Robustness of width-fitting without pre-peak data

Ascertain whether the DES reference-curve fitting method that estimates each Type Ia supernova light-curve width by chi-square minimization against a stacked, de-redshifted reference light curve yields unbiased and reliable results when the target light curve lacks pre-peak photometric observations, and quantify any bias or increased uncertainty relative to fits that include pre-peak data.

Background

The paper measures cosmological time dilation using Type Ia supernovae from the Dark Energy Survey by constructing stacked reference light curves matched in rest-frame wavelength and fitting individual supernova light curves via a width parameter. While most supernovae have multi-band coverage, some light curves may lack pre-peak observations, which could affect the accuracy of width estimation and, consequently, the inferred time-dilation signal.

The authors performed a test requiring pre-peak observations and found minimal numerical impact on the fitted exponent b, but they note that this reduction in sample size does not allow a definitive assessment of robustness without pre-peak data. They suggest that future analyses could artificially remove pre-peak data to directly test for bias and stability of the method under such conditions.