Near surface wind—conventionally measured 10 m above the ground, U10—is a fundamental variable for many processes acting in bays and estuaries. For practical reasons, wind in bays and estuaries is often measured at the land‐water interface, e.g., on docks, piers, and jetties. By analyzing time series of U10 from 15 stations in the Mississippi Delta, Louisiana, we show that wind measurements at the land‐water interface are biased and not representative of the wind over estuarine waters. Specifically, measurements at the land‐water interface underestimate the speed of winds blowing from land, because even short (<1 km) stretches of land with large bed roughness can significantly reduce U10. Wind measured next to large channels is biased not only in speed but also in direction, which tends to align with the channel axis. We suggest that very local (<1 km) variability in the boundary layer properties contributes to the large spatial variability of the measured estuarine wind field. Simple anisotropic corrections based on bed roughness should be applied to measurements made at the land‐water interface to estimate wind over adjacent estuarine waters. Alternatively, wind measured at offshore stations—even if located hundreds of km away—provide reliable estimates of estuarine wind. A more systematic use of wind measurements would likely increase the accuracy of estuarine numerical models and allow for more direct comparison between different estuarine systems.