Earlier studies have conjectured that redshift drift is described by spatially averaged quantities and thus becomes positive if the average expansion of the Universe accelerates. This conclusion is reevaluated here by considering exact light propagation in a simple toy-model with average accelerated expansion. The toy-model and light propagation setup is explicitly designed for concordance between spatial averages and averages along light rays. While it is verified that redshift-distance relations are well described by average quantities in this setup, it is found that the redshift drift is not. Specifically, the redshift drift is negative despite the on-average late-time accelerated expansion of the model. This result implies that measuring redshift drift signals at low redshifts gives the potential for directly falsifying the backreaction conjecture. However, the results are based on a toy-model so it is in principle possible that the result is an artifact and that redshift drift is in reality well described by spatially averaged quantities. The result therefore highlights the importance of developing \emph{exact} solutions to the Einstein equations which exhibit average accelerated expansion without local expansion so that the relation between spatial averages and observations can be firmly established.

Effects of inhomogeneities on observations have been vastly studied using both perturbative methods, N-body simulations and Swiss cheese solutions to the Einstein equations. In nearly all cases, such studied setups assume vanishing spatial background curvature. While a spatially flat Friedmann-Lemaitre-Robertson-Walker model is in accordance with observations, a non-vanishing curvature is not ruled out. It is therefore important to note that, as has been pointed out in the literature, 1 dimensional averages might not converge to volume averages in non-Euclidean space. If this is indeed the case, it will affect the interpretation of observations in spacetimes with non-vanishing average spatial curvature. This possibility is therefore studied here by computing the integrated expansion rate and shear, the accumulated density contrast, and fluctuations in the redshift-distance relation in Swiss cheese models with different background curvatures. It is found that differences in mean and dispersion of these quantities in the different models are small and naturally attributable to differences in background expansion rate and density contrasts. Thus, the study does not yield an indication that the relationship between 1 dimensional spatial averages and volume averages depends significantly on background curvature.

Weak lensing will play an important role in future cosmological surveys, including e.g. Euclid and SKA. Sufficiently accurate theoretical predictions are important for correctly interpreting these surveys and hence for extracting correct cosmological parameter estimations. We quantify for the first time in a relativistic setting how many post-Born and lens-lens coupling corrections are required for sub-percent accuracy of the theoretical weak lensing convergence for $z\le 2$ (the primary weak lensing range for Euclid and SKA). We do this by ray-tracing through a fully relativistic exact solution of the Einstein Field Equations which consists of randomly packed mass-compensated underdensities of realistic amplitudes. We find that including lens-lens coupling terms and post-Born corrections up to second and third order respectively is sufficient for sub-percent accuracy of the convergence along $94\%$ of the studied light rays. We also find that a significant percentage of the studied rays have post-Born corrections of size over $10\%$ of the usual gravitational convergence, $\kappa^{(1)}$, and several rays even have post-Born corrections several times the size of $\kappa^{(1)}$ at $z = 2$.

We study models in which neutrino masses are generated dynamically at cosmologically late times. Our study is purely phenomenological and parameterized in terms of three effective parameters characterizing the redshift of mass generation, the width of the transition region, and the present day neutrino mass. We also study the possibility that neutrinos become strongly self-interacting at the time where the mass is generated. We find that in a number of cases, models with large present day neutrino masses are allowed by current CMB, BAO and supernova data. The increase in the allowed mass range makes it possible that a non-zero neutrino mass could be measured in direct detection experiments such as KATRIN. Intriguingly we also find that there are allowed models in which neutrinos become strongly self-interacting around the epoch of recombination.

Light propagation in two Swiss cheese models based on anisotropic Szekeres structures is studied and compared with light propagation in Swiss cheese models based on the Szekeres models' underlying LTB models. The study shows that the anisotropy of the Szekeres models has only a small effect on quantities such as redshift-distance relations, projected shear and expansion rate along individual light rays. \newline\indent The average angular diameter distance to the last scattering surface is computed for each model. Contrary to earlier studies, the results obtained here are (mostly) in agreement with perturbative results. In particular, a small negative shift, $\delta D_A:=\frac{D_A-D_{A,bg}}{D_{A,bg}}$, in the angular diameter distance is obtained upon line-of-sight averaging in three of the four models. The results are, however, not statistically significant. In the fourth model, there is a small positive shift which has an especially small statistical significance. The line-of-sight averaged inverse magnification at $z = 1100$ is consistent with $1$ to a high level of confidence for all models, indicating that the area of the surface corresponding to $z = 1100$ is close to that of the background.

The effects of local inhomogeneities on low redshift $H_0$ determinations are studied by estimating the redshift-distance relation of mock sources in N-body simulations. The results are compared to those obtained using the standard approach based on Hubble's law. The comparison shows a clear tendency for the standard approach to yield lower values of $H_0$ than the approach based on the scheme using light rays. The difference is, however, small. More precisely, it is found that the overall effect of inhomogeneities on the determination of $H_0$ is a small increase in the local estimates of about $0.3\%$ compared to the results obtained with Hubble's law, when based on a typical distribution of supernovae in the redshift range $0.01 < z < 0.1$. The overall conclusion of the study is a verification of the results that have earlier been obtained by using Hubble's law: The effects of inhomogeneities on local $H_0$ estimates are not significant enough to make it plausible that differences in high- and low-redshift estimates of $H_0$ are due to small inhomogeneities within the setting of standard cosmology.

An expression for the average redshift drift in a statistically homogeneous and isotropic dust universe is given. The expression takes the same form as the expression for the redshift drift in FLRW models. It is used for a proof-of-principle study of the effects of backreaction on redshift drift measurements by combining the expression with two-region models. The study shows that backreaction can lead to positive redshift drift at low redshifts, exemplifying that a positive redshift drift at low redshifts does not require dark energy. Moreover, the study illustrates that models without a dark energy component can have an average redshift drift observationally indistinguishable from that of the standard model according to the currently expected precision of ELT measurements. In an appendix, spherically symmetric solutions to Einstein's equations with inhomogeneous dark energy and matter are used to study deviations from the average redshift drift and effects of local voids.

The simplest standard ray tracing scheme employing the Born and Limber approximations and neglecting lens-lens coupling is used for computing the convergence along individual rays in mock N-body data based on Szekeres swiss cheese and onion models. The results are compared with the exact convergence computed using the exact Szekeres metric combined with the Sachs formalism. A comparison is also made with an extension of the simple ray tracing scheme which includes the Doppler convergence. The exact convergence is reproduced very precisely as the sum of the gravitational and Doppler convergences along rays in Lemaitre-Tolman-Bondi models. This is not the case when the models are based on non-symmetric Szekeres models. For such models, there is a significant deviation between the exact and ray traced paths and hence also the corresponding convergences.

It is proposed to use exact, cosmologically relevant solutions to Einstein's equations to accurately quantify the precision of ray tracing techniques through Newtonian N-body simulations. As an initial example of such a study, the recipe in (Green & Wald, 2012) for going between N-body results and a perturbed FLRW metric in the Newtonian gauge is used to study light propagation through quasi-spherical Szekeres models. The study is conducted by deriving a set of ODEs giving an expression for the angular diameter distance in the Newtonian gauge metric. The accuracy of the results obtained from the ODEs is estimated by using the ODEs to determine the distance-redshift relation in mock N-body data based on quasi-spherical Szekeres models. The results are then compared to the exact relations. From this comparison it is seen that the obtained ODEs can accurately reproduce the distance-redshift relation along both radial and non-radial geodesics in spherically symmetric models. The reproduction of geodesics in non-symmetric Szekeres models is slightly less accurate, but still good. These results indicate that the employment of perturbed FLRW metrics for standard ray tracing techniques yields fairly accurate results, at least regarding distance-redshift relations. It is possible though, that this conclusion will be rendered invalid if other typical ray tracing approximations are included and if light is allowed to travel through several structures instead of just one.