A numerical model of the internal tide in knight inlet,British Columbia

Abstract

A two‐dimensional, hydrostatic numerical model of the tides in Knight Inlet is compared with observations of velocity and density obtained from three cyclesonde moorings. The observations from a fourth cyclesonde mooring were used to provide boundary data at the open end of the model. The time period in the fjord that the model simulates was a period of high, freshwater runoff, so that the fjord had a distinct, surface layer. The use of high, vertical resolution was avoided by attaching a homogeneous, fresh, surface layer to the top of the model. The density equation was linearized about a mean, fixed density field, and the mixing of density was not allowed.

The model reproduces the semidiurnal (M₂, S₂ and N₂) and diurnal (K₁ and O₁) velocity and density signals in the inlet. The shallow‐water constituents (M₄ and MK₃) are reproduced even though the density equation has been linearized. The fortnightly constituent (MSf) is poorly simulated. When the advection terms in the momentum equation are set to zero, the basic features of the semidiurnal and diurnal constituents are still reproduced, but the shallow‐water constituents are poorly simulated.

The energy flux along the inlet of the M₂ internal tide is insensitive to the advective terms in the momentum equation. The total rate of dissipation of M₂ energy is similar to the energy flux in the M₂ internal tide near the sill, which implies that, according to the model, most of the energy removed from the barotropic tide is fed into the internal tide. The majority of the energy in the M₂ internal tide is dissipated close to the sill of the inlet, but enough of the energy makes its way to the head of the inlet to reflect and set up a recognizable standing wave pattern.