Abstract Density gradients in coastal regions with significant freshwater input are large and variable and are a major control of nearshore circulation. However their measurement is difficult, especially where the gradients are largest, close to the coast, with significant uncertainties because of a variety of factors – time and spatial (horizontal and vertical) scales are small, tidal currents are strong and water depths shallow. Whilst temperature measurements are relatively straightforward, measurement of salinity (the dominant control of spatial variability for density) can be less reliable in turbid coastal waters. The nearshore density gradients in Liverpool Bay are investigated using an integrated multi-year data set from an in situ buoy, instrumented ferry and HF radar. The ferry is particularly useful for estimating coastal density gradients since measurements are made right from the mouth of Mersey, where gradients are on average 3×10−4kgm−4. Using measurements at the single in situ site by the Mersey Bar, 17km from land, density gradients can be estimated from the tidal excursion or by using ferry data; both giving average values of 5×10−5kgm−4. Nine years of surface salinity measurements there show no evidence of predominant periodicities, although there is a weak annual cycle, and no consistent relation with storms or floods, leading to the conclusion that the majority of the Mersey plume, for most of the time, lies closer to the English shore than the Mersey Bar. Liverpool Bay's circulation is the dominant factor, with wind forcing tending to reinforce it for wind speeds greater than 5–10ms−1. Near bed currents are consistently shoreward and near surface currents northward.