BULK-PHASE VISCOELASTIC PROPERTIES OF SEAWATER

Deformation and mixing in any medium are controlled by its viscosity a nd elasticity. The present study provides preliminary information, at oceanic shear rates, on the viscosity and elasticity in the bulk phase of seawater. Thirty-two samples of seawater, obtained on cruises at f our times of year in a meso-oligotrophic part of the Mediterranean, we re measured for bulk-phase viscoelastic properties in an oscillating-s hear Couette flow with measuring gap 0.5 mm, over shear rates gamma fr om 0.0021 to 0.286 s-1. Fifteen samples were obtained on a single crui se in the German Bight, North Sea, when Phaeocystis blooms were taking place, and these were similarly measured at gamma from 0.0021 to 0.97 3 s-1. The bulk-phase measurements suffered interference contributed b y the surface film. The interference mechanism was investigated, and a method was developed to correct for it, which permitted extraction of the bulk-phase viscosity. While extraction of bulk-phase elastic effe cts was possible only in the most thickened Mediterranean samples, in the generally thicker North Sea samples bulk-phase elasticity was clos ely related to bulk-phase excess (polymeric) viscosity.At gamma = 0.00 21 s-1, viscosity, eta, in the Mediterranean samples ranged from 0.17 to 19 (mean 3.5, n = 23) times the average solution viscosity, eta(w)B AR which is gamma-independent and contributed principally by water and salt. Corresponding values for the German Bight samples were 0.99 to 127 (mean 52, n = 5) times eta(w)BAR. For each cruise, the overall exc ess viscosity, eta(E) = eta - eta(w), showed a power-law relationship with gamma such that eta(E) = k.gamma(-P). For the five different crui ses, P varied from 1.1 to 1.5. Such high values of P indicate that thi ckening is contributed principally by cross-linked polymer gel rather than by overlapping chains. The elastic modulus G' was very variable, with sample maxima of 100 and 300 muPa in the Mediterranean and North Seas respectively. Thickening (determined as both eta(E) and G') was h eterogeneous, with a variability coefficient (SD/mean) from 1 to 3 for different cruises and values of gamma. In the three Mediterranean cru ises in which duplicate measurements were made for the same samples, a ll the variability could be accounted for by in-sample variability, su ggesting that centimetre-scale flocculation-type processes were respon sible. Gaussian standard deviation SD(G) was constant (indicating self -similarity) for the higher values of eta(E) both the Mediterranean an d North Sea studies a well as for G' in the North Sea study. At the lo wer end of the eta(E) and G' distributions, SD(G) was higher, due to e xperimental variability. Over a range of appropriate shear rates, the turbulent (Kolmogorov) length scale L was calculated using the mean vi scosities measured. The calculations suggest that, when turbulence is low or biological activity high, L is considerably higher than values previously supposed, and mixing correspondingly reduced. The heterogen eous nature of both eta(E) and G' must furthermore increase intermitte nce in both turbulence and mixing. It is concluded that a significant part of the sea functions as a lumpy, biopolymeric gel in which small- scale (less-than-or-equal-to 100.L) flow and mixing are under strong b iological influence.