Title: Hubble tension tomography: BAO vs SnIa distance tension

Abstract: We investigate the redshift dependence of the Hubble tension by comparing the luminosity distances obtained using an up-to-date BAO dataset (including the latest DESI data) calibrated with the CMB-inferred sound horizon, and the Pantheon+ SnIa distances calibrated with Cepheids. Using a redshift tomography method, we find: 1) The BAO-inferred distances are discrepant with the Pantheon+ SnIa distances across all redshift bins considered, with the discrepancy level varying with redshift. 2) The distance discrepancy is more pronounced at lower redshifts ($z \in [0.1,0.8]$) compared to higher redshifts ($z\in [0.8,2.3]$). The consistency of $\Lambda$CDM best fit parameters obtained in high and low redshift bins of both BAO and SnIa samples is investigated and we confirm that the tension reduces at high redshifts. Also a mild tension between the redshift bins is identified at higher redshifts for both the BAO and Pantheon+ data with respect to the best fit value of $H_0$ in agreement with previous studies which find hints for an 'evolution' of $H_0$ in the context of $\Lambda$CDM. These results confirm that the low redshift BAO and SnIa distances can only become consistent through a re-evaluation of the distance calibration methods. An $H(z)$ expansion rate deformation alone is insufficient to resolve the tension. Our findings also hint at a possible deviation of the expansion rate from the Planck18/$\Lambda$CDM model at high redshifts $z\gtrsim 2$. We show that such a deformation is well described by a high redshift transition of $H(z)$ like the one expressed by $\Lambda_s$CDM even though this alone cannot fully resolve the Hubble tension due to its tension with intermediate/low $z$ BAO data.