The scatter, residual correlations and curvature of the SPARC baryonic Tully-Fisher relation

In recent work, Lelli et al. (2016) argue that the tightness of the baryonic Tully-Fisher relation (BTFR) of the SPARC galaxy sample, and the weakness of the correlation of its residuals with effective radius, pose challenges to LCDM cosmology. In this Letter we calculate the statistical significance of these results in the framework of halo abundance matching, which imposes a canonical galaxy-halo connection. Taking full account of sample variance among SPARC-like realisations of the parent halo population, we find the scatter in the predicted BTFR to be 3.6 sigma too high, but the correlation of its residuals with galaxy size to be naturally weak. Further, we find abundance matching to generate BTFR curvature in 3.0 sigma disagreement with the data, and a fraction of galaxies with non-flat rotation curves somewhat larger than observed.