Salt intrusion in the Tweed Estuary

Results are presented from a 2-week field programme in the Tweed Estuary, U.K. Maximum values of the empirically based Estuarine Richardson Number, Ri(E), occurred during neap tides, and minimum values occurred during spring tides. Estimated values of Ri(E) varied between 0.3 and 2.3, suggesting the occurrence of partially mixed to stratified conditions, depending on tidal state and freshwater inflow. These relatively large values of Ri, were consistent with both observed strong salinity stratification and large salt fluxes due to vertical shear transport. Low values (< 0.25) of the estimated gradient Richardson Number, Ri, generally occurred close to the bed on the flood, suggestive of tidal mixing there, and higher (> 0.5) values in the halocline. A velocity maximum occurred within the halocline during the early flood. Wave-like spatial oscillations of the halocline occurred on the ebb. The oscillation troughs were situated above deep holes located just down-estuary of the rail and old road bridges. There was an indication that the constricted flow between the bridges' arches resulted in enhanced mixing of near-surface waters and a thickening of the halocline. It is also possible that these wave-like structures were stationary, near-critical internal lee waves, triggered by the deep holes. Trapping of high-salinity waters occurred on the ebb. Saline pools were isolated within a deep hole or deeper section of bed by the falling halocline. When the salt wedge moved further down-estuary, the 'trapped' waters were subjected to strongly ebbing, overlying freshwater, and were subsequently entrained and flushed. The salinity intrusion was a strong function of spring-neap tidal state and a weaker function of freshwater inflow. The estimated salinity intrusion varied from about 4.7 to 7.6 km during the fieldwork period. The strong dependence on tidal range followed from the comparable lengths of the tidal excursion and salinity intrusion. Long excursion lengths were also partly responsible for the short residence (or flushing) times and their strong dependence on spring-neap tidal state. For typical summer freshwater inflows of 40 m(3) s(-1), the estimated residence times varied from one tide at large spring tides (removal on one ebb) to two tides (removal over two ebbs) at neap tides. Computed residence times were shorter for typical winter inflows. (C) 1996 Academic Press Limited