Effects of Rotation on Convective Plumes from Line Segment Sources

View More View Less
  • 1 NOAA Pacific Marine Environmental Laboratory, Seattle, Washington
  • | 2 Institute of Marine Sciences, University of California, Santa Cruz, Santa Cruz, California
© Get Permissions
Full access

Abstract

Effects of rotation on finite-length line plumes are studied with a three-dimensional nonhydrostatic numerical model. Geophysical convection with this source geometry occurs, for example, as the result of fissure releases of hot hydrothermal fluids at the seafloor from terrestrial release of hot gases and ash during volcanic activity along fissures and in the descent from the sea surface of brines formed during freezing of ice leads at high latitudes. Here the model treats the case of a starting plume of dense fluid descending into a rotating environment. Results are compared with laboratory experiments so that the validity of the model, particularly the nonlinear subgrid-scale mixing formulation, might first be established. Differences in plumes caused by varying rotation rate, &ohm, and buoyancy flux, B0, are the primary focus, with experiments in fluid of depth h spanning a convective Rossby number [B01/3/(2Ωh)] of 0.01−1.0. Rotation initiates spiraling of the descending plumes but it has little effect on the speed of plume descent; the latter depends on the strength of turbulent mixing. Low rotation rates allow the descending plume cap to be broad and the stem to be narrow. Higher rotation rates retard the lateral spread of the plume cap and widen the plume stem. Updraft at the stem edge is very much larger at higher rotation rates, and that appears to be instrumental in determining stem and cap width. Values of turbulent mixing coefficients within the plume are dependent on B0 but not on Ω. Thus rotational effects on turbulence are not needed to account for differences in plume structure arising solely from Ω variation. Agreement between model and laboratory results did not occur without a nonlinear time- and space-dependent subgrid-scale mixing parameterization, suggesting that model applications to convective geophysical problems identified above require the same.

Abstract

Effects of rotation on finite-length line plumes are studied with a three-dimensional nonhydrostatic numerical model. Geophysical convection with this source geometry occurs, for example, as the result of fissure releases of hot hydrothermal fluids at the seafloor from terrestrial release of hot gases and ash during volcanic activity along fissures and in the descent from the sea surface of brines formed during freezing of ice leads at high latitudes. Here the model treats the case of a starting plume of dense fluid descending into a rotating environment. Results are compared with laboratory experiments so that the validity of the model, particularly the nonlinear subgrid-scale mixing formulation, might first be established. Differences in plumes caused by varying rotation rate, &ohm, and buoyancy flux, B0, are the primary focus, with experiments in fluid of depth h spanning a convective Rossby number [B01/3/(2Ωh)] of 0.01−1.0. Rotation initiates spiraling of the descending plumes but it has little effect on the speed of plume descent; the latter depends on the strength of turbulent mixing. Low rotation rates allow the descending plume cap to be broad and the stem to be narrow. Higher rotation rates retard the lateral spread of the plume cap and widen the plume stem. Updraft at the stem edge is very much larger at higher rotation rates, and that appears to be instrumental in determining stem and cap width. Values of turbulent mixing coefficients within the plume are dependent on B0 but not on Ω. Thus rotational effects on turbulence are not needed to account for differences in plume structure arising solely from Ω variation. Agreement between model and laboratory results did not occur without a nonlinear time- and space-dependent subgrid-scale mixing parameterization, suggesting that model applications to convective geophysical problems identified above require the same.

Save