Volume 48 (1996) Number 2 
155 
Two parameters are essential for such esti 
mates: bottom roughness, i. e. the question whether 
there are hydrodynamically rough or smooth condi 
tions, and friction velocity «. which can be estimated 
by the relation 
with U g the geostrophic velocity above the BBL 
(Armi and Millard [1976]). The D1 moorings give ty 
pical values for U g , about 3 cm/s for periods of slow 
motion and up to 10 cm/s for periods with strong 
currents (Klein [1993]). 
Regarding bottom roughness, we distinguish 
between areas covered with fine sediments and 
areas with manganese nodule coverage. The diffe 
rence between hydrodynamically rough and smooth 
conditions is given by the Reynolds number: 
where * s is the roughness element or grain dia 
meter, and v the kinematic molecular viscosity for 
water (=1 x 10 -6 m 2 s~ 1 ). Under smooth conditions, 
R c is smaller than 3, in rough conditions R c is >70. 
The critical factor is the grain size, i.e. the nodule 
diameter. Even during periods of relatively strong 
currents >10 cm/s R c is smaller than 3 (hydrodyna 
mically smooth conditions) if the area is covered 
only with sediment. Rough conditions occur only if 
the bottom is covered with manganese nodules. 
The deepest layer close to a smooth sea 
bottom is the very thin viscous sublayer (VSL). Its 
thickness Z^ s is only about 1-2 cm at U s = 3 cm/s 
(Wimbush [1976]) but this is nevertheless significant 
for animals living close to the seabed (Gage and 
Tyler [1991]). 2^ s will decrease at an increase of U r 
If any structure, e. g. a manganese nodule or marine 
organism, exeeds one third of 2the viscous flow is 
disrupted and becomes turbulent. 
The logarithmic layer (LL) lies above the 
VSL. The velocity at height z, i/(z), is given by 
U{z) = -\n^r (3) 
where * is the von Karmans constant (= 0.4) and Z„ 
the roughness element (Bowden [1978]). Under hy 
drodynamically rough conditions 4 amounts to 
*/30, in the hydrodynamically smooth case Z 0 is 
9v!u„ The height of LL is about 1 m. Within LL, the 
Coriolis force is negligible compared to viscosity 
and the stress is assumed to be constant. 
The height of the turbulent Ekman Layer (EL) 
can be approximated by 
h t = 0.4 — (4) 
/ 
with Coriolis parameter /= 1.65 x 10~ 5 1/s at the 
SEDIPERU site, and/= 1.77 x 1Q- 5 1/s at the DEA 
(Bowden [1978]). Ekman’s theory predicts a 
decrease of velocity from the top of EL towards the 
bottom and veering of the current direction by 45° to 
the right (at southern latitudes) compared to the 
geostrophic flow above EL This veering cannot be 
observed in our data because there is no informa 
tion about the current direction below 13 mab (see 
Table 2). /j e versus U e is shown in Fig. 7. Another re 
lation for h c is by the Ekman theory: 
h -- x i/W (5) 
where A z is the vertical eddy diffusivity. Using [4] 
and [5] it is possible to estimate the order of magni 
tude of A z inside the EL For U g = 1 cm/s A z is of the 
order of 1 cm 2 /s, for U s = 10 cm/s A z amounts to 
about 50 cm 2 /s. 
Including both LL and EL, the bottom mixed 
layer (BML) is defined by its hydrographical fine 
structure. Therefore, CTD probes and/or water 
samples are necessary to trace the height of this 
layer. Inside the BML, potential temperature, sali 
nity, and other physical parameters are well mixed 
due to bottom friction. 
CTD profiles from DEA obtained in 1992 give 
BML heights between 100 and 200 m (Klein 
[1993]). In 1996, during R/V ‘Sonne’ cruise 106/2,