The effect of bedforms on frictional roughness felt by the overlying flow is crucial to the regional modelling of estuaries and coastal seas. Bedforms are also a key marker of palaeoenvironments. Experiments have shown that even modest biotic and abiotic cohesion in sand inhibits bedform formation, modifies bedform size, and slows bedform development, but this has rarely been tested in nature. The present study used a comprehensive dataset recorded over a complete spring–neap cycle on an intertidal flat to investigate bedform dynamics controlled by a wide range of wave and current conditions, including the effects of wave–current angle and bed cohesion. A detailed picture of different bedform types and their relationship to the flow, be they equilibrium, non-equilibrium, or relict, was produced, and captured in a phase diagram that integrates wave-dominated, current-dominated, and combined wave–current bedforms. This bedform phase diagram incorporates a substantially wider range of flow conditions than previous phase diagrams, including bedforms related to near-orthogonal wave–current angles, such as ladderback ripples. Comparison with laboratory-derived bedform phase diagrams indicates that washed-out ripples, lunate interference ripples and upper-stage plane beds replace the subaqueous dune field; such bedform distributions may be a key characteristic of intertidal flats. The field data also provide a means of predicting the dimensions of these bedforms, which can be transferred to other areas and grain sizes. We show that an equation for the prediction of equilibrium bedform size is sufficient to predict the roughness, even though the bedforms are highly variable in character and only in equilibrium with the flow for approximately half the time. Whilst the effect of cohesive clay is limited under more active spring conditions, clay does play a role in reducing the bedform dimensions under more quiescent neap conditions. We also investigated which combinations of waves, currents, and bed clay contents in the intertidal zone have the highest potential for bedform preservation in the geological record. This shows that combined wave–current bedforms have the lowest preservation potential and equilibrium current ripples have the highest preservation potential, even in the presence of moderate and storm waves. Hence, the absence of wave ripples and combined-flow bedforms and their primary stratification in sedimentary successions cannot be taken as evidence that waves were absent at the time of deposition.