A simple theory that predicts the vertical structure and offshore spreading of a localized buoyant inflow onto a continental shelf is formulated. The theory is based on two competing mechanisms that move the buoyant fluid offshore: 1) the radial spread of the lighter water over the ambient water, being deflected by the Coriolis force and producing an anticyclonic cyclostrophic plume, and 2) offshore transport of buoyant water in the frictional bottom boundary layer that moves the entire plume offshore while maintaining contact with the bottom. The surface expression of the cyclostrophic plume moves offshore a distance y s ϭ 2(3gЈh 0 ϩ)/(2gЈh 0 ϩ) 1/2 f, 2 2 i i where gЈ is reduced gravity based on the inflow density anomaly, h 0 is the inflow depth, i is the inflow velocity, and f is the Coriolis parameter. The plume remains attached to the bottom to a depth given by h b ϭ (2L i h 0 f /gЈ) 1/2 , where L is the inflow width. Both scales are based solely on parameters of the buoyant inflow at its source. There are three possible scenarios. 1) If the predicted h b is shallower than the inflow depth, then the bottom boundary layer does not transport buoyancy offshore, and a purely surface-advected plume forms, which extends offshore a minimum of more than four Rossby radii. 2) If the h b isobath is farther offshore than y s , then transport in the bottom boundary layer dominates and a purely bottom-advected plume forms, which is trapped along the h b isobath. 3) If the h b isobath is deeper than the inflow depth but shoreward of y s , then an intermediate plume forms in which the plume detaches from the bottom at h b and spreads offshore at the surface to y s. The theory is tested using a primitive equation numerical model. All three plume types are reproduced with scales that agree well with the theory. The theory is compared to a number of observational examples. In all cases, the prediction of plume type is correct, and the length scales are consistent with the theory.

Abstract. The impact of buoyant discharge variations on the dynamics of coastal buoyancydriven currents is studied using a primitive equation numerical model (SPEM5). First, variable discharge is introduced as harmonic fluctuations of the inflow velocity at the tidal (period 12 hours) and subinertial (period 10 days) frequencies. Tidal fluctuations produce only minor effects on the buoyant plume compared to the case of constant inflow, while subinertial fluctuations substantially modify the anticyclonic bulge. A partially detached anticyclonic plume forms when discharge subsides after reaching its peak value. Such a plume has maximum offshore extension some distance downstream of the mouth with the lightest water separated from the coast. A secondary bulge forms during the low runoff interval. When high discharge resumes, this secondary bulge is shifted offshore and enhanced for some time. An individual high-discharge event is next considered, where both the net transport of the inflow and the absolute value of its density anomaly increase and then return to their initial (background) values over 5 and 10 day time intervals. This event also generates a partially detached plume (especially with the 10 day duration). In this case, the lightest water occupies the downstream part of the bulge and is separated not only from the coast but also from the mouth. The effect of variable discharge is more dramatic with a uniform downstream current of 0.1 m s -•. Under such conditions, constant buoyant discharge does not form a well-pronounced anticyclonic bulge. In contrast, variable discharge produces an almost circular anticyclone during the high-runoff interval. As runoff decreases, this anticyclone separates from the source and either continues to propagate downstream as an individual eddy or is modified by the next cycle of increasing discharge. Observational evidence for both the partially detached bulge near the mouth and the anticyclone propagating downstream from its source is presented in this study. One feature was observed at the mouth of the Columbia River estuary; the second feature was observed off the southern New Jersey coast -150 km south of its source, the Hudson estuary.

No abstract

Subinertial dynamics on the inner New Jersey shelf is examined using time series of the forcing agents (atmospheric pressure, wind stress, and Hudson River streamflow), adjusted sea level (ASL) along the southern part of the Mid-Atlantic Bight, and mooring data collected during the summer of 1996. High-frequency (period 1-3 days) transient wind-driven events were evident both in ASL and alongshelf current data. ASL events propagated southward with remarkably high speed (ϳ10 m s Ϫ1) in the manner of free coastally trapped waves (CTW). However, these transients were forced by the wind events within the study domain: both ASL and alongshelf current fluctuations were coherent with the local alongshore wind stress. ASL amplitude substantially increased downshelf (southward). These transient flows propagated from the corner in the coastline formed by the southern Long Island and northern New Jersey coasts. This bend of the coastline created a discontinuity in the alongshore wind stress component that caused the generation of CTW pulses at this location. During the period of observations, enhanced buoyant flows arrived at the site of the moorings. They were associated with increased Hudson River discharge. These buoyant flows and transient wind-driven events strongly interacted: transient wind-driven currents were dramatically amplified in the buoyant water while the buoyant water was spread offshore. Amplified transient currents were not associated with the enhanced vertical shear. Lower-frequency wind forcing generated upwelling events with typical duration of 8-10 days. During the upwelling, temperature dropped through the whole water column, but the stratification remained significant (5Њ-6ЊC in 8-10 m of water). Even though upwelling-favorable winds dominated, record-mean currents in the upper layer were weak (2-5 cm s Ϫ1) due to the close competition between wind and buoyancy forcing.