretical standpoint in a systematic fashion.

of certain
We adopt the wind model in F10 to simulate various wind morphologies, which are then coupled to post-process radiative transfer calculations with xstar (Kallman & Bautista 2001) to solve for ionization equilibrium of irradiated outflows.As a consequence, an emitted UV/X-ray continua are absorbed and obscured by ionized winds from which we compute synthetic spectra via calculated ionic column densities of C iv ions as the radiation penetrates progressively through winds.Our primary focus in the present work is to explore observable correlations (or their absence) between the degree of X-ray obscuration and corresponding UV absorption signature traced via C iv line.
In §2, we briefly summarize the basics of MHD wind models from F10.In §3, we present our results demonstrating its synergistic concordance to the derived observational facts.We discuss the implications of our model and results in §4, followed by Summary and Discussion in §5.

BASICS OF THE STRATIFIED WIND MODEL
As discussed in detail in F10, a classical type of self-similar MHD winds employed in this work (i.e.Blandford & Payne 1982;Contopoulos & Lovelace 1994) exhibits a characteristic profile (see also Kazanas et al. 2012;Kraemer et al. 2018;Kazanas 2019;Jacquemin-Ide et al. 2020).The most essential part of the model is the radial and angular dependence of the wind density n(r, θ) and where n 12 is the wind density normalization (in units of 10 12 cm −3 ) on the footpoint of the innermost wind streamline anchored at r = R in , being typically identified with the innermost stable circular orbit (ISCO), and v K is the Keplerian velocity at r = R in .The angular functions, f (θ) and g(θ), are numerically calculated by solving the Grad-Shafranov equation in conjunction with the ideal MHD equations2 (see F10).It is reminded that the wind motion along a streamline is smoothly and rapidly converted from toroidal (v ϕ ) to poloidal (v r,θ ) direction across the Alfvén point, which is mainly described by g(θ).On a separate note, the model predicts v LoS ∝ ξ 1/(4−2p) (F10) in agreement with some AGN X-ray warm absorbers when p ∼ 1.1 − 1.2 (e.g.Behar 2009;Detmers et al. 2011;Laha et al. 2021).
To better exploit various wind structures allowed by the model, we first consider six distinct wind morphologies as shown in Table 1 for different sets of plasma parameters being conserved along a streamline; i.e. particle flux to magnetic flux (F o ) and total angular momentum of plasma (H o ).The total plasma energy of the wind particles, J o , is roughly conserved as J o ≃ −1.5 in all cases.
The magnetic field orientation on the foot point of winds (i.e.tangential slope of a field line on the disk surface) is defined by (dR/dZ) o ≡ dR/dZ(r = R o , Z ∼ 0) in each case.The wind's poloidal geometry is described by opening angle of wind at the Alfvén point (θ A ) and at the distance of z/R ISCO = 10 4 (θ 4 ).Detailed definition and description of the model parameters are provided in F10.With n 12 = 2.4 and p = 1.1, the innermost wind density along 30 • LoS, n in (30 • ), is computed for wind 1-3 in Table 1.In order to avoid additional model dependencies, we make two assumptions; (i) p = 1.1 not arbitrarily but for the observational implications from AMD analysis as described above.Winds with p ≪ 1 or p ≫ 1 would also unrealistically overproduce or underproduce UV absorbers, which would be inconsistent with UV/X-ray data.(ii) θ = 30 • as often conceived for Seyfert 1 AGNs in the conventional unification paradigm.We note, for example, that the wind 2 considered here is characterized by τ ∼ 0.026 ≪ 1 in the innermost layer at 30 • , while the base of the wind near the disk surface is characterized by Thomson thick with τ ∼ 13.5 ≫ 1 as expected.
In Figure 1 the poloidal configuration of each wind (wind 1-6) is shown; color-coded (normalized) density distribution n(r, θ), density contours (dashed curves), poloidal magnetic field lines (solid curves) and poloidal velocity field (white arrows).LoS of 30 • is given in dashed line for reference.
As expected, the wind density can vary by many orders of magnitude where it is highest near the disk surface at Z ∼ 0 (i.e. a reservoir of matter) and lowest near the funnel region close to the symmetry axis towards the inner edge of wind.Large-scale wind morphology, characterized by θ 4 and θ A (in Table 1), indicates a paraboloidal structure as a generic geometry.Considering a fiducial viewing angle of 30 • for canonical Seyfert 1 AGNs, note that the LoS does not intersect over a wide range of distance with winds of large opening angle; i.e. wind 4-6, allowing X-rays to transmit through wind media without causing obscuration at all.While wind structure can be diverse depending on those variables listed in Table 1, as we are interested in disk winds as obscurers in this paper, we further focus only on wind 1-3 below.
Note that our wind model (in F10) is steady-state and hence we don't follow its time evolution (Blandford & Payne 1982;Contopoulos & Lovelace 1994).Subsequent calculations are made as "snapshot" at each state, given that the typical timescale of X-ray obscuration spans over many years to a decade, much longer than dynamical timescale for accretion and outflows (unlike local stochastic, turbulent variabilities in the gas).Hence, the wind medium is assumed to reach a quasiasymptotic, semi-equilibrium state for a given wind morphology with an observer situated at infinity.

MODELING OF X-RAY OBSCURATION WITH STRATIFIED DISK WINDS
Given the physical properties obtained from wind 1-3 as discussed in §2, radiative transfer calculations are performed with xstar with an input (unabsorbed) spectral energy distribution (SED).To this end, we follow the findings from Mehdipour et al. (2015) for NGC 5548 during its 2013 epoch by employing their reconstructed broadband SED with L X = 1.17 × 10 44 erg s −1 (see inset in Fig. 2a). Figure 2a shows the predicted X-ray spectra as a function of the outer extent of disk-wind (r = R 17 10 17 cm) for wind 2 assuming n 12 = 2.4, p = 1.1 and θ = 30 • .In our model, the wind is perfectly smooth and continuous corresponding to a full covering fraction (i.e.C UV,X = 1) with solar abundances.For n 12 = 2.4, the obscuration does not become effective until r ∼ 10 17 cm.This is simply because a total cumulative wind column at r ≪ 10 17 cm is not sufficient for causing moderate obscuration of X-rays.Under this condition, the inner edge of the wind material is highly photoionized at log ξ ∼ 6.9 − 7.1 with gas temperature of T ∼ 10 8 K.With p = 1.1, the wind column density falls off very slowly with distance as dN H /d(log ξ) ∝ r 1−p ∝ ξ (p−1)/(2−p) (e.g.Kazanas et al. 2012;Fukumura et al. 2021).We find that the magnitude of obscuration is consistent with the wind morphology (most characterized by opening angle); i.e. the more extreme obscuration in wind 1 and the least case in wind 3.As expected, soft X-ray continuum becomes progressively more obscured with increasing wind outer radius R 17 in all cases.
The fraction of obscured X-ray flux (0.1-20 keV) relative to the intrinsic (unobscured) one is given in Figure 2b as a function of the outer wind radius R out in each case.Wind of smaller opening angle is found to obscure X-ray more effectively at smaller distances; i.e. r ∼ 10 14−14.5 cm in wind 1 in comparison to r > ∼ 10 16−18 cm in wind 2 and wind 3. We find that X-ray is completely blocked in wind 1 for a given condition, while only a small fraction of the intrinsic X-ray is absorbed in wind 3. To further investigate obscured X-ray spectral shape relevant for observations, therefore, we follow up on wind 2 as a fiducial case.
In Figure 3a we calculate hardness ratio defined as HR ≡ (H−S)/(H+S) where H (or S) is 2-10 keV (or 0.2-2 keV) flux.As expected, higher wind density (n 12 = 85) naturally enhances obscuration at smaller distances.From Figures 1-3a we learn that X-ray continuum can be very efficiently obscured through the wind, which is also manifested in a quick rise of HR.This is well understood by the fact that the continuum flux suffers progressively from attenuation by the stratified wind eventually reaching a crucial point where the accumulated column over the LoS becomes substantial.We also calculate the fractional changes in H by ∆f H ≡ |∆H|/H unobs (in blue) and S by ∆f S ≡ |∆S|/S unobs (in red) in Figure 3a for comparison where "unobs" denotes unobscured (intrinsic) flux and ∆H and ∆S are the flux changes relative to their unobscured values.The variation of HR (thus X-ray hardening) is initially caused because both soft (S) and hard (H) X-ray fluxes are obscured as the wind radius R out increases up to ∼ 10 17.5 cm.As R out extends beyond ∼ 10 17.5 cm, hard X-ray suppression roughly saturates while soft X-ray continues to get obscured, which results in the change in HR shown in Figure 3a.
We stress here again that the timescale of this change in obscuration is very long (e.g.years; ∼ 9 years in NGC 5548) and it is reasonable to consider that accretion and wind are in quasi-equilibrium as R 17 varies.These are therefore "snapshot" calculations.The light-crossing time between ∼ 10 16 cm to 10 18 cm is about one year, although the global wind morphological change (i.e.wind size) should take place at much slower rate.
In parallel, we examine the broadband spectral shape by computing the optical-to-X-ray strength 3 , α OX , in wind 2 in Figure 3b for n 12 = 2.4.As radiation passes through the wind with increasing R out , X-ray hardening occurs due to obscuration in soft band (i.e.Γ ∼ 1.5 → 0) accompanied by increasing UV counterpart relative to X-ray (i.e.α OX ∼ −1.2 → −1.6) as a consequence.

PREDICTED X-RAY VS. UV RELATIONS
In relation to X-ray, we consider the transmission of UV continuum due to the same disk winds at larger distances.As a diagnostic proxy in UV band better suited to study BLR, we calculate C iv absorption line (a doublet at 1548 in conjunction with X-ray spectra in wind 2 for n 12 = 2.4, by simply assuming that the outer BLR is located within r ∼ 10 17 cm (i.e.∼ 10 4 Schwarzschild radii) with the dynamical timescale of ∼ 30 days or less (e.g.Kaspi et al. 2005;Homayouni et al. 2023).In Figure 4a we show the signature of broad C iv absorption imprinted on Gaussian C iv emission (assuming a source distance at z = 0.01) as a function of the normalized distance R 17 of the outer edge of the wind where where r = 10 17 R 17 cm.This thus shows UV counterpart to X-ray obscuration given in Figure 2a.(Tananbaum et al. 1979).An interesting feature is that X-ray is continuously obscured at distances through R 17 ∼ 2.1 − 6.9 (Fig. 2a), while absorption feature in C iv region is only marginally present.This may explain why in some obscured AGNs the observed C iv absorption is too weak to be detected; e.g.NGC 3227 (Mao et al. 2022a) and MR 2251-178 (Mao et al. 2022b).Beyond R 17 ∼ 6.9, the soft X-ray continuum being already suppressed remains nearly unchanged with increasing distance.On the other hand, the C iv absorption is very sensitive to distance; e.g. the absorption continues to become deeper and broader at R 17 < ∼ 8.15 with v LoS > ∼ 5, 400 km s −1 and can be almost saturated at R 17 > ∼ 8.7 with v LoS < ∼ 4, 800 km s −1 , whereas the spectral appearance of X-ray obscuration seems to vary little.This is because the gas of higher column at inner part of the wind can "self-shield" the outer part of the gas near BLR to lower ionization state for C iv.Hence, UV counterparts such as C iv line can only emerge at distant locations where initially strong ionizing X-rays are sufficiently weakened by wind media.To the best of our knowledge, we are not aware of any theoretical models trying to account for the observed X-ray obscuration in conjunction with UV absorbers.
The absorption signature becomes less blueshifted (i.e.shifting towards longer wavelength) with increasing R 17 because of the underlying wind kinematics being v LoS ∝ r −1/2 as expressed in equation (1).The feature can get relatively broad suppressing almost the entire blue tail of emission (e.g.R 17 ∼ 8.7 − 8.99).These spectral features, X-ray obscuration and UV absorbers, thus appear to be consistent with multi-wavelength observations of NGC 5548 and NGC 3783 (e.g.Kaastra et al. 2014;Kriss et al. 2019;M22).
Lastly, we exploit our results in UV/X-ray by extracting a correlation between the equivalent width (EW) of broad C iv absorption and X-ray HR as the outer extent of the wind R 17 varies.A correlation is clearly seen in Figure 4b where the following five distinct regimes are identified; (i) unobscured state (e.g.R 17 < ∼ 2.1) with no/little X-ray obscuration nor UV absorption, (ii) mildly obscured state (e.g.R 17 > ∼ 2.1 − 4.0) exhibiting moderate obscuration above ∼ 1 keV with little/no UV absorption, and (iii) obscured state (e.g.R 17 ∼ 4.0 − 5.6) showing significant obscuration above ∼ 1 keV with little/no UV absorption and (iv) heavily obscured state (R 17 > ∼ 5.6 − 8.0) even below ∼ 1 keV together with pronounced UV absorption signatures, and (v) saturated state (R 17 > ∼ 8.0) in both UV and X-ray band.This trend is partially understood also by the variation of HR shown in Figure 3a.As stated in §3, attenuation (in X-ray and UV) can be sensitive to the wind size R 17 because of the stratified structure of continuous density field as characterized as n ∝ r −p (from denser layer to tenuous layer outward in density) where incoming X-rays are progressively absorbed along the LoS.Thus, the larger the wind size is, the more the obscuration is.For comparison, the observed correlation from NGC 5548 using HST/COS and Swift is also shown (in red from M22), which is broadly consistent with the prediction setting aside the exact details (e.g.normalization and slope).

SUMMARY & DISCUSSION
We consider the dual effects of accretion disk winds of ionized plasma in AGNs in the context of the variable X-ray obscuration and UV absorption features often observed in Seyfert 1 galaxies such as NGC 5548 and NGC 3783.As a proof of concept study, we focus on a smooth and continuous winds of stratified structure (as opposed to clumpy winds) by employing an MHD wind launched over a large radial extent of a disk surface.Being coupled to post-process radiative transfer calculations, we compute obscured X-ray spectrum and its UV counterparts.We find that winds of a small opening angle can provide a sufficient obscuring column along a LoS (e.g.θ = 30 • ) depending on the density normalization n 12 at the base of the wind.By simulating a series of broadband spectra as a function of the outer extent of the wind, we demonstrate the following tangible consequences; (1) efficient X-ray obscuration by Compton-thick part of the inner wind at smaller distances near AGN and (2) UV absorption signatures via C iv line attributed to the outer part of the wind around BLR.It is shown that the predicted correlation between X-ray HR and EW of C iv absorption line in our model is (at least qualitatively) viable with a number of multi-wavelength data of NGC 5548 (e.g.M22). Figure 5 illustrates the dual role (UV and X-ray) of the continuous disk wind launched over an extended disk surface with a stratified structure in both density and velocity, which might be well represented by MHD winds as considered in this work where the size of the wind (the outer edge) plays a role in characterizing the degree of the dual effect.
In particular, the calculated C iv absorption, which can be quite broad depending on the outer extent of the wind, is found to be present exclusively on the blueshifted side of C iv emission line profile (see Figure 4a) in good agreement with the observed HST/STIS spectra from NGC 3783 (e.g.Kriss et al. 2019) and NGC 5548 (e.g.Kaastra et al. 2014;M22).The predicted wind velocities of v LoS < ∼ 6, 000 km s −1 , responsible for these absorption features, are in consistence with the inferred values from HST/COS data in NGC 5548 (e.g.Kaastra et al. 2014).We stress that the inner part of the wind helps naturally "self-shield" (by obscuring soft X-ray photons) the outer part of the same wind from being strongly ionized (to allow UV features at BLR distances) in this framework.
It has been suggested that the observed UV/X-ray phenomena can be well explained by obscuring winds in the form of a clumpy medium through fragmentation perhaps due to thermal instabilities (Kaastra et al. 2014;Mehdipour et al. 2015;Kriss et al. 2019;M22;Waters et al. 2022).In this scenario, the behavior of UV/X-ray variabilities (i.e.HR and EW of UV absorption) can be interpreted primarily by variable covering fractions (C UV,X ), while the obscuring column remains relatively high  4b since no theoretical predictions are made in their framework to date.Our present work alternatively suggests that the observed UV/X-ray variabilities can be naturally regulated by the change of disk wind morphology and/or its outer extent (i.e. the outer edge) without needing fragmentation of winds, as demonstrated in Figure 4b.In other words, the outer wind edge (i.e.R out ) in our work serves a purpose similar to the UV/X-ray covering factors (e.g.C UV,X ), as illustrated in Figure 5 where narrow/broad UV/X-ray outflows, multi-phase, multi-layer obscurers and ultra-fast outflows (UFOs) are partially coexisting at different distances (see Zaidouni et al. 2024 (submitted) for a similar observational implication in Mrk 817).Partington et al. (2023) has found that the observed variability in X-ray obscuring column N H (from NICER) is closely tied to the change in EW of Si iv absorption line from HST/COS in Mrk 817, speculating that the X-ray and UV continua are both impacted by the same obscurers.Our calculations also support this coupling (e.g.obscuring N H vs. UV line EW), which can be attributed to the change in R out of the stratified disk wind, but not necessarily in a clumpy form.
It is also known that the observed duration of obscuration is different between NGC 3783 (∼ weeks/months) and NGC 5548 (∼ decade), for example.Unfortunately, NGC 5548 is the only source so far to provide us with a long-term historical evolution of X-ray obscuration in conjunction with corresponding UV absorption.Furthermore, other variables such as BH mass and Eddington ratio can play a role in influencing the ionization state of obscuring material and its location, regardless of whether it is continuous wind or clumpy wind (or MHD driven or radiation driven).A more detailed understanding of obscuration timescale and duration has to wait until fully dynamical wind simulations can self-consistently integrate multi-dimensional radiative transfer calculations to handle realistic photoionization equilibrium in the gas, which is beyond the scope of the current work.
Such a difference can also be explained by the spatial extent of the wind in the present framework (see also M22).
Our predicted UV/X-ray correlation shown in Figure 4b, while qualitatively consistent, does not quantitatively match the inferred slope and the normalization for NGC 5548 (M22).This discrepancy can be reduced (if not fully resolved) by a different choice of n 12 and p in the model.As a preliminary study, our results are presented in this work assuming a set of wind parameters mainly represented by n 12 = 2.4 and p = 1.1 assuming θ = 30 • .We note that the energy range used to compute HR is slightly different between our work and M22, but this does not change the end result.In the context of our obscuring wind model, we see that higher value of n 12 would enhance both obscuration and UV absorption, whereas higher slope of p would lead to a more centrally concentrated wind near AGN.Hence, it is expected that one can reproduce the observed correlation by, for example, decreasing p-value (e.g.p ∼ 0.9 − 1.0) with the same n 12 value in order to leave sufficient gas at larger distances around BLR to slightly enhance C iv EW.We note that other physical variables such as covering fraction, filling factor, intrinsic change in the continuum can also play some roles in determining the observed correlation, which is beyond the scope of this work.A detailed quantitative analysis with multi-wavelength spectroscopy in the future should incorporate these dependences, but it is quite plausible for one to realistically rearrange these variables to better reproduce the derived UV/X-ray correlation.
Historically, Seyfert 2 AGNs are believed to be heavily obscured by Compton-thick gas comprising a "dusty torus" at ∼ 1 − 10 parsec-scale (e.g.Marchesi et al. 2018Marchesi et al. , 2022)).Its geometrical distribution is still debated, but occultation events and infrared observations favor a "clumpy torus" viewpoint (e.g.Jaffe et al. 2004;Elitzur & Shlosman 2006).While not emphasized in this work, it is imaginable that a more permanent obscuration seen in Seyfert 2s, rather than transient one in Seyfert 1s, may well be closely connected to disk-winds of high column considered in this work.A detailed calculation will be needed for further discussion.
In our calculations, we adopt MHD-driven disk wind models (e.g.F10) to demonstrate its viability with data.It is reminded that a pivotal feature of winds needed is a stratified structure (i.e.density and velocity) of a large solid angle over many decades in radius (see Fig. 1) and not necessarily restricted to MHD driving alone.While possible, it is never demonstrated explicitly whether winds driven by other equally promising mechanisms (such as thermal driving and radiation driving) may also possess those physical characteristics required for obscuration.
Physics of obscuring disk wind in Seyfert 1s, as either continuous or clumpy media, is yet to be fully explored while providing us with a handful of intriguing insights.For interpreting multi-epoch spectroscopy, it will be very useful to have a long-term monitoring of expected transient and episodic disk winds similar to those available in NGC 5548 and NGC 3783, for example, in an attempt to collectively understand the underlying variable nature of obscurers.A future effort should be able to unambiguously reveal the elusive identity of obscuring winds.

Figure 1 .
Figure 1.Poloidal morphology in (R, Z)-coordinates of six simulated MHD-driven disk winds in the innermost region (up to ∼ 10 3 R o ) for various sets of conserved quantities, as listed in Table1, showing color-coded normalized density distribution n(r, θ), density contours (dashed curves), poloidal magnetic field lines (solid curves) and poloidal velocity field (white arrows).Distances are given in units of R o ≡ R ISCO for M/M ⊙ = 3 × 10 7 as a fiducial black hole mass in NGC 3783(Vestergaard & Peterson 2006) assuming p = 1.1.LoS of 30 • is given in dashed line.Wind 4-6 would produce no obscuration at all for low inclination (∼ 30 • ), hence excluded from further discussion in §3-5.See the explanation in §2.

Figure 2 .
Figure 2. (a) Calculations of obscured X-ray spectra for (wind 2) at different outer extent of the wind R out for n 12 = 2.4, p = 1.1 and θ = 30 • .Inset in (a) shows the intrinsic (unobscured) SED (in Jy Hz) used for radiative transfer calculations (Mehdipour et al. 2015).(b) Predicted fractional obscuration in 0.1-20 keV flux relative to the unobscured flux as a function of R out for wind 1-3 in comparison.

Figure 3 .
Figure 3. (a) Calculated HR of X-ray spectra (dark thick) with the fractional decrease of hard flux ∆f H (2 − 10 keV in blue) and soft flux ∆f S (0.2 − 2 keV in red) as a function of the outer extent of the wind R out in wind 2. HR with n 12 = 85 is shown in inset.(b) Predicted spectral slope, Γ (in red) and α OX (in dark), for n 12 = 2.4 in wind 2.

3
The spectral index α OX ≡ 0.384 log(f 2keV /f 2500 • A ) measures the X-ray-to-UV relative brightness where f 2keV and f 2500 • A are respectively 2 keV and 2500 • A flux densities

Figure 4 .
Figure 4. (a) Calculations of broad C iv absorption imprinted on Gaussian emission for different outer extent of the wind R 17 in wind 2 for n 12 = 2.4 assuming a source redshift of z = 0.01 (UV counterpart to Figure 2a).Inset shows a progression of C iv absorption line with increasing R 17 (indicated by numbers).(b) Predicted correlation between X-ray HR and EW (in

Table 1 .
Stratified Disk Wind Parameters † See the text in §2 for the physical significance of the parameters.‡ Wind density at its inner edge for 30 • LoS.We assume p = 1.1 for all the wind solutions.
Schematic of obscuring accretion disk wind with dual role (UV+X-ray) in Seyfert 1s in this scenario.General characteristic property is indicated.over multiple epochs; e.g.C UV ∼ 0.01 − 0.3 and C X ∼ 0.5 − 0.95 while N H ∼ 10 23 cm −2 over a decade in NGC 5548 (M22).As the changes in the observed continuum and luminosities are only moderate, it may be thought that the observed variabilities of X-ray continuum and UV absorption are intrinsically caused by a clumpy nature of obscuring disk winds instead of by varying coronae or disk radiation.On the other hand, it is not clear whether those clumpy wind models are in fact capable of reproducing quantitatively the observed correlation in Figure