Turbulence closure modeling of estuarine stratification

Abstract

Transient stratification in estuaries reflects competition between the stratifying influences of the vertical gravitational circulation and elastic straining of the longitudinal density gradient by vertical shear, set against the mixing influence of, principally, tidally generated turbulence. A one‐dimensional (vertical) numerical model of an estuary, using seven different parameterizations of vertical mass and momentum mixing coefficients from the literature, is used to make general predictions about the nature of the time‐dependent signal of stratification in an estuary. These predictions show good agreement with published predictions from simpler models and they demonstrate that, at higher frequencies (e.g., first and second species tides) stratification is produced primarily from elastic straining but that, at lower frequencies (due to the fortnightly spring‐neap modulation, for example) the estuarine gravitational circulation tends to dominate. The same model, with rotation included, is then used to hindcast a time series of stratification from Spencer Gulf, South Australia, in order to test parameterizations of mixing coefficients against real data. The character of the semidiurnal and diurnal stratification signal is better predicted by some schemes than others. Against qualitative and quantitative criteria, the Mellor and Yamada (1982) and Pacanowski and Philander (1981) schemes perform best. All schemes, however, underpredict the amplitude of the stratification signal at all timescales. At longer timescales, associated with the springs‐neaps cycle, this is believed to be due to a difference in the degree of geostrophic adjustment of the model and the observations, produced by the three‐dimensional topography of the observational site (Spencer Gulf, South Australia).

Improving the parameterisation of horizontal density gradient in one-dimensional water column models for estuarine circulation

“…Furthermore, they generate simple results, which permits to easily understand the key processes and quickly establish diagnoses. However, one common failure of these models is the generation of runaway stratification: when the tidal amplitude is low, stratification tends to grow without bound due to an inadequate parameterisation of horizontal density gradient (Nunes Vaz and Simpson, 1994;Warner et al, 2005). This paper shows that simple analytical developments can lead to a new version of the model which keeps stratification under control.…”

“…Several models were applied to simulate and understand the evolution of stratification in estuaries. Linear prescriptive models were first used (Simpson et al, 1991;Nunes Vaz and Simpson, 1994;Scott, 2004). Then, several authors turned to one-dimensional water column non linear models Monismith and Fong, 1996;Nunes Vaz and Simpson, 1994;Lucas et al, 1998).…”

“…Linear prescriptive models were first used (Simpson et al, 1991;Nunes Vaz and Simpson, 1994;Scott, 2004). Then, several authors turned to one-dimensional water column non linear models Monismith and Fong, 1996;Nunes Vaz and Simpson, 1994;Lucas et al, 1998). Recently, three-dimensional models were used to simulate estuarine flows (Burchard and Baumert, 1998;Hetland and Geyer, 2004;Warner et al, 2005).…”

“…Such finite-element models and their advantages were described by Hanert et al (2006Hanert et al ( , 2007. As mixing is a key player in the evolution of the stratification (Nunes Vaz and Simpson, 1994), we use the Mellor and Yamada level 2.5 turbulence closure Yamada, 1974, 1982;Galperin et al, 1988) which is well suited for the prediction Fig. 1.…”

“…of stratification in estuaries (Nunes Vaz and Simpson, 1994). This turbulence closure was recently implemented using the finite-element method for one-dimensional and three-dimensional models.…”