Comment on "Discrete Boltzmann Modeling of Sediment Transport in Dam-Break Flows Over Erodible Beds" by Peng et al. (2026): Reproducibility Failures Across Five Validation Cases

The six validations in Peng et al. (2026, https://doi.org/10.1029/2025WR042529) must establish that the stated equations and parameters produce the reported dam-break and sediment-transport responses. An audit of source-code release 1.0.0, controlled reruns, and published figures identifies failures in five cases. In Case 4, all ten configurations terminate by 0.0432 s, before the published 0.50–1.00 s validation states; independent central-processor runs reproduce every graphics-processor stopping iteration. The original calculation produces negative sediment concentration and a bed elevation of −55.9429 m from an initial +0.1 m. In Case 5, aligning three article–code differences increases mean error against the published gauge curves 2.6-fold and changes physical surface elevation by up to 0.0181 m at P2. The program also exports an incorrect surface-elevation quantity. In Case 6, the configuration aligning four differences with the article stops at 0.290 s; correcting the continuity sign alone permits completion but increases error against the published curves. Case 1 uses an inconsistent downstream initialization, and Case 2 combines incompatible specifications with a missing program. Case 3 reproduces the published benchmark. These results invalidate the attribution of the published erodible-bed validation to the stated equations and parameters.