The light-cone (LC) effect causes the mean as well as the statistical properties of the redshifted 21-cm signal $T_{\rm b}(\hat{\bf n},\nu)$ to change with frequency $\nu$ (or cosmic time). Consequently, the statistical homogeneity (ergodicity) of the signal along the line of sight (LoS) direction is broken. This is a severe problem particularly during the Epoch of Reionization (EoR) when the mean neutral hydrogen fraction ($\bar{x}_{\rm HI}$) changes rapidly as the universe evolves. This will also pose complications for large bandwidth observations. These effects imply that the 3D power spectrum $P(k)$ fails to quantify the entire second-order statistics of the signal as it assumes the signal to be ergodic and periodic along the LoS. As a proper alternative to $P(k)$, we use the multi-frequency angular power spectrum (MAPS) ${\mathcal C}_{\ell}(\nu_1,\nu_2)$ which does not assume the signal to be ergodic and periodic along the LoS. Here, we study the prospects for measuring the EoR 21-cm MAPS using future observations with the upcoming SKA-Low. Ignoring any contribution from the foregrounds, we find that the EoR 21-cm MAPS can be measured at a confidence level $\ge 5\sigma$ at angular scales $\ell \sim 1300$ for total observation time $t_{\rm obs} \ge 128\,{\rm hrs}$ across $\sim 44\,{\rm MHz}$ observational bandwidth. We also quantitatively address the effects of foregrounds on MAPS detectability forecast by avoiding signal contained within the foreground wedge in $(k_\perp, k_\parallel)$ plane. These results are very relevant for the upcoming large bandwidth EoR experiments as previous predictions were all restricted to individually analyzing the signal over small frequency (or equivalently redshift) intervals.

We present the prospects of extracting information about the Epoch of Reionization by identifying the remaining neutral regions, referred to as islands, in tomographic observations of the redshifted 21-cm signal. Using simulated data sets we show that at late times the 21-cm power spectrum is fairly insensitive to the details of the reionization process but that the properties of the neutral islands can distinguish between different reionization scenarios. We compare the properties of these islands with those of ionized bubbles. At equivalent volume filling fractions, neutral islands tend to be fewer in number but larger compared to the ionized bubbles. In addition, the evolution of the size distribution of neutral islands is found to be slower than that of the ionized bubbles and also their percolation behaviour differs substantially. Even though the neutral islands are relatively rare, they will be easier to identify in observations with the low-frequency component of the Square Kilometre Array (SKA-Low) due to their larger size and the lower noise levels at lower redshifts. The size distribution of neutral islands at the late stages of reionization is found to depend on the source properties, such as the ionizing efficiency of the sources and their minimum mass. We find the longest line of sight through a neutral region to be more than 100 comoving Mpc until very late stages (90-95 per cent reionized), which may have relevance for the long absorption trough at $z=5.6-5.8$ in the spectrum of quasar ULAS J0148+0600.

We use thirty-eight high-resolution simulations of galaxy formation between redshift 10 and 5 to study the impact of a 3 keV warm dark matter (WDM) candidate on the high-redshift Universe. We focus our attention on the stellar mass function and the global star formation rate and consider the consequences for reionization, namely the neutral hydrogen fraction evolution and the electron scattering optical depth. We find that three different effects contribute to differentiate warm and cold dark matter (CDM) predictions: WDM suppresses the number of haloes with mass less than few $10^9$ M$_{\odot}$; at a fixed halo mass, WDM produces fewer stars than CDM; and finally at halo masses below $10^9$ M$_{\odot}$, WDM has a larger fraction of dark haloes than CDM post-reionization. These three effects combine to produce a lower stellar mass function in WDM for galaxies with stellar masses at and below $\sim 10^7$ M$_{\odot}$. For $z > 7$, the global star formation density is lower by a factor of two in the WDM scenario, and for a fixed escape fraction, the fraction of neutral hydrogen is higher by 0.3 at $z \sim 6$. This latter quantity can be partially reconciled with CDM and observations only by increasing the escape fraction from 23 per cent to 34 per cent. Overall, our study shows that galaxy formation simulations at high redshift are a key tool to differentiate between dark matter candidates given a model for baryonic physics.

The upcoming radio interferometer Square Kilometre Array (SKA) is expected to directly detect the redshifted 21-cm signal from the neutral hydrogen present during the Cosmic Dawn. Temperature fluctuations from X-ray heating of the neutral intergalactic medium can dominate the fluctuations in the 21-cm signal from this time. This heating depends on the abundance, clustering, and properties of the X-ray sources present, which remain highly uncertain. We present a suite of three new large-volume, 349\,Mpc a side, fully numerical radiative transfer simulations including QSO-like sources, extending the work previously presented in Ross et al. (2017). The results show that our QSOs have a modest contribution to the heating budget, yet significantly impact the 21-cm signal. Initially, the power spectrum is boosted on large scales by heating from the biased QSO-like sources, before decreasing on all scales. Fluctuations from images of the 21-cm signal with resolutions corresponding to SKA1-Low at the appropriate redshifts are well above the expected noise for deep integrations, indicating that imaging could be feasible for all the X-ray source models considered. The most notable contribution of the QSOs is a dramatic increase in non-Gaussianity of the signal, as measured by the skewness and kurtosis of the 21-cm probability distribution functions. However, in the case of late Lyman-$\alpha$ saturation, this non-Gaussianity could be dramatically decreased particularly when heating occurs earlier. We conclude that increased non-Gaussianity is a promising signature of rare X-ray sources at this time, provided that Lyman-$\alpha$ saturation occurs before heating dominates the 21-cm signal.

We use cosmological hydrodynamical galaxy formation simulations from the NIHAO project to investigate the impact of the threshold for star formation on the response of the dark matter (DM) halo to baryonic processes. The fiducial NIHAO threshold, $n=10\, {\rm cm}^{-3}$, results in strong expansion of the DM halo in galaxies with stellar masses in the range $10^{7.5} < M_{star} < 10^{9.5} M_{\odot}$. We find that lower thresholds such as $n=0.1$ (as employed by the EAGLE/APOSTLE and Illustris/AURIGA projects) do not result in significant halo expansion at any mass scale. Halo expansion driven by supernova feedback requires significant fluctuations in the local gas fraction on sub-dynamical times (i.e., < 50 Myr at galaxy half-light radii), which are themselves caused by variability in the star formation rate. At one per cent of the virial radius, simulations with $n=10$ have gas fractions of $\simeq 0.2$ and variations of $\simeq 0.1$, while $n=0.1$ simulations have order of magnitude lower gas fractions and hence do not expand the halo. The observed DM circular velocities of nearby dwarf galaxies are inconsistent with CDM simulations with $n=0.1$ and $n=1$, but in reasonable agreement with $n=10$. Star formation rates are more variable for higher $n$, lower galaxy masses, and when star formation is measured on shorter time scales. For example, simulations with $n=10$ have up to 0.4 dex higher scatter in specific star formation rates than simulations with $n=0.1$. Thus observationally constraining the sub-grid model for star formation, and hence the nature of DM, should be possible in the near future.

In this third paper of the series, we investigate the effects of warm dark matter with a particle mass of $m_\mathrm{WDM}=3\,\mathrm{keV}$ on the smallest galaxies in our Universe. We present a sample of 21 hydrodynamical cosmological simulations of dwarf galaxies and 20 simulations of satellite-host galaxy interaction that we performed both in a Cold Dark Matter (CDM) and Warm Dark Matter (WDM) scenario. In the WDM simulations, we observe a higher critical mass for the onset of star formation. Structure growth is delayed in WDM, as a result WDM haloes have a stellar population on average two Gyrs younger than their CDM counterparts. Nevertheless, despite this delayed star formation, CDM and WDM galaxies are both able to reproduce the observed scaling relations for velocity dispersion, stellar mass, size, and metallicity at $z=0$. WDM satellite haloes in a Milky Way mass host are more susceptible to tidal stripping due to their lower concentrations, but their galaxies can even survive longer than the CDM counterparts if they live in a dark matter halo with a steeper central slope. In agreement with our previous CDM satellite study we observe a steepening of the WDM satellites' central dark matter density slope due to stripping. The difference in the average stellar age for satellite galaxies, between CDM and WDM, could be used in the future for disentangling these two models.

We present analysis of the normalised 21-cm bispectrum from fully-numerical simulations of intergalactic-medium heating by stellar sources and high-mass X-ray binaries (HMXB) during the cosmic dawn. Lyman-$\alpha$ coupling is assumed to be saturated, we therefore probe the nature of non-Gaussianities produced by X-ray heating processes. We find the evolution of the normalised bispectrum to be very different from that of the power spectrum. It exhibits a turnover whose peak moves from large to small scales with decreasing redshift, and corresponds to the typical separation of emission regions. This characteristic scale reduces as more and more regions move into emission with time. Ultimately, small-scale fluctuations within heated regions come to dominate the normalised bispectrum, which at the end of the simulation is almost entirely driven by fluctuations in the density field. To establish how generic the qualitative evolution of the normalised bispectrum we see in the stellar + HMXB simulation is, we examine several other simulations - two fully-numerical simulations that include QSO sources, and two with contrasting source properties produced with the semi-numerical simulation 21cmFAST. We find the qualitative evolution of the normalised bispectrum during X-ray heating to be generic, unless the sources of X-rays are, as with QSOs, less numerous and so exhibit more distinct isolated heated profiles. Assuming mitigation of foreground and instrumental effects are ultimately effective, we find that we should be sensitive to the normalised bispectrum during the epoch of heating, so long as the spin temperature has not saturated by $z \approx 19$.

The NIHAO cosmological simulations form a collection of a hundred high-resolution galaxies. We used these simulations to test the impact of stellar feedback on the morphology of the HI distribution in galaxies. We ran a subsample of twenty of the galaxies with different parameterizations of stellar feedback, looking for differences in the HI spatial distribution and morphology. We found that different feedback models do leave a signature in HI, and can potentially be compared with current and future observations. These findings can help inform future modeling efforts in the parameterization of stellar feedback in cosmological simulations of galaxy formation and evolution.

The upcoming radio interferometer Square Kilometre Array is expected to directly detect the redshifted 21-cm signal from the Cosmic Dawn for the first time. In this era temperature fluctuations from X-ray heating of the neutral intergalactic medium can impact this signal dramatically. Previously, in (Ross et al, 2017), we presented the first large-volume, 244 $h^{-1}$Mpc = 349 Mpc a side, fully numerical radiative transfer simulations of X-ray heating. This work is a follow-up where we now also consider QSO-like sources in addition to high mass X-ray binaries. Images of the two cases are clearly distinguishable at SKA1-LOW resolution and have RMS fluctuations above the expected noise. The inclusion of QSOs leads to a dramatic increase in non-Gaussianity of the signal, as measured by the skewness and kurtosis of the 21-cm signal. We conclude that this increased non-Gaussianity is a promising signature of early QSOs.

The upcoming SKA1-Low radio interferometer will be sensitive enough to produce tomographic imaging data of the redshifted 21-cm signal from the Epoch of Reionization. Due to the non-Gaussian distribution of the signal, a power spectrum analysis alone will not provide a complete description of its properties. Here, we consider an additional metric which could be derived from tomographic imaging data, namely the bubble size distribution of ionized regions. We study three methods that have previously been used to characterize bubble size distributions in simulation data for the hydrogen ionization fraction - the spherical-average, mean-free-path and friends-of-friends methods - and apply them to simulated 21-cm data cubes. Our simulated data cubes have the (sensitivity-dictated) resolution expected for the SKA1-Low reionization experiment and we study the impact of both the light-cone and redshift space distortion effects. To identify ionized regions in the 21-cm data we introduce a new, self-adjusting thresholding approach based on the K-Means algorithm. We find that the fraction of ionized cells identified in this way consistently falls below the mean volume-averaged ionized fraction. From a comparison of the three bubble size methods, we conclude that all three methods are useful, but that the mean-free-path method performs best in terms of tracking the progress of reionization and separating different reionization scenarios. The light-cone effect is found to affect data spanning more than about 10~MHz in frequency ($\Delta z\sim0.5$). We find that redshift space distortions only marginally affect the bubble size distributions.

We introduce a novel technique, called "granulometry", to characterize and recover the mean size and the size distribution of HII regions from 21-cm tomography. The technique is easy to implement, but places the previously not very well defined concept of morphology on a firm mathematical foundation. The size distribution of the cold spots in 21-cm tomography can be used as a direct tracer of the underlying probability distribution of HII region sizes. We explore the capability of the method using large-scale reionization simulations and mock observational data cubes while considering capabilities of SKA1-low and a future extension to SKA2. We show that the technique allows the recovery of the HII region size distribution with a moderate signal-to-noise ratio from wide-field imaging ($\rm SNR\lesssim3$), for which the statistical uncertainty is sample variance dominated. We address the observational requirements on the angular resolution, the field-of-view, and the thermal noise limit for a successful measurement. To achieve a full scientific return from 21-cm tomography and to exploit a synergy with 21-cm power spectra, we suggest an observing strategy using wide-field imaging (several tens of square degrees) by an interferometric mosaicking/multi-beam observation with additional intermediate baselines (~2-4 km).

Radiative feedback from stars and galaxies has been proposed as a potential solution to many of the tensions with simplistic galaxy formation models based on $\Lambda$CDM, such as the faint end of the UV luminosity function. The total energy budget of radiation could exceed that of galactic winds and supernovae combined, which has driven the development of sophisticated algorithms that evolve both the radiation field and the hydrodynamical response of gas simultaneously, in a cosmological context. We probe self-feedback on galactic scales using the adaptive mesh refinement, radiative transfer, hydrodynamics, and $N$-body code. Unlike previous studies which assume a homogeneous UV background, we self-consistently evolve both the radiation field and gas to constrain the halo baryon fraction during cosmic reionization. We demonstrate that the characteristic halo mass with mean baryon fraction half the cosmic mean, $M_{\mathrm{c}}(z)$, shows very little variation as a function of mass-weighted ionization fraction. Furthermore, we find that the inclusion of metal cooling and the ability to resolve scales small enough for self-shielding to become efficient leads to a significant drop in $M_{\mathrm{c}}$ when compared to recent studies. Finally, we develop an Artificial Neural Network that is capable of predicting the baryon fraction of haloes based on recent tidal interactions, gas temperature, and mass-weighted ionization fraction. Such a model can be applied to any reionization history, and trivially incorporated into semi-analytical models of galaxy formation.

Large variations in the effective optical depth of the He II Ly$\alpha$ forest have been observed at $z\gtrsim2.7$, but the physical nature of these variations is uncertain: either the Universe is still undergoing the process of He II reionization, or the Universe is highly ionized but the He II-ionizing background fluctuates significantly on large scales. In an effort to build upon our understanding of the latter scenario, we present a novel model for the evolution of ionizing background fluctuations. Previous models have assumed the mean free path of ionizing photons to be spatially uniform, ignoring the dependence of that scale on the local ionization state of the intergalactic medium (IGM). This assumption is reasonable when the mean free path is large compared to the average distance between the primary sources of He II-ionizing photons, $\gtrsim L_\star$ quasars. However, when this is no longer the case, the background fluctuations become more severe, and an accurate description of the average propagation of ionizing photons through the IGM requires additionally accounting for the fluctuations in opacity. We demonstrate the importance of this effect by constructing 3D semi-analytic models of the helium ionizing background from $z=2.5$-$3.5$ that explicitly include a spatially varying mean free path of ionizing photons. The resulting distribution of effective optical depths at large scales in the He II Ly$\alpha$ forest is very similar to the latest observations with HST/COS at $2.5 \lesssim z \lesssim 3.5$.

Upcoming observations of the 21-cm signal from the Epoch of Reionization will soon provide the first direct detection of this era. This signal is influenced by many astrophysical effects, including long range X-ray heating of the intergalactic gas. During the preceding Cosmic Dawn era the impact of this heating on the 21-cm signal is particularly prominent, especially before spin temperature saturation. We present the largest-volume (349\,Mpc comoving=244~$h^{-1}$Mpc) full numerical radiative transfer simulations to date of this epoch that include the effects of helium and multi-frequency heating, both with and without X-ray sources. We show that X-ray sources contribute significantly to early heating of the neutral intergalactic medium and, hence, to the corresponding 21-cm signal. The inclusion of hard, energetic radiation yields an earlier, extended transition from absorption to emission compared to the stellar-only case. The presence of X-ray sources decreases the absolute value of the mean 21-cm differential brightness temperature. These hard sources also significantly increase the 21-cm fluctuations compared the common assumption of temperature saturation. The 21-cm differential brightness temperature power spectrum is initially boosted on large scales, before decreasing on all scales. Compared to the case of the cold, unheated intergalactic medium, the signal has lower rms fluctuations and increased non-Gaussianity, as measured by the skewness and kurtosis of the 21-cm probability distribution functions. Images of the 21-cm signal with resolution around 11~arcmin still show fluctuations well above the expected noise for deep integrations with the SKA1-Low, indicating that direct imaging of the X-ray heating epoch could be feasible.

Observations of the Milky Way (MW), M31, and their vicinity, known as the Local Group (LG), can provide clues about the sources of reionization. We present a suite of radiative transfer simulations based on initial conditions provided by the Constrained Local UniversE Simulations (CLUES) project that are designed to recreate the Local Universe, including a realistic MW-M31 pair and a nearby Virgo. Our box size (91 Mpc) is large enough to incorporate the relevant sources of ionizing photons for the LG. We employ a range of source models, mimicking the potential effects of radiative feedback for dark matter haloes between $10^{8}-10^{9}$ M$_{\odot}$. Although the LG mostly reionizes in an inside-out fashion, the final 40 per cent of its ionization shows some outside influence. For the LG satellites, we find no evidence that their redshift of reionization is related to the present-day mass of the satellite or the distance from the central galaxy. We find that less than 20 per cent of present-day satellites for MW and M31 have undergone any star formation prior to the end of global reionization. Approximately five per cent of these satellites could be classified as fossils, meaning the majority of star formation occurred at these early times. The more massive satellites have more cumulative star formation prior to the end of global reionization, but the scatter is significant, especially at the low-mass end. Present-day mass and distance from the central galaxy are poor predictors for the presence of ancient stellar populations in satellite galaxies.

Observations of the epoch of reionization give us clues about the nature and evolution of the sources of ionizing photons, or early stars and galaxies. We present a new suite of structure formation and radiative transfer simulations from the PRACE4LOFAR project designed to investigate whether the mechanism of radiative feedback, or the suppression of star formation in ionized regions from UV radiation, can be inferred from these observations. Our source halo mass extends down to $10^8 M_\odot$, with sources in the mass range $10^8$ to $10^9 M_\odot$ expected to be particularly susceptible to feedback from ionizing radiation, and we vary the aggressiveness and nature of this suppression. Not only do we have four distinct source models, we also include two box sizes (67 Mpc and 349 Mpc), each with two grid resolutions. This suite of simulations allows us to investigate the robustness of our results. All of our simulations are broadly consistent with the observed electron-scattering optical depth of the cosmic microwave background and the neutral fraction and photoionization rate of hydrogen at $z\sim6$. In particular, we investigate the redshifted 21-cm emission in anticipation of upcoming radio interferometer observations. We find that the overall shape of the 21-cm signal and various statistics are robust to the exact nature of source suppression, the box size, and the resolution. There are some promising model discriminators in the non-Gaussianity and small-scale power spectrum of the 21-cm signal.

The observed 21-cm signal from the epoch of reionization will be distorted along the line-of-sight by the peculiar velocities of matter particles. These redshift-space distortions will affect the contrast in the signal and will also make it anisotropic. This anisotropy contains information about the cross-correlation between the matter density field and the neutral hydrogen field, and could thus potentially be used to extract information about the sources of reionization. In this paper, we study a collection of simulated reionization scenarios assuming different models for the sources of reionization. We show that the 21-cm anisotropy is best measured by the quadrupole moment of the power spectrum. We find that, unless the properties of the reionization sources are extreme in some way, the quadrupole moment evolves very predictably as a function of global neutral fraction. This predictability implies that redshift-space distortions are not a very sensitive tool for distinguishing between reionization sources. However, the quadrupole moment can be used as a model-independent probe for constraining the reionization history. We show that such measurements can be done to some extent by first-generation instruments such as LOFAR, while the SKA should be able to measure the reionization history using the quadrupole moment of the power spectrum to great accuracy.

A proposed method for dealing with foreground emission in upcoming 21-cm observations from the epoch of reionization is to limit observations to an uncontaminated window in Fourier space. Foreground emission can be avoided in this way, since it is limited to a wedge-shaped region in $k_{\parallel}, k_{\perp}$ space. However, the power spectrum is anisotropic owing to redshift-space distortions from peculiar velocities. Consequently, the 21-cm power spectrum measured in the foreground avoidance window---which samples only a limited range of angles close to the line-of-sight direction---differs from the full spherically-averaged power spectrum which requires an average over \emph{all} angles. In this paper, we calculate the magnitude of this "wedge bias" for the first time. We find that the bias is strongest at high redshifts, where measurements using foreground avoidance will over-estimate the power spectrum by around 100 per cent, possibly obscuring the distinctive rise and fall signature that is anticipated for the spherically-averaged 21-cm power spectrum. In the later stages of reionization, the bias becomes negative, and smaller in magnitude ($\lesssim 20$ per cent). The effect shows only a weak dependence on spatial scale and reionization topology.

We propose a novel technique to separate the late-time, post-reionization component of the kinetic Sunyaev-Zeldovich (kSZ) effect from the contribution to it from a (poorly understood and probably patchy) reionization history. The kSZ effect is one of the most promising probe of the {\em missing baryons} in the Universe. We study the possibility of reconstructing it in three dimensions (3D), using future spectroscopic surveys such as the Euclid survey. By reconstructing a 3D template from galaxy density and peculiar velocity fields from spectroscopic surveys we cross-correlate the estimator against CMB maps. The resulting cross-correlation can help us to map out the kSZ contribution to CMB in 3D as a function of redshift thereby extending previous results which use tomographic reconstruction. This allows the separation of the late time effect from the contribution owing to reionization. By construction, it avoids contamination from foregrounds, primary CMB, tSZ effect as well as from star forming galaxies. Due to a high number density of galaxies the signal-to-noise (S/N) for such cross-correlational studies are higher, compared to the studies involving CMB power spectrum analysis. Using a spherical Bessel-Fourier (sFB) transform we introduce a pair of 3D power-spectra: ${\cal C}^{\parallel}_\ell(k)$ and ${\cal C}^{\perp}_\ell(k)$ that can be used for this purpose. We find that in a future spectroscopic survey with near all-sky coverage and a survey depth of $z\approx 1$, reconstruction of ${\cal C}^{\perp}_\ell(k)$ can be achieved in a few radial wave bands $k\approx(0.01-0.5 h^{-1}\rm Mpc)$ with a S/N of upto ${\cal O}(10)$ for angular harmonics in the range $\ell=(200-2000)$ (abrdiged).

Recent He II Lyman-alpha forest observations from 2.0 < z < 3.2 show large fluctuations in the optical depth at z > 2.7. These results point to a fluctuating He-ionizing background, which may be due to the end of helium reionization of this era. We present a fast, semi-numeric procedure to approximate detailed cosmological simulations. We compute the distribution of dark matter halos, ionization state of helium, and density field at z = 3 in broad agreement with recent simulations. Given our speed and flexibility, we investigate a range of ionizing source and active quasar prescriptions. Spanning a large area of parameter space, we find order-of-magnitude fluctuations in the He II ionization rate in the post-reionization regime. During reionization, the fluctuations are even stronger and develop a bimodal distribution, in contrast to semi-analytic models and the hydrogen equivalent. These distributions indicate a low-level ionizing background even at significant He II fractions.

We examine large-scale fluctuations in the HeII Lyman-alpha forest transmission during and after HeII reionization. We use a simple Monte Carlo model to distribute quasars throughout a large volume and compute the resulting radiation field along one-dimensional skewers. In agreement with previous studies, we find that the rarity of these sources induces order unity fluctuations in the mean optical depth after reionization, even when averaged over large segments (~10-100 Mpc across). We compare our models to existing data along five HeII Lyman-alpha forest lines of sight spanning z~2-3.2. The large cosmic variance contained in our model plausibly explains many of the observed fluctuations at z<2.5. But our models cannot accommodate the large fluctuations toward high optical depths on ~30 Mpc scales observed at z~2.7-2.9, and the measured optical depths (tau>4) at z>2.9 are difficult to explain with a smoothly-evolving mean radiation field. In order to better understand this data, we construct a toy model of HeII reionization, in which we assume that regions with the smallest radiation fields in a post-reionization Universe (or farthest from strong ionizing sources) are completely dark during reionization. The observed fluctuations fit much more comfortably into this model, and we therefore argue that, according to present data, HeII reionization does not complete until z<2.9.

Recent observations suggest that helium became fully ionized around redshift z ~ 3. The HeII optical depth derived from the Lyman-alpha forest decreases substantially from this period to z ~ 2; moreover, it fluctuates strongly near z ~ 3 and then evolves smoothly at lower redshifts. From these opacities, we compute, using a semi-analytic model, the evolution of the mean photoionization rate and the attenuation length for helium over the redshift range 2.0 < z < 3.2. This model includes an inhomogeneous metagalactic radiation background, which is expected during and after helium reionization. We find that assuming a uniform background underestimates the required photoionization rate by up to a factor ~2. When averaged over the (few) available lines of sight, the effective optical depth exhibits a discontinuity near z = 2.8, but the measurement uncertainties are sizable. This feature translates into a jump in the photoionization rate and, provided the quasar emissivity evolves smoothly, in the effective attenuation length, perhaps signaling the helium reionization era. We then compute the evolution of the effective optical depth for a variety of simple helium reionization models, in which the measured quasar luminosity function and the attenuation length, as well as the evolving HeIII fraction, are inputs. A model with reionization ending around redshift z ~ 2.7 is most consistent with the data, although the constraints are not strong thanks to the sparseness of the data.