Laboratory Models of Double-Diffusive 

 Processes in the Ocean 



J. Stswart Turnsr 



Australian National University 

 Canberra, Australia 



ABSTRACT 



There is now good observational evidence to support 

 the ideas that double-diffusive processes', i.e., 

 those for which the differential diffusion of heat 

 and salt are important, can affect the rates of 

 vertical transport of these properties in the ocean, 

 and are responsible for the formation of certain 

 types of microstructure. Much of our detailed 

 understanding of these effects has come from related 

 laboratory experiments, but new phenomena are still 

 being discovered which are as yet untested by direct 

 measurements in the ocean. It is the purpose of 

 this paper first to review the background to this 

 subject, and then to describe the more recent experi- 

 ments which suggest further double-diffusive effects 

 likely to be significant in various oceanographic 

 contexts . 



A convenient laboratory technique has been to 

 use two solutes (commonly salt and sugar) to model 

 the T-S variations; some of these experiments with 

 closer diffusivities are in fact directly relevant 

 to the ocean. When more than two diffusing compo- 

 nents are present it has been shown that even small 

 differences in molecular diffusivity can signifi- 

 cantly affect the relative rates of transfer of 

 solutes through an interface, and this should be 

 considered more carefully in geochemical studies. 

 Strong double-diffusive layering is often associated 

 with large horizontal gradients of T and S, and 

 related effects have been studied in our laboratory 

 in three different geometries: the circulation 

 produced by a block of ice in a salinity gradient; 

 a line source of one fluid intruding at its own 

 density level into a gradient with different prop- 

 erties; and the spreading across a frontal surface 

 separating two fluids having the same vertical 

 density but different T-S structures. 



1 . INTRODUCTION 



It is past the stage when the relevance of double- 

 diffusive effects has to be justified ab initio to 

 an audience of oceanographers . Over the last few 

 years , there have been many observations of fine- 

 structure and microstructure in the deep ocean 

 which can only be explained in these terms. Wherever 

 there is a systematic association between T and S 

 variations, with both properties increasing or 

 decreasing together (so that their effect on the 

 density is in opposite senses) , then it is clear 

 that the difference in molecular diffusivities for 

 heat and salt can affect the vertical structure 

 and the transports of the two properties. It is 

 not then sufficient to base predictions of mixing 

 on the net density distribution alone. 



Our understanding of these processes has been 

 greatly influenced by related laboratory experiments 

 [see Turner (1973, 1974)]. Much of the detailed 

 work has concentrated on the properties of sharp in- 

 terfaces separating relatively well-mixed layers: 

 it has been shown that when there are compensating 

 T-S gradients, a smoothly stratified water column 

 typically breaks up into a series of steps, and 

 molecular processes must be more important across 

 such interfaces . Once layers have formed there 

 remains little doubt that the coupled transports 

 can be estimated using the laboratory results. It 

 is much less certain, however, that the processes 

 of formation of layers have always been adequately 

 modelled in the laboratory, where most of the experi- 

 ments have been one-dimensional in form. 



More recent experiments [Turner and Chen (1974) , 

 Huppert and Turner (1978) , Turner (1978) ] have begun 

 to explore a variety of two-dimensional effects, and 

 it is these which will be given most attention in 

 the verbal presentation of this paper. It should be 



596 



