Cosmic acceleration may be due to modified gravity, with effective field theory or property functions describing the theory. Connection to cosmological observations through practical parametrization of these functions is difficult and also faces the issue that not all assumed time dependence or parts of parameter space give a stable theory. We investigate the relation between parametrization and stability in Horndeski gravity, showing that the results are highly dependent on the function parametrization. This can cause misinterpretations of cosmological observations, hiding and even ruling out key theoretical signatures. We discuss approaches and constraints that can be placed on the property functions and scalar sound speed to preserve some observational properties, but find that parametrizations closest to the observations, e.g. in terms of the gravitational strengths, offer more robust physical interpretations. In addition we present an example of how future observations of the B-mode polarization of the cosmic microwave background from primordial gravitational waves can probe different aspects of gravity.

Cosmic spatial curvature is a fundamental geometric quantity of the Universe. We investigate a model independent, geometric approach to measure spatial curvature directly from observations, without any derivatives of data. This employs strong lensing time delays and supernova distance measurements to measure the curvature itself, rather than just testing consistency with flatness. We define two curvature estimators, with differing error propagation characteristics, that can crosscheck each other, and also show how they can be used to map the curvature in redshift slices, to test constancy of curvature as required by the Robertson-Walker metric. Simulating realizations of redshift distributions and distance measurements of lenses and sources, we estimate uncertainties on the curvature enabled by next generation measurements. The results indicate that the model independent methods, using only geometry without assuming forms for the energy density constituents, can determine the curvature at the $\sim6\times10^{-3}$ level.

Adding to our previous method for dealing with gravitational modifications at redshift $z\gtrsim3$ through a single parameter, we investigate treatment of lower redshift modifications to linear growth observables. We establish subpercent accurate fits to the redshift space distortion observable $f\sigma_8(a)$ using two parameters binned in redshift, testing the results for modifications with time dependence that rises, falls, is nonmonotonic, is multipeaked, and corresponds to $f(R)$ and braneworld gravity. The residuals are then propagated to cosmological parameter biases for DESI observations, and found to cause a shift in the dark energy joint confidence contour by less than the equivalent of $\sim0.1\sigma$. The proposed 2--3 parameter modified gravity description also can reveal physical characteristics of the underlying theory.

Cosmic growth of large scale structure probes the entire history of cosmic expansion and gravitational coupling. To get a clear picture of the effects of modification of gravity we consider a deviation in the coupling strength (effective Newton's constant) at different redshifts, with different durations and amplitudes. We derive, analytically and numerically, the impact on the growth rate and growth amplitude. Galaxy redshift surveys can measure a product of these through redshift space distortions and we connect the modified gravity to the observable in a way that may provide a useful parametrization of the ability of future surveys to test gravity. In particular, modifications during the matter dominated era can be treated by a single parameter, the "area" of the modification, to an accuracy of $\sim0.3\%$ in the observables. We project constraints on both early and late time gravity for the Dark Energy Spectroscopic Instrument and discuss what is needed for tightening tests of gravity to better than 5% uncertainty.