arXiv subjects

Tiziana Di Matteo

Publications and source records attributed to Tiziana Di Matteo.

At least 127 records · Page 7Linked to original sources

arXiv2015

Luminosity function of [OII] emission-line galaxies in the MassiveBlack-II simulation

We examine the luminosity function (LF) of [OII] emission-line galaxies in the high-resolution cosmological simulation MassiveBlack-II (MBII). From the spectral energy distribution of each galaxy, we select a sub-sample of star-forming galaxies at $0.06 \le z \le 3.0$ using the [OII] emission line luminosity L([OII]). We confirm that the specific star formation rate matches that in the GAMA survey. We show that the [OII] LF at z=1.0 from the MBII shows a good agreement with the LFs from several surveys below L([OII])=$10^{43.0}$ erg/s while the low redshifts ($z \le 0.3$) show an excess in the prediction of bright [OII] galaxies, but still displaying a good match with observations below L([OII])=$10^{41.6}$ erg/s. Based on the validity in reproducing the properties of [OII] galaxies at low redshift ($z \le 1$), we forecast the evolution of the [OII] LF at high redshift ($z \le 3$), which can be tested by upcoming surveys such as the HETDEX and DESI. The slopes of the LFs at bright and faint ends range from -3 to -2 showing minima at z=2. The slope of the bright end evolves approximately as 1/(z+1) at z=2 while the faint end evolves as ~3/(z+1) at $0.6 \le z \le 2$. In addition, a similar analysis is applied for the evolution of [OIII] LFs, which is to be explored in the forthcoming survey WFIRST-AFTA. Finally, we show that the auto-correlation function of [OII] and [OIII] emitting galaxies shows a rapid evolution from z=2 to 1.

astro-ph.GA↗

arXiv2015

Multi-scaling of wholesale electricity prices

We empirically analyze the most volatile component of the electricity price time series from two North-American wholesale electricity markets. We show that these time series exhibit fluctuations which are not described by a Brownian Motion, as they show multi-scaling, high Hurst exponents and sharp price movements. We use the generalized Hurst exponent (GHE, $H(q)$) to show that although these time-series have strong cyclical components, the fluctuations exhibit persistent behaviour, i.e., $H(q)>0.5$. We investigate the effectiveness of the GHE as a predictive tool in a simple linear forecasting model, and study the forecast error as a function of $H(q)$, with $q=1$ and $q=2$. Our results suggest that the GHE can be used as prediction tool for these time series when the Hurst exponent is dynamically evaluated on rolling time windows of size $\approx 50 - 100$ hours. These results are also compared to the case in which the cyclical components have been subtracted from the time series, showing the importance of cyclicality in the prediction power of the Hurst exponent.

q-fin.ST↗

arXiv2015

The formation of Milky Way-mass disk galaxies in the first 500 million years of a cold dark matter universe

Whether among the myriad tiny proto-galaxies there exists a population with similarities to present day galaxies is an open question. We show, using BlueTides, the first hydrodynamic simulation large enough to resolve the relevant scales, that the first massive galaxies to form are %in fact predicted to have extensive rotationally-supported disks. Although their morphology resembles in some ways Milky-way types seen at much lower redshifts, these high-redshift galaxies are smaller, denser, and richer in gas than their low redshift counterparts. From a kinematic analysis of a statistical sample of 216 galaxies at redshift $z=8-10$ we have found that disk galaxies make up 70\% of the population of galaxies with stellar mass $10^{10} M_\odot$ or greater. Cold Dark Matter cosmology therefore makes specific predictions for the population of large galaxies 500 million years after the Big Bang. We argue that wide-field satellite telescopes (e.g. WFIRST) will in the near future discover these first massive disk galaxies. The simplicity of their structure and formation history should make possible new tests of cosmology.

astro-ph.GA↗

arXiv2015

Galaxy shapes and alignments in the MassiveBlack-II hydrodynamic and dark matter-only simulations

We compare the shapes and intrinsic alignments of galaxies in the MassiveBlack-II cosmological hydrodynamic simulation (MBII) to those in a dark matter-only (DMO) simulation performed with the same volume (100$h^{-1}$Mpc)$^{3}$, cosmological parameters, and initial conditions. Understanding the impact of baryonic physics on galaxy shapes and alignments and their relation to the dark matter distribution should prove useful to map the intrinsic alignments of galaxies from hydrodynamic to dark matter-only simulations. We find that dark matter subhalos are typically rounder in MBII, and the shapes of stellar matter in low mass galaxies are more misaligned with the shapes of the dark matter of the corresponding subhalos in the DMO simulation. At $z=0.06$, the fractional difference in the mean misalignment angle between MBII and DMO simulations varies from $\sim 28 \% - 12 \%$ in the mass range $10^{10.8} - 6.0 \times 10^{14} h^{-1}M_{\odot}$. We study the dark matter halo shapes and alignments as a function of radius, and find that while galaxies in MBII are more aligned with the inner parts of their dark matter subhalos, there is no radial trend in their alignments with the corresponding subhalo in the DMO simulation. This result highlights the importance of baryonic physics in determining the alignment of the galaxy with respect to the inner parts of the halo. Finally, we compare the ellipticity-direction (ED) correlation for galaxies to that for dark matter halos, finding that it is suppressed on all scales by stellar-dark matter misalignment. In the projected shape-density correlation ($w_{δ+}$), which includes ellipticity weighting, this effect is partially canceled by the higher mean ellipticities of the stellar component, but differences of order $30-40\%$ remain on scales $> 1$ Mpc over a range of subhalo masses, with scale-dependent effects below $1$ Mpc.

astro-ph.CO↗

arXiv2015

Intrinsic alignments of galaxies in the MassiveBlack-II simulation: analysis of two-point statistics

The intrinsic alignment of galaxies with the large-scale density field is an important astrophysical contaminant in upcoming weak lensing surveys. We present detailed measurements of the galaxy intrinsic alignments and associated ellipticity-direction (ED) and projected shape ($w_{g+}$) correlation functions for galaxies in the cosmological hydrodynamic MassiveBlack-II (MB-II) simulation. We carefully assess the effects on galaxy shapes, misalignment of the stellar component with the dark matter shape and two-point statistics of iterative weighted (by mass and luminosity) definitions of the (reduced and unreduced) inertia tensor. We find that iterative procedures must be adopted for a reliable measurement of the reduced tensor but that luminosity versus mass weighting has only negligible effects. Both ED and $w_{g+}$ correlations increase in amplitude with subhalo mass (in the range of $10^{10} - 6.0\times 10^{14}h^{-1}M_{\odot}$), with a weak redshift dependence (from $z=1$ to $z=0.06$) at fixed mass. At $z \sim 0.3$, we predict a $w_{g+}$ that is in reasonable agreement with SDSS LRG measurements and that decreases in amplitude by a factor of $\sim 5$--18 for galaxies in the LSST survey. We also compared the intrinsic alignments of centrals and satellites, with clear detection of satellite radial alignments within their host halos. Finally, we show that $w_{g+}$ (using subhalos as tracers of density) and $w_{δ+}$ (using dark matter density) predictions from the simulations agree with that of non-linear alignment models (NLA) at scales where the 2-halo term dominates in the correlations (and tabulate associated NLA fitting parameters). The 1-halo term induces a scale dependent bias at small scales which is not modeled in the NLA model.

astro-ph.CO↗

arXiv2015

Relation between Financial Market Structure and the Real Economy: Comparison between Clustering Methods

We quantify the amount of information filtered by different hierarchical clustering methods on correlations between stock returns comparing it with the underlying industrial activity structure. Specifically, we apply, for the first time to financial data, a novel hierarchical clustering approach, the Directed Bubble Hierarchical Tree and we compare it with other methods including the Linkage and k-medoids. In particular, by taking the industrial sector classification of stocks as a benchmark partition, we evaluate how the different methods retrieve this classification. The results show that the Directed Bubble Hierarchical Tree can outperform other methods, being able to retrieve more information with fewer clusters. Moreover, we show that the economic information is hidden at different levels of the hierarchical structures depending on the clustering method. The dynamical analysis on a rolling window also reveals that the different methods show different degrees of sensitivity to events affecting financial markets, like crises. These results can be of interest for all the applications of clustering methods to portfolio optimization and risk hedging.

q-fin.ST↗

arXiv2014

Scaling relations between black holes and their host galaxies: comparing theoretical and observational measurements, and the impact of selection effects

We use the high-resolution simulation MassiveBlackII to examine scaling relations between black hole mass (MBH) and host galaxy properties (sigma, M*, and LV), finding good agreement with observational data, especially at the high-mass end. The simulations have less intrinsic scatter than observations, and the MBH-LV correlation has the largest scatter, suggesting it may the the least fundamental of the three relations. We find Gaussian scatter about all three relations, except among the highest mass galaxies, which host more massive black holes. Below z~2 the slopes for the full population remain roughly z-independent, and only steepen by 50% by z~4. The normalization of the sigma, LV relations evolve by 0.3, 0.43 dex, while the MBH correlation does not evolve to at least z~2. Testing for selection biases, we find samples selected by MBH or M* have steeper slopes than randomly selected samples. If unaccounted for, such a selection function would find faster evolution than inferred from a randomly selected sample, as objects at the highend of the relation tend to evolve more rapidly. We find a potential bias among high-LBH subsamples (tending to reside in higher mass galaxies), but these bright-AGN exhibit no intrinsic bias relative to fainter ones in equivalent-mass hosts, nor is there a significant difference between active- and inactive-samples. Finally we characterize the evolution of individual black holes along the scaling planes. Below the local relation, black holes grow faster than their host (72% of black holes 0.3 dex below the mean relation have a MBH-M* trajectory steeper than the local relation), while those above have shallower trajectories (only 14% are steeper than local). Black holes tend to grow faster than their hosts until surpassing the local relation, at which point their growth is suppressed while their hosts continue to grow, returning them to the mean relation.

astro-ph.GA↗

arXiv2014

Risk diversification: a study of persistence with a filtered correlation-network approach

The evolution with time of the correlation structure of equity returns is studied by means of a filtered network approach investigating persistences and recurrences and their implications for risk diversification strategies. We build dynamically Planar Maximally Filtered Graphs from the correlation structure over a rolling window and we study the persistence of the associated Directed Bubble Hierarchical Tree (DBHT) clustering structure. We observe that the DBHT clustering structure is quite stable during the early 2000' becoming gradually less persistent before the unfolding of the 2007-2008 crisis. The correlation structure eventually recovers persistence in the aftermath of the crisis settling up a new phase, distinct from the pre-cysts structure, where the market structure is less related to industrial sector activity. Notably, we observe that - presently - the correlation structure is loosing again persistence indicating the building-up of another, different, phase. Such dynamical changes in persistence and their occurrence at the unfolding of financial crises rises concerns about the effectiveness of correlation-based portfolio management tools for risk diversification.

q-fin.PM↗

arXiv2014

The role of Compton heating in radiation-regulated accretion on to black holes

We investigate the role of Compton heating in radiation-regulated accretion on to black holes from a neutral dense medium using 1D radiation-hydrodynamic simulations. We focus on the relative effects of Compton-heating and photo-heating as a function of the spectral slope α, assuming a power-law spectrum in the energy range of 13.6 eV--100 keV. While Compton heating is dominant only close to the black hole, it can reduce the accretion rate to 0.1 % ($l \propto \dot{m}^2$ model)--0.01 % ($l \propto \dot{m}$ model) of the Bondi accretion rate when the BH radiation is hard (α ~ 1), where $l$ and $\dot{m}$ are the luminosity and accretion rate normalised by Eddington rates, respectively. The oscillatory behaviour otherwise typically seen in simulations with α > 1, become suppressed when α ~ 1 only for the $l \propto \dot{m}$ model. The relative importance of the Compton heating over photo-heating decreases and the oscillatory behaviour becomes stronger as the spectrum softens. When the spectrum is soft (α > 1.5), photo-heating prevails regardless of models making the effect of Compton heating negligible. On the scale of the ionization front, where the gas supply into the Strömgren sphere from large scale is regulated, photo-heating dominates. Our simulations show consistent results with the advection-dominated accretion flow ($l \propto \dot{m}^2$) where the accretion is inefficient and the spectrum is hard (α ~ 1).

astro-ph.HE↗

arXiv2014

Galaxy Shapes and Intrinsic Alignments in The MassiveBlack-II Simulation

The intrinsic alignment of galaxy shapes with the large-scale density field is a contaminant to weak lensing measurements, as well as being an interesting signature of galaxy formation and evolution (albeit one that is difficult to predict theoretically). Here we investigate the shapes and relative orientations of the stars and dark matter of halos and subhalos (central and satellite) extracted from the MassiveBlack-II simulation, a state-of-the-art high resolution hydrodynamical cosmological simulation which includes stellar and AGN feedback in a volume of $(100{h^{-1}\mathrm{Mpc}})^3$. We consider redshift evolution from $z=1$ to $0.06$ and mass evolution within the range of subhalo masses, $10^{10} -6.0 \times 10^{14.0}{h^{-1}M_{\odot}}$. The shapes of the dark matter distributions are generally more round than the shapes defined by stellar matter. The projected root-mean-square (RMS) ellipticity per component for stellar matter is measured to be $e_\text{rms} = 0.28$ at $z=0.3$ for $M_{subhalo}> 10^{12.0}{h^{-1}M_{\odot}}$, which compares favourably with observational measurements. We find that the shapes of stellar and dark matter are more round for less massive subhalos and at lower redshifts. By directly measuring the relative orientation of the stellar matter and dark matter of subgroups, we find that, on average, the misalignment between the two components is larger for less massive subhalos. The mean misalignment angle varies from $\sim 30^{\circ}-10^{\circ}$ for $M \sim 10^{10} - 10^{14} {h^{-1}M_{\odot}}$ and shows a weak dependence on redshift. We also compare the misalignment angles in central and satellite subhalos at fixed subhalo mass, and find that centrals are more misaligned than satellites. We present fitting formulae for the shapes of dark and stellar matter in subhalos and also the probability distributions of misalignment angles.

astro-ph.CO↗

arXiv2014

The MassiveBlack-II Simulation: The Evolution of Halos and Galaxies to z~0

(Abridged for arXiv)We investigate the properties of halos, galaxies and blackholes to z=0 in the high resolution hydrodynamical simulation MassiveBlack-II (MBII) which evolves a LCDM cosmology in a comoving volume Vbox=100(Mpc/h)^3. MBII is the highest resolution simulation of this size which includes a self-consistent model for star formation, black hole accretion and associated feedback. We provide a simulation browser web application which enables interactive search and tagging of halos, subhalos and their properties and publicly release our galaxy catalogs. Our analysis of the halo mass function (MF) in MBII reveals that baryons have strong effects, with changes in the halo abundance of 20-35% below the knee of the MF (Mhalo < 10^13.2 Msun/h at z=0) when compared to fits based on dark matter only simulations. We provide a fitting function for the halo MF out to redshift z=11 and discuss how the onset of non-universality in the MF limits the accuracy of our fit. We study the halo occupation distribution and clustering of galaxies, in particular the evolution and scale dependence of stochasticity and bias finding reasonable agreement with observational data. The shape of the cosmic spectral energy distribution predicted by MBII is consistent with observations, but lower in amplitude. The Galaxy Stellar Mass Function (GSMF) function is broadly consistent with observations at z>=2. At z<2, the population of passive low mass (for M*<10^9 Msun) galaxies in MBII makes the GSMF too steep compared to observations whereas at the high mass end (M*>10^11 Msun) galaxies hosting bright AGNs make significant contributions to the GSMF. The quasar bolometric luminosity function is also largely consistent with observations. We note however that more efficient AGN feedback (beyond simple thermal coupling used here) is likely necessary for the largest, rarest objects/clusters at low redshifts.

astro-ph.CO↗

arXiv2013

High redshift supermassive blackholes: accretion through cold flows

We use zoom-in techniques to re-simulate three high-redshift (z > 5.5) halos which host 10^9 solar mass blackholes from the ~ Gpc volume, MassiveBlack cosmological hydrodynamic simulation. We examine a number of factors potentially affecting supermassive blackhole growth at high redshift in cosmological simulations. These include numerical resolution, feedback prescriptions and formulation of smoothed particle hydrodynamics. We find that varying the size of the region over which feedback energy is deposited directly, either for fixed number of neighbours or fixed volume makes very little difference to the accretion history of blackholes. Changing mass resolution by factors of up to 64 also does not change the blackhole growth history significantly. We find that switching from the density-entropy formulation to the pressure-entropy formulation of smoothed particle hydrodynamics slightly increases the accretion rate onto blackholes. In general numerical details appear to have small effects on the main fueling mechanism for blackholes at these high redshifts. We examine the fashion by which this occurs, finding that the insensitivity to simulation technique seems to be a hallmark of the cold flow feeding picture of these high-z supermassive blackholes. We show that the gas that participates in critical accretion phases, in these massive objects at z > 6~7 is in all cases colder, denser, and forms more coherent streams than the average gas in the halo. This is also mostly the case when the blackhole accretion is feedback regulated (z < 6), however the distinction is less prominent. For our resimulated halos, cold flows appear to be a viable mechanism for forming the most massive blackholes in the early universe, occurring naturally in LambdaCDM models of structure formation. Not requiring fine tuning of numerical parameters, they seem to be physically inevitable in these objects.

astro-ph.CO↗

arXiv2013

Rayleigh-Taylor instability of ionization front around black holes

We examine the role of ionizing radiation emitted from black holes (BHs) in suppressing the growth of the Rayleigh-Taylor instability (RTI) across the ionization front (I-front) that forms when the gas fueling the BH is neutral. We use radiation-hydrodynamic simulations to show that the RTI is suppressed for non-accelerating fronts on all scales resolved in our simulations. A necessary condition for the stability of the I-front is that the radius of the Strömgren sphere is larger than the Bondi radius. When this condition is violated the I-front collapses producing an accretion luminosity burst. Transient growth of the RTI occurs only during the accretion burst when the effective acceleration in the frame of reference of the I-front increases significantly due to the rapid expansion of the Strömgren sphere.

astro-ph.CO↗

arXiv2013

Theoretical predictions for the effect of nebular emission on the broad band photometry of high-redshift galaxies

By combining optical and near-IR observations from the Hubble Space Telescope with NIR photometry from the Spitzer Space Telescope it is possible to measure the rest-frame UV-optical colours of galaxies at z=4-8. The UV-optical spectral energy distribution of star formation dominated galaxies is the result of several different factors. These include the joint distribution of stellar masses, ages, and metallicities, and the subsequent reprocessing by dust and gas in the ISM. Using a large cosmological hydrodynamical simulation we investigate the predicted spectral energy distributions of galaxies at high-redshift with a particular emphasis on assessing the potential contribution of nebular emission. We find that the average pure stellar UV-optical colour correlates with both luminosity and redshift such that galaxies at lower-redshift and higher-luminosity are typically redder. Assuming the escape fraction of ionising photons is close to zero, the effect of nebular emission is to redden the UV-optical 1500-V_w colour by, on average, 0.4 mag at z=8 declining to 0.25 mag at z=4. Young and low-metallicity stellar populations, which typically have bluer pure stellar UV-optical colours, produce larger ionising luminosities and are thus more strongly affected by the reddening effects of nebular emission. This causes the distribution of 1500-V_w colours to narrow and the trends with luminosity and redshift to weaken. The strong effect of nebular emission leaves observed-frame colours critically sensitive to the source redshift. For example, increasing the redshift by 0.1 can result in observed frame colours changing by up to ~0.6. These predictions reinforce the need to include nebular emission when modelling the spectral energy distributions of galaxies at high-redshift and also highlight the difficultly in interpreting the observed colours of individual galaxies without precise redshifts.

astro-ph.CO↗

arXiv2013

Hydrogen Self Shielding in a Quasar Proximity Zone

We calculate the distribution of HI within 750 proper kpc/h of a quasar, Lbol = 1.62e13 Lsun, powered by an SMBH, Mbh = 4.47e8 Msun, at z = 3. Our numerical model includes a cosmological hydrodynamic simulation that tracks the self consistent growth and thermal feedback of black holes calculated using GADGET-3 as well as a detailed post-processing ray tracing treatment of the non-uniform ionizing radiation field calculated using SPHRAY, which naturally accounts for the self shielding of optically thick systems. We show that the correct treatment of self shielding introduces a flattening feature into the neutral column density distribution around Log NHI = 20 and that regions with the lowest neutral fractions are not those with the highest density gas. For comparison, we solve a Ricatti equation which determines the equilibrium Hydrogen ionization fractions in the presence of a radiation field that falls off as 1/r^2 with regions above a given gas density threshold completely shielded from ionizing radiation. We demonstrate that these semi analytic models cannot reproduce the HI field calculated using SPHRAY. We conclude by comparing our models of this single proximity zone to observations by Hennawi and Prochaska of the absorption spectra of background quasars which are coincident on the sky with foreground quasars in their Quasars Probing Quasars (QPQ) series of papers. Compared to the QPQ sample, we find a factor of 3 fewer optically thick (Log NHI > 17.2) systems around our quasar, however the dark matter halo that hosts our simulated quasar, Mhalo = 5.25e12 Msun, is less massive than the typical QPQ host halo by a factor of four. Allowing for a linear scaling between halo mass, baryonic overdensity and number of absorbers, we estimate the typical host halo mass in the QPQ sample as 1.92e13 Msun.

astro-ph.CO↗

arXiv2013

Interpreting the observed UV continuum slopes of high-redshift galaxies

The observed UV continuum slope of star forming galaxies is strongly affected by the presence of dust. Its observation is then a potentially valuable diagnostic of dust attenuation, particularly at high-redshift where other diagnostics are currently inaccesible. Interpreting the observed UV continuum slope in the context of dust attenuation is often achieved assuming the empirically calibrated Meurer et al. (1999) relation. Implicit in this relation is the assumption of an intrinsic UV continuum slope ($β=-2.23$). However, results from numerical simulations suggest that the intrinsic UV continuum slopes of high-redshift star forming galaxies are bluer than this, and moreover vary with redshift. Using values of the intrinsic slope predicted by numerical models of galaxy formation combined with a Calzetti et al. (2000) reddening law we infer UV attenuations ($A_{1500}$) $0.35-0.5\,{\rm mag}$ ($A_{V}$: $0.14-0.2\,{\rm mag}$ assuming Calzetti et al. 2000) greater than simply assuming the Meurer relation. This has significant implications for the inferred amount of dust attenuation at very-high ($z\approx 7$) redshift given current observational constraints on $β$, combined with the Meurer relation, suggest dust attenuation to be virtually zero in all but the most luminous systems.

astro-ph.CO↗

arXiv2013

Confronting predictions of the galaxy stellar mass function with observations at high-redshift

We investigate the evolution of the galaxy stellar mass function at high-redshift ($z\ge 5$) using a pair of large cosmological hydrodynamical simulations: {\em MassiveBlack} and {\em MassiveBlack-II}. By combining these simulations we can study the properties of galaxies with stellar masses greater than $10^{8}\,{\rm M_{\odot}}\,h^{-1}$ and (co-moving) number densities of $\log_{10}(ϕ\, [{\rm Mpc^{-3}\,dex^{-1}}\,h^{3}])>-8$. Observational determinations of the galaxy stellar mass function at very-high redshift typically assume a relation between the observed UV luminosity and stellar mass-to-light ratio which is applied to high-redshift samples in order to estimate stellar masses. This relation can also be measured from the simulations. We do this, finding two significant differences with the usual observational assumption: it evolves strongly with redshift and has a different shape. Using this relation to make a consistent comparison between galaxy stellar mass functions we find that at $z=6$ and above the simulation predictions are in good agreement with observed data over the whole mass range. Without using the correct UV luminosity and stellar mass-to-light ratio, the discrepancy would be up to two orders of magnitude for large galaxies $>10^{10}\,{\rm M_{\odot}}\,h^{-1}$. At $z=5$, however the stellar mass function for low mass $<10^{9}\,{\rm M_{\odot}}\,h^{-1}$ galaxies is overpredicted by factors of a few, consistent with the behaviour of the UV luminosity function, and perhaps a sign that feedback in the simulation is not efficient enough for these galaxies.

astro-ph.CO↗

arXiv2012

Growth and anisotropy of ionization fronts near high redshift quasars in the MassiveBlack simulation

We use radiative transfer to study the growth of ionized regions around the brightest, z=8 quasars in a large cosmological hydrodynamic simulation that includes black hole growth and feedback (the MassiveBlack simulation). We find that in the presence of the quasar s the comoving HII bubble radii reach 10 Mpc/h after 20 My while with the stellar component alone the HII bubbles are smaller by at least an order of magnitude. Our calculations show that several features are not captured within an analytical growth model of Stromgren spheres. The X-ray photons from hard quasar spectra drive a smooth transition from fully neutral to partially neutral in the ionization front. However the transition from partially neutral to fully ionized is significantly more complex. We measure the distance to the edge of bubbles as a function of angle and use the standard deviation of these distances as a diagnostic of the isotropy of ionized regions. We find that the overlapping of nearby ionized regions from clustered halos not only increases the anisotropy, but also is the main mechanism which allows the outer radius to grow. We therefore predict that quasar ionized bubbles at this early stage in the reionization process should be both significantly larger and more irregularly shaped than bubbles around star forming galaxies. Before the star formation rate increases and the Universe fully reionizes, quasar bubbles will form the most striking and recognizable features in 21cm maps.

astro-ph.CO↗