A short note on: Solari, S. & Alonso, R. (2025) “A full wind bivariate bias correction method”. Applied Ocean Research, 161, 104686. https://doi.org/10.1016/j.apor.2025.104686

Solari and Alonso (2025) proposed a two-step bivariate bias correction method for near-surface winds in which the iterative random projection method (IRPM) is first used to match a modelled multivariate distribution to an observed one and a Gaussian process emulator is then fitted so the learned correction can be applied to new data. This comment clarifies how that workflow relates to the existing multivariate bias correction by n -dimensional probability density function transfer (MBCn) algorithm of Cannon (2018): (i) IRPM already forms the core iterative loop of MBCn, and (ii) the quantile delta mapping plus rank-reordering extension of MBCn can correct new data while preserving modelled changes in the original marginal quantiles, without requiring a separate Gaussian process emulator. Unlike a fixed emulator, the MBCn correction of a new vector depends on the empirical sample in which it is embedded. In this context, emulators can still be useful for the sake of computational efficiency, or when a continuous, differentiable, or smoothly interpolated state-space mapping is needed.