Dynamics of diffusive layering and chemocline formation in Lake Kivu and brine pools

Abstract

To manage Lake Kivu in Central Africa to avoid its future eruption because of accumulated methane and carbon dioxide, a better understanding was required of the diffusive layering (DL) mode of double-diffusive convection (DDC) that governs the vertical transport mechanisms below the upper mixed layer (mixolimnion). Through emphasis on the physical mechanisms involved in this phenomenon, this work proposes a new understanding of which are the independent parameters in DL systems together with a hitherto missing explanation of the formation and long-term stability of subsea brine pools.Based on a review of observational data and of the basis for the existing hypothesis on the vertical transport of heat and mass in DL systems, a new hypothesis is proposed according to which the independent parameters are dβS0/dz, the solute density gradient, and geothermal heat flux, different from those hitherto believed (T and Rρ, the density ratio). Furthermore, diffusive and convective heat transport mechanisms balance to generate a constant heat flux in the DL systems. This causes an imbalance in the mass transport capacity that slowly causes some zones to mix and expand, thereby thinning the gradient zones. The steady state of this slow dynamic mechanism is the formation and maintaining of the completely mixed zones and the ultimately thinned gradient zones, that retain solutes below them and are found in deep brine pools, e.g. in the Mediterranean Sea and in the Red Sea.These results are of importance for understanding vertical transport phenomena in DL systems; all literature on this subject since 1965 likely needs to be revisited. Much theory likely will not hold and experimental results likely are insufficient if heat flux and the dimensions of the single ΔT and ΔS interfaces have not been measured. Furthermore, attempting to describe both the fingering and the DL modes of DDC by the same transport equations seems inappropriate. These conclusions probably also apply to DL systems in the Arctic Ocean, thus potentially influencing models of water circulation in the Atlantic Ocean.