Wave Climate Signals Are Obscured by Observational Errors and Sampling Bias of Voluntary Observing Ships

Abstract

Visual wave observations from Voluntary Observing Ships (VOSs) have long been regarded as a useful data source for studying the long‐term evolution of the global wave climate due to their unique multi‐decadal continuity. However, by integrating satellite altimetry, buoy records, and an ERA5‐based Virtual Observation framework, this study demonstrates that the observational errors and sampling bias significantly obscure long‐term trends in ocean waves in VOS data. Results reveal substantial intrinsic uncertainties in VOS observations, with overall significant wave height uncertainty exceeding 1.0 m and exceeding 2 m in some regions. While the Law of Large Numbers allows for reasonable approximations of long‐term mean sea states in high‐traffic corridors, high‐percentiles (90th and 99th percentiles) are systematically underestimated due to the inherent statistical challenges of capturing transient extreme events through sparse sampling. Crucially, even if the uncertainty of VOS observations is ignored, VOS sampling patterns can introduce substantial discrepancies in ERA5‐derived long‐term trend estimates, with sampling‐induced artifacts sometimes comparable to climate signals in the reference field. We conclude that while VOSs remain an invaluable historical record for the pre‐satellite era, they possess severe limitations for monitoring global wave climate changes.