Zurück zur Übersicht

Double diffusion in Lake Kivu

Publikationsart Nicht peer-reviewed
Publikationsform Andere Publikationen (nicht peer-reviewed)
Publikationsdatum 2013
Autor/in Sommer Tobias,
Projekt Lake Kivu - turbulence and double diffusion in permanent stratification
Alle Daten anzeigen

Andere Publikationen (nicht peer-reviewed)

Buch Double diffusion in Lake Kivu
Verlag ETH Zürich, Zürich


Double diffusion enhances the vertical transport within a stratified water body by transforming gradual density gradients into staircases of convectively mixed layers separated by high-gradient and gravitationally stable interfaces. Those staircases can develop when two agents, which diffuse at different rates, contribute in opposing ways to the vertical density gradient. This study focuses on the diffusive regime of double diffusion, where the slower diffusing agent (here dissolved substances) is stabilizing. Estimating the vertical fluxes through double-diffusive staircases based on easily accessible parameters has been one goal of double diffusion research in the past. So far, flux laws are based on scaling arguments which are calibrated by empirical results from laboratory experiments and their validity in natural environments is questionable. Double-diffusive heat fluxes are of particular interest in the Arctic Ocean, where they control the heat exchange between a warm intrusion from the North Atlantic and the overlying sea ice. A more detailed understanding of double diffusion in natural environments is thus essential to test well-established theories. In this study the presently most detailed in-situ measurements of natural double- diffusive staircases are combined with cutting-edge Direct Numerical Simulations (DNS) in order to improve our understanding of the dynamics between interfaces and mixed layers and to test existing flux laws. Lake Kivu is an ideal study environment for double diffusion with up to 300 clearly distinguishable double-diffusive steps distributed almost over the entire permanently stratified water column below ~100 m depth. It is located close to the equator in the East African Rift Valley at the border between Rwanda and the Democratic Republic of the Congo. During two field campaigns in 2010 and 2011, 225 microstructure profiles of temperature and conductivity were measured resulting in a total profiling length of ~55 km. Our data set comprises 9,401 fully resolved interfaces and adjacent mixed layers. This study shows that a full resolution of the smallest interface thicknesses requires a correction for the responses of the microstructure sensors. The frequency responses of the FP07 temperature and the SBE-7 conductivity sensor are derived from observed differences in the statistical distributions of interface thicknesses at various profiling speeds. Heat and salt fluxes through double-diffusive staircases are found to be sensitive to the frequency response corrections. This implies that the conductivity sensor is not “infinitely” fast, as previously assumed. Temperature interfaces are shown to be thicker than salinity interfaces. The buoyancy stored in the resulting boundary layers is transported by plumes to the mixed layers. The interaction between the mixed layer and the boundary layers is time dependent and additionally influenced by mixed layer motions and thus cannot be understood by linear stability analysis. The most common heat flux parameterization is tested and a correction introduced that depends on the Rayleigh number of the double-diffusive system. Applying the correction to studies in the Arctic reduces previous heat flux estimates by up to a factor of four. Two-dimensional DNS is shown to reproduce the interface thicknesses of in-situ microstructure profiles and molecular heat fluxes through interfaces capture the total vertical heat fluxes for density ratios larger than three. In summary, this study demonstrates the value of in-situ measurements, in particular in combination with DNS, for testing existing theories. With increasing computation power and sensor development, such interdisciplinary approaches will become even more attractive in the future and might eventually lead to new models and theories which are closer to reality than existing ones.