Salinity Intrusion in Estuaries

View More View Less
  • 1 Institute of Oceanographic Sciences, Bidston Observatory, Merseyside, England L43 7RA
© Get Permissions
Full access

Abstract

One dimensional time-averaged solutions are examined for salinity intrusion in estuaries with a breadth variation BL(X/λ)n and a depth variation H(X/λ)m, where X is the distance from the head of the estuary. These solutions emphasize the importance of the rate of change of cross-sectional area in determining salinity distribution.

Assuming a constant longitudinal-dispersion coefficient Dx = D, the salinity distribution is shown to be highly dependent on the dimensionless parameter V′ = U1 × X1/D, with U1 the velocity of the fresh-water flow at position X1, where the estuary is effectively at oceanic salinity. [This parameter V′ is equivalent to the flushing number F introduced by Arons and Stommel (1951) for the case of an estuary of rectangular cross section.] For eight estuaries, comparisons are made between calculated and observed salinity distributions, where for each estuary the value of D in the calculated distribution was chosen to produce the best agreement with the observed distribution. For six of the eight estuaries, the chosen value of D was within the range 50 m2 s−1<D<500 m2 s−1, in good agreement with corresponding values found in previous studies. However, it is shown that the salinity distribution is highly sensitive to the specified value of D, implying that the usefulness of the one-dimensional, time-averaged solutions may be somewhat restricted.

Theoretical distributions of salinity also were obtained for Dx = D1dc/dx and Dx = D2(dc/dx)2, where dc/dx represents the time-averaged longitudinal salinity gradient and D1 and D2 are constant coefficients. While reasonable agreement is again obtained with observed distributions, certain limitations in the application of these two forms for Dx are shown.

Attempts to derive a more rational dimensionless form for Dx in terms of gross estuarine parameters proved unsuccessful.

Abstract

One dimensional time-averaged solutions are examined for salinity intrusion in estuaries with a breadth variation BL(X/λ)n and a depth variation H(X/λ)m, where X is the distance from the head of the estuary. These solutions emphasize the importance of the rate of change of cross-sectional area in determining salinity distribution.

Assuming a constant longitudinal-dispersion coefficient Dx = D, the salinity distribution is shown to be highly dependent on the dimensionless parameter V′ = U1 × X1/D, with U1 the velocity of the fresh-water flow at position X1, where the estuary is effectively at oceanic salinity. [This parameter V′ is equivalent to the flushing number F introduced by Arons and Stommel (1951) for the case of an estuary of rectangular cross section.] For eight estuaries, comparisons are made between calculated and observed salinity distributions, where for each estuary the value of D in the calculated distribution was chosen to produce the best agreement with the observed distribution. For six of the eight estuaries, the chosen value of D was within the range 50 m2 s−1<D<500 m2 s−1, in good agreement with corresponding values found in previous studies. However, it is shown that the salinity distribution is highly sensitive to the specified value of D, implying that the usefulness of the one-dimensional, time-averaged solutions may be somewhat restricted.

Theoretical distributions of salinity also were obtained for Dx = D1dc/dx and Dx = D2(dc/dx)2, where dc/dx represents the time-averaged longitudinal salinity gradient and D1 and D2 are constant coefficients. While reasonable agreement is again obtained with observed distributions, certain limitations in the application of these two forms for Dx are shown.

Attempts to derive a more rational dimensionless form for Dx in terms of gross estuarine parameters proved unsuccessful.

Save