While recent supernova cosmology research has benefited from improved measurements, current analysis approaches are not statistically optimal and will prove insufficient for future surveys. This paper discusses the limitations of current supernova co
smological analyses in treating outliers, selection effects, shape- and color-standardization relations, unexplained dispersion, and heterogeneous observations. We present a new Bayesian framework, called UNITY (Unified Nonlinear Inference for Type-Ia cosmologY), that incorporates significant improvements in our ability to confront these effects. We apply the framework to real supernova observations and demonstrate smaller statistical and systematic uncertainties. We verify earlier results that SNe Ia require nonlinear shape and color standardizations, but we now include these nonlinear relations in a statistically well-justified way. This analysis was primarily performed blinded, in that the basic framework was first validated on simulated data before transitioning to real data. We also discuss possible extensions of the method.
NICMOS 2 observations are crucial for constraining distances to most of the existing sample of z > 1 SNe Ia. Unlike the conventional calibration programs, these observations involve long exposure times and low count rates. Reciprocity failure is know
n to exist in HgCdTe devices and a correction for this effect has already been implemented for high and medium count-rates. However observations at faint count-rates rely on extrapolations. Here instead, we provide a new zeropoint calibration directly applicable to faint sources. This is obtained via inter-calibration of NIC2 F110W/F160W with WFC3 in the low count-rate regime using z ~ 1 elliptical galaxies as tertiary calibrators. These objects have relatively simple near-IR SEDs, uniform colors, and their extended nature gives superior signal-to-noise at the same count rate than would stars. The use of extended objects also allows greater tolerances on PSF profiles. We find ST magnitude zeropoints (after the installation of the NICMOS cooling system, NCS) of 25.296 +- 0.022 for F110W and 25.803 +- 0.023 for F160W, both in agreement with the calibration extrapolated from count-rates 1,000 times larger (25.262 and 25.799). Before the installation of the NCS, we find 24.843 +- 0.025 for F110W and 25.498 +- 0.021 for F160W, also in agreement with the high-count-rate calibration (24.815 and 25.470). We also check the standard bandpasses of WFC3 and NICMOS 2 using a range of stars and galaxies at different colors and find mild tension for WFC3, limiting the accuracy of the zeropoints. To avoid human bias, our cross-calibration was blinded in that the fitted zeropoint differences were hidden until the analysis was finalized.
