INTRODUCTION

It is well known that there is a close relationshipbetween physical and biological processes in theocean. However, despite this knowledge, it is sur-prising that physical oceanography still plays a verymodest part in some marine biological milieu. Wetherefore believe that there is a need to stimulate tomore cooperation between physicists and biologiststo prevent future marine scientific programs whichsuffer due to lack of cooperation.

Lack of integration of physical oceanography inbiological investigations often has economical rea-sons. Ecological field experiments are time con-suming and costly and biologists are tempted tolimit the physical oceanographic part to a mini-mum and make “a best guess” based on e.g. a fewCTD-stations. Such data, however, tells usually

next to nothing about the physical processes whichare necessary to interpret the biological data in asatisfactory way.

It is usually decisive that the physical data aresampled simultaneously with the biological data.This is especially important in areas with strong gra-dients where the processes take place on small timeand space scales, e.g. in fronts, current shear,upwelling.

The possibility of cooperation in field experi-ments has been facilitated in recent years. Work isin continous progress on the development of instru-ments which have improved the measuring tech-nique and increased the efficiency of data sampling,which has therefore become considerably less timeconsuming than just a few years ago. Some of thebiological data may today be collected with a cov-erage and resolution comparable with the best phys-ical data. Also the development of ecological mod-els (coupled hydrodynamical and biological mod-els) have been an inspiration for cooperations.

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 93

SCI. MAR., 61 (Supl. 1): 93-108 SCIENTIA MARINA 1997

LECTURES ON PLANKTON AND TURBULENCE, C. MARRASÉ, E. SAIZ and J.M. REDONDO (eds.)

Physical oceanography and marineecosystems: some illustrative examples*

HARALD SVENDSEN

Geophysical Institute, University of Bergen, 5007 Bergen, Norway.

SUMMARY: The role of physical oceanography on different length and time scales in marine ecosystems is elucidated bymeans of some illustrative examples. Special focus is put on time and length scales, physical mechanisms underlying pro-duction and distribution of biological matter, contact rate and small-scale turbulence, vertical migration and larval settle-ment in tidal flow and the influence of advection on the distribution of plankton and fish eggs.

Key words: physical mechanisms, length and time scales, advection, turbulence, tidal flow, contact rate, migration, settle-ment.

*Received December 12, 1995. Accepted May 25, 1996.

LENGTH AND TIME SCALES

The nature of the relationship between physicaland biological processes is very complex. Bothrely on solar energy, biologically through photo-synthesis and physically, direct through heating ofthe water and indirectly as momentum transmittedto the ocean surface from the wind. An appropri-ate way to approach the subject is to get a feelingfor the length and time scales of the organisms andphenomena to be discussed. Fig. 1 shows somecharacteristic size range of organisms and physi-cal length scales and time scales for some physi-cal and biological features. Somewhat simplified,it seems that while physical features determine thespatial scales of ecological processes, the organ-isms determine the time scales: e.g. from the baro-clinic Rossby radius of deformation which is thetypical width of coastal upwelling areas and oceancurrents to the biologically important mixed layerdepth (≈ 102 m) to the viscous length (Kolmogo-roff length) in the lower end of the scale (≈ 10–2

m) where viscosity starts to smooth out turbulentfluctuations. The time scale may range from 100years, the life span of a mammal, to a few yearsfor fish and further to hours, the doubling time ofbacterias.

SOME PHYSICAL MECHANISMS UNDERLYING PRODUCTION AND DISTRIBUTION OF BIOLOGICAL MATTER

It may be appropriate to start with a description ofsome important mechanisms which govern produc-tion and transport of biological matter in the ocean.The main mechanisms in this respect are advection,turbulent diffusion, wind mixing, the tide and tidalfronts, Ekman pumping, upwelling and convectionprocesses due to evaporation or surface cooling andfreezing. The significance of the different mecha-nismn varies both in time and geographically.

Turbulence

Shear-generated turbulence

As a rule of thumb production of turbulencetakes place in a vertical current shear when the gra-dient Richardson number is less than 1/4 (Miles,1961):

(Richardson number)Ri =

g

ρ∂ρ∂z

∂u

∂z⎛⎝

⎞⎠

2

94 H. SVENDSEN

10-6 10-5 10-4 10-3 10-2 100 102 103 106 (m)(1000 km)(1 km) Earth(100 m)(1 m)(1 cm)(1 mm)(1 µm)

SECOND

MINUTE

HOUR

DAY

WEEK

MONTH

YEAR

MammalFish

Mixed�layer

Zooplankton

Phytoplankton

Bacteria

Swells

Seiches�in fjords

Internal�grav. waves

Inter. osc.�Internal tide

Coastal�upwelling

Barotropic�tide

Geostrophic�turbulence�

Rossby waves

General�circulation

Capillar�wave�

Micro turbulence

HUNDRED�YEAR

FIG. 1. – Characteristic size range of organisms and physical length and time scales.

where ρ = density, g = acceleration of gravity, u =horizontal velocity and z = vertical coordinate.

Furthermore,

represents the stabilizing force and

(shear stress) represents the destabilizing force (seeFig. 2). Even for weak current shear the conditionsfor turbulence production may be present when thestratification is weak. The vertical shear is usuallystrongest in the surface and bottom layers which arethus the main areas for shear generated turbulence inthe ocean.

Production of turbulence in a stratified fluid isnot a continuous process but usually has an inter-mittent character. The turbulence appear as “bursts”in depth levels with very low Richardson number. Inthe depth levels where the “bursts” appear the wateris strongly mixed and the stratification may changeconsiderably. After some time new conditionsfavourable for “burst” generation are built up againand turbulence is generated.

An example illustrating the effect of the intermit-tency character of turbulence on the stratification isshown in Fig. 3. The vertical profiles of density andvelocity are sampled once every hour at a station

∂u

∂z⎛⎝

⎞⎠

2g

ρ∂ρ∂z

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 95

FIG. 2. – Shear generated turbulence in a stratified fluid.

FIG. 3. – Velocity (––––) and density (------) profiles from Station H in Fig. 18a. The numbers are bulk Richardson numbers.

(H1) about 3 km from the outlet of a power plant sit-uated at the head of a narrow fjord (Hylsfjorden) onthe west coast of Norway (see Fig. 18a). The densi-ty is calculated based on CTD data sampled from aship, anchored near a buoy rig at Station H1 carry-ing seven SD-2000 current meters in 1, 2, 3, 4, 5, 6,and 7 m depth respectively. The width of the fjord atthe station site is about 1 km. The discharge from thepower plant was 200 m3 s–1. There was no wind orprecipitation during the experiment and the tide wasvery weak in the fjord. It is seen in the figure that thechanges in the stratification was much more pro-nounced in the depth levels with very low values ofthe Richardson numbers (the numbers are bulkRichardson number).

Breaking of internal waves

A well known source of turbulence is related tointernal wave activity. Turbulent energy is transferredfrom the waves when the waves become unstable andbreak either in the interior of the flow or along slop-ing boundaries (Fig. 4). A condition for generation ofinternal waves is that the water-masses are stratified.Consequently, the contribution of turbulent energyfrom this source varies with time, especially at high-er latitudes. A typical example is the Barents Sea. Insummer and autumn, when the ocean is stratified,internal wave activity may contribute to the turbulentenergy level in parts of the Barents Sea, while in win-ter when the watermasses are almost homogeneousthis source is absent.

Wind generated turbulence

Wind energy supplied to the surface is trans-ferred to turbulent energy by two main mechanisms:shear instabilities as described above and breaking

of surface waves. The latter usually injects turbulentenergy at a much smaller scale than the former (e.g.Phillips, 1966). There is however an exception.When Langmuir circulations appear, which are dri-ven by breaking waves, mixing takes place on muchlarger scales than individual wave-breaking events.

Mixing caused by tides

Mixing related to tidal action is especially impor-tant in shelf waters and has a marked periodic char-acter related to the tidal cycle. Mixing due to tidalcurrents show small variations between each tidalcycle in shelf areas where the difference betweenspring and neap tides is small. However, some areashave a very pronounced spring-neap tidal cycle ase.g. the south Australian Shelf where there is almostno mixing at neap tide while at spring tide the watermasses are exposed to strong mixing (Provis andLennon, 1983). Simpson and Hunter (1974) sug-gested that the marked frontal structure occuring onthe shallow European continental shelf during thesummer months is produced by variations in thelevel of tidal mixing (Fig. 5). The location of thefronts is essentially determined by the parameter

96 H. SVENDSEN

FIG. 4. – Breaking internal waves a) in the interior of the flow (left) and b) along a sloping bottom (right).

FIG. 5. – Front caused by tidal mixing. Q is heat flux.

h/u3, where h is the water depth and u is the ampli-tude of the tidal current.

Mixing in geostrophic fronts

Strong mixing may take place in geostrophicfronts e.g. in the Polar front between Arctic water andAtlantic water. The shape of the front and cross-frontal mixing is affected by variable wind stress.Frontal waves may be trigged by variable wind stressand governed by baroclinicity and bottom topogra-phy. Frontal wave instability and eddy shedding (Fig.6), is the origin of the often numerous pockets ofwater (eddies) which appear in some frontal areas. Insome of these eddies considerable upwelling takesplace.

Vertical advection

Ekman suction

Wind stress on the sea surface varies from placeto place. This leads to convergence of water mass insome places and divergence in other places. Thehorizontally divergent Ekman drift across the sides

of a given area must be replaced by water which is“sucked” vertically up into the surface layer (Fig. 7).The suction may take place in connection with slow-ly moving or quasi-stationary cyclons over an oceanarea.

Applying the Ekman mass transport equations(see Gill, 1982):

(Τx = ρCDWx2, Τy = ρCDWx

2; Wx and Wy are windvelocity components) and the integrated equation ofcontinuity:

and assuming stationary conditions, the Ekman vol-ume transport is:

Then for a stationary or slowly moving low over theocean the Ekman suction velocity, WE, is:

The two terms in the bracket represent the curl ofwind stress. Applying monthly mean wind stress datafor November in the Barents Sea (Martinsen et al.,1992), we got a vertical velocity 0.01 - 0.03 cm s–1,or water “sucked” vertically about 10-30 m day–1.

Since the phenomenon described is closely relat-ed to slowly moving wind fields such events arerather episodic.

Coastal upwelling

A wind blowing along a coast in the northernhemisphere with the coast to the left of the winddirection causes a transport of upper layer water outfrom the coast, so called Ekman drift (e.g. Gill, 1982;Cushman-Roisin, 1994 and Fig. 8). To compensate

WE =1ρf

∂Τy

∂x−∂Τx

∂y

⎛

⎝⎜

⎞

⎠⎟

VE = −1ρf

ΤxUE =1ρf

Τy

⇒∂UE

∂x+∂VE

∂y− WE = 0

∂u∂x

+∂v∂y

+∂w∂z

⎛

⎝⎜

⎞

⎠⎟

z

0

∫ dz = 0

ρ∂VE

∂t+ fUE

⎛⎝

⎞⎠

= Τyρ∂UE

∂t− fVE

⎛⎝

⎞⎠

= Τx

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 97

FIG. 6. – Mixing in fronts and formation of frontal waves.

FIG. 7. – Ekman suction. WE: Ekman suction velocity.

for this water transport, upwelling of usually nutrientrich water takes place along the coast. The interfacebetween the waters upwelled along coast and the off-shore waters constitutes an upwelling front (Fig. 8).The amount of water (the depth of the water column)which is affected by the upwelling and the positionof the front depends on the stratification and thestrength of the wind. Readers who are not familiarwhith this general subject are referred to the volumeedited by Richards (1981).

Upwelling in fronts

A frequent phenomenon in the ocean are theapperance of quasi-geostrophic fronts where a pres-sure gradient perpendicular to the front drives a cur-rent parallel to the front. Due to the friction in thewater the pressure gradient and the Coriolis forceare not exactly balanced (quasi-geostrophic) causingdivergence to the left of the center of the front andconvergence to the right with a resulting verticalflow between. For details about this process, seeGarret and Loder (1981).

Upwelling along an ice edge

Along an ice edge upwelling may also take placecaused by wind stress which exerts different stress-es on ice and open water which cause the moving iceto exert stress on the ocean beneath. If the seawateris homogeneous or weakly stratified the verticaltransport may affect a considerable part of the upperwater masses and transport water vertically 10-20 mday–1. For more details the reader is referred toHäkkinen (1990).

Thermal convection and convection due to sea iceformation

Convection caused by cooling of the surface andsea ice formation takes place in high latitudes, in theArctic and Antartic. Due to intense surface coolingconvective mixing may reach great depths duringthe winter. When sea ice is formed, salt brine drainsout and increases the density of the underlyingwater. This in turn triggers convection which mayreach great depths.

Deep water formation may also be caused byother mechanisms as in the Eastern Mediterranean.Evaporation due to heating of the surface increasesthe salinity of the surface layer water thus precondi-tioning the formation of Mediterranean deep waterdue to winter-time heat loss.

SOME EXAMPLES OF RELATIONSHIPS BETWEEN PHYSICAL AND BIOLOGICALPROCESSES IN THE OCEAN

It is not possible in a short article to cover allaspects related to physical-biological processes inthe ocean. I will therefore limit the presentation to afew subjects where it is obvious that thoroughknowledge of the physical processes is crucial tounderstanding the biological processes. (Two of thesubjects, 1 and 3, are closely related to the topic ofthis volume).

The subjects are the following:1. Plankton contact rate and small-scale turbu-

lence.2. Vertical migration in tidal flow. 3. Larval settlement in turbulent tidal flow.4. The influence of advection on distribution of

plankton, cod eggs and larvae.5. Impact of man-made environmental changes

on ecosystems.The presentation within each subject is mostly

based on published papers in the last decenium. Dueto lack of reliable field experiments, most of sub-jects 1 and 3 are based on models, both numericaland analytical.

Plankton contact rate and small-scale turbulence

The first study of the relation between small-scale turbulence and plankton contact rates was ananalytical study by Rothshild and Osborn (1988).

98 H. SVENDSEN

Fig. 8.– Coastal upwelling

Before that time focus was put on some functionalstudies, in laboratory experiments without turbu-lence, of the relative density of predator-and-preyplankton. Rothshild and Osborn (1988) estimatedthe components of predator-and-prey contact thatare due to small-scale turbulence. The main resultsare summarized in Fig. 9, where v is the velocity ofthe predator, u the velocity of the prey and w theroot-mean-square turbulent velocity. As is obviousfrom the results in the figure, turbulence enhancesthe contact rates.

The results above are broadly confirmed byYamazaki et al. (1991), who carried out directnumerical simulation of planktonic contact in turbu-lent flow. Yamazaki et al. (1991) based their study onthe fact that the only way to generate a dynamicallycorrect turbulent field on the scale of planktonicorganisms is to simulate the full set of the Navier-Stokes equations directly. The turbulence is generat-ed by a pseduo-spectral method of Rogallo (1981)

modified by Squires and Eaton (1990) and the sec-ond-order Runge-Kutta method is applied to traceparticles in a turbulent flow and simulate planktonicswimming with a random walk. Fig. 10a shows tra-jectories of three particles in a realistic oceanic tur-bulence field without random walk (passive parti-cles) and Fig. 10b-d show the same with increasingrandom walk (active particles). The random walk ofthe active particles has to be intensified considerablybefore the effect of the turbulence on movementbecomes negligible. As is obvious from the figure,knowledge about the turbulent field is crucial to sep-arate the random walk from the combined effect ofrandom walk and turbulence.

A question which has attracted increasinglyattention is illustrated in Fig. 11. Is there a correla-tion between turbulence intensity and feeding rate ofcopepods?

The first field study of the influence of turbu-lence on plankton encounter rates was carried out by

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 99

FIG. 9. – Contours showing the increase in the contact rate as a function of prey velocity (u) and preda-tor velocity (v) for an uncorrelated rms turbulent velocity (w). (Redrawn from Rothschild and

Osborn, 1988).

FIG. 10. – Three-dimensional trajectories of three particles; a) without random walk and c-d) with increasing random walk. (Redrawn fromYamazaki et al., 1991).

Sundby and Fossum (1990). Based on the resultsfrom this study and others [MacKenzie and Kiørboe,1993 (laboratory experiment) and MacKenzie andLeggett, 1993 (model)], Sundby et al. (1994)answered one of the questions which was raised pre-viously in Sundby and Fossum (1990): Is there anoptimum level of wind-induced turbulence abovewhich capture of food declines because the relativemotion between predator and prey becomes so largethat the predator gets too short a time to attack? Intheir study of encounter rates between first-feedingcod larvae and their prey during moderate to strongwind generated mixing, Sundby et al. (1994) foundthat the feeding rate increases by a factor of about 7as wind increases from 2 to 10 m s-1. It was con-cluded that a possible optimal level must be found athigher wind speeds than 10 m s-1. The subject is dis-cussed elsewhere in this volume.

Vertical migration and settlement in tidal flow

Vertical migration occurs at a variety of periodsincluding tidal periods. There are several reasonsfor planktonic migration: predation avoidance,bioenergetic advantages and when vertical shear ispresent, vertical migration may be used to re-estab-lish horizontal position. Vertical migration is usu-ally combined with forced movements related tothe flow field: e.g. vertical migration in a sheared,oscillatory tidal current can bring about long-termtransport of larvae known as selective tidal streamtransport.

A detailed study of migration in such an ocillato-ry tidal current demand a comprehensive fieldexperiment to determine both the migration of theorganisms and the flow field. Up to now, as far as isknown by the author, only a few experiments, all ofmodest extent, have been carried out. Some theoret-ical studies, however, have been carried out in recentyears.

Horizontal motion induced by vertical migrationin oscillatory currents were first studied by Hill(1991). His work is an excellent demonstration of acombined effect of a physical and a biologicalprocess in the ocean.

Assuming the horizontal velocity U(z,t) of theflow (see Fig. 12):

and the height of the particle, zp(t), above the sea bedat any time t:

zp(t) = D + asin(ωmt)

where Uo = surface velocity, h = total depth, ω =angular frequency of the tidal ocillation, φ =phase angle, ωm = angular frequency of the migra-tion, D = mean height of the organism above thesea bed and a = vertical migration amplitude of anindividual zooplankter. The velocity of the organ-ism is then:

U z, t( ) =Uoz

hsin ωt + φ( )

100 H. SVENDSEN

FIG. 11. – Suggested correlation between “appetite andwell-being” of copepods and turbulence intensity. 1. Strong intensi-ty - stressed copepods; 2. Low intensity - starving copepods; 3.

Medium intensity – well-being copepods.

FIG. 12. – Velocity profile in an oscillatory flow with linear shear. Uo= mean velocity, D = mean pos. of particle, a = vertical migrationamplitude of an individual zooplankter. (Redrawn rom Hill, 1991).

up = U(zp,t) = [D + asin(ωmt)]sin(ωt + φ)

The integration of the equation is straight for-ward and it is shown that the horizontal displace-ment provided that ω ≠ ωm lies within the range (seeFig. 13):

xD cos(φ + ε) – A ≤ xp ≤ xD cos(φ + ε) + A

where xD = mean (time-independent) displacement,A = amplitude of an oscillating time-dependent dis-placement and ε = a phase correction.

As is seen in the Fig. 13a,b there is no net long-term horizontal displacement for either diel (24 h) orsemi-diel (12 h) vertical migration of the particlewhen the period of the tide is semi-diurnal. The dis-

placement is however much larger for the semi-diel(12 h) than for the diel (24 h) period and approach-es infinity as the migration period approaches thesemi-diurnal tidal period (12.42 h) (Fig. 13c). Hill(1991) also showed that for nonlinear forms ofvelocity shear, unidirectional transport is possiblefor multiple tidal migration periods.

In a study of the impact of advective processeson displacement of zooplankton biomass in aNorth Norwegian fjord system, Falkenhaug et al.(1995) showed that tidal displacement may havelarge consequences for the observed distributionof the zooplankton if organisms migrate throughdifferent layers during a 24-hour cycle. The cur-rent measurements, however, show that with a 24-hour migration cycle an organism will experiencecurrents of opposite directions and thus mayremain in the system.

Uo

h

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 101

FIG. 13. – a. Horizontal displacement of a vertically migrating organism; diel migration, semidiurnal tide; b. same as a. but semi-diel migra-tion; c. Maximum horizontal displacement from the initial position x=0. (Redrawn from Hill, 1991).

Larval settlement in turbulent tidal flow

The diverse factors that affect recruitment ofbenthic invertebrates have been devoted attention innumerous papers which have greatly expanded theunderstanding of the diverse physical factors thataffect the recruitment and the extent to which thesefactors may operate. Butman (1986) and others haveshown by simple scaling arguments that most larvaedisperse horizontally by currents essentially in apassive manner with only partial control of their ver-tical position. This means that hydrodynamicprocesses, i.e. vertical advection, turbulent mixingand shear and boundary layer processes, affect theintensities of the settlement.

To quantitively evaluate the relative influencesof these processes upon settlement let us considersome of the results from Gross et al. (1992) whodeveloped a time-dependent model of the tidal bot-tom boundary layer. They based their model on:- the concentration equation:

- the momentum equations of u and v:

and

- the k-ε (TKE-closure) turbulent kinetic energyequation:

where C = concentration of larvae, Av = eddy diffu-

sivity, Ae = turbulent energy diffusivity, wf = larval

fall velocity, both u,v = velocity components, P=pressure, ρ = density, and f = Coriolis parameter.

The solutions obtained with the TKE closure areshown in Fig. 14d. The suspended particle load sim-ulated with the TKE closure is expected to be closeto reality. However, in most existing models muchsimpler closures (eddy diffusitives) are used.

To investigate how many solutions with simplerparameterization of the eddy diffusive parameterdiffer from the TKE closure, Gross et al. (1992) ranthe model with three different eddy diffusitivesintroduced by Soulsby (1990).

Their model run with the three different eddy dif-fusivities are shown in Fig. 14a-c. The constant dif-fusivity gave poor results near the bed compared tosimulation with TKE closure, but adequate in theouter regime while it was opposite for the linear dif-fusity. The best results were obtained with the mixedmethod while the constant coefficient gave verypoor results as is seen in Fig. 14.

Constant Av(z,t) = CdUd(t)

Linear Av(z,t) = κz[CdUd(t)]

1/2

Mixed Av(z,t) = κz[C

dU

d(t)]1/2 z < Zδbl

Av(z,t) = κZδbl[CdUd(t)]1/2 z > Zδbl

(Zδbl is the boundary layer thickness).There are a large number of parameters required

to specify a sinusoidally time dependent boundary

∂ke

∂t= Av

∂u∂z

⎡⎣⎢

⎤⎦⎥

2

+∂v∂z

⎡⎣⎢

⎤⎦⎥

2⎧⎨⎪

⎩⎪

⎫⎬⎪

⎭⎪+

∂∂z

Ae

∂ke

∂z⎡⎣⎢

⎤⎦⎥− ε

∂v∂t

+ fu =1ρ∂P∂y

+∂∂z

Av

∂v∂z

⎡⎣⎢

⎤⎦⎥

∂u∂t

− fv =1ρ∂P∂x

+∂∂z

Av

∂u∂z

⎡⎣⎢

⎤⎦⎥

∂C∂t

+∂ w f C( )

∂z+

∂∂z

Av

∂C∂z

⎡⎣⎢

⎤⎦⎥

102 H. SVENDSEN

FIG. 14. – Eddy diffusivity (Av) and concentration profiles (zC) over time during a tidal oscillation for four eddy diffusivities: a. cons-tant; b. linear; c. mixed constant and linear and d. TKE closure (from Gross et al., 1992).

layer, larval suspension and settlement. An appro-priate method to examine these parameters is bydimensional analysis to reveal fundamental nondi-mensional parameters to examine the behavior ofthe model. [For details see Gross et al. (1992)]. Themomentum field can be fully nondimensionalizedby specification of the following four variables:

u*max Maximum shear stress at the bed.

T Period of ocillation.h Total depth.zo Roughness length scale of sea bed.

while concentration field and bottom boundary con-dition are specified with:

Ctotal Total initial number of larvae in sus-

pension per area.w

f Fall velocity of particles.ws Settlement velocity (velocity out of

the near bed region to the bed).

Many different field conditions can now beexamined by using these variables to form the fol-lowing parameters:

, , and

Relative roughness h/zo:

Kinetic energy closure requires a large Reynolds’number in a flow layer (h) over a hydrodynamicalrough bed layer with constant (z

o);

[Re/Reo = u*h/ν/u*zoν = h/zo]. h/zo is usually > 104.

Boundary layer depth/Full depth δbl

/h:This relation gives an indication of the height

above the bed to which the turbulence generatednear the bed will have an effect.

The height (δbl) of an oscillating turbulent bound-ary layer depends on the turbulent strength and theperiod of oscillation:

where κ is a constant.

(κu*max: turbulence strength).

If < 1 then the depth of the boundary layer is

less than the water column.

If > 1, turbulence is able to suspend material

all the way to the surface.

Rouse Number wf /κu*max:

This number gives a measure of the relativestrength of downward flux due to fall velocity of anorganism versus the upward turbulent suspensionflux. A large Rouse number indicates that the larvaeremain close to the bed.

Settlement velocity/fall velocity ws /wf:

If the three mechanisms of larval movement: theadvective fall flux w

fC, the diffusive flux Aν ∂C/∂zand the bed settlement flux w

sC are out of balance

δbl

h

δbl

h

δbl =κu*maxT

2π

ws

wf

u*maxT

h

w f

u*max

h

zo

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 103

FIG. 15. – Proportion of larvae (Czc) below zc (the distance in which a larvae is close enough to the bed tointeract with the bed) as a function of phase of tide. There is a strong dependence of the near bottom

concentration profile on wf/κu

*maxand a weak dependence on δ

bl/h (from Gross et. al, 1992).

then the ratio of settlement flux (ws C) to eitheradvective flux (w

f C) or diffusive flux (Aν ∂C/∂z)

becomes important. Either ws

/u*max

or ws

/wf

maydescribe this ratio.

Larval settlement is also controlled hydrody-namically through the supply of larvae to thedepth zc where settlement may take place within“quiet” intervals between intense bursts of turbu-lence. Fig. 15 shows the dependence of the rela-tive pool (relative to the total amount of larvae) ofpotential settlers Czc for different values of turbu-lence levels and the relation fall velocities versusupward flux. Obviously there is a strong depen-dence of the near bottom concentration profile onw

f /κu*max and a weak dependence on the boundarylayer depth (δ

bl /h).As is obvious from the discussion above, knowl-

edge about the bottom shear stress is crucial to be ableto estimate bottom settlement of larvae. The develop-ment of high-resolution current meters (acousticdoppler current meters) which has taken place inrecent years has made it possible to measure bottomstress in the ocean and by such one is able to estimatethe diffusive flux based on measurements.

The influence of advection on distribution offish eggs and larvae

Another area which has been devoted attentionfor many decades is the link between horizontaladvection and distribution of plankton, fish eggs andlarvae. This has been studied in many ocean areasaround the world. It is shown that there is a close

relationship between the distribution of eggs andearly stages of fish and advection processes. Toillustrate this relation some results from one of themost intensively investigated areas, Georges Bankin the Gulf of Maine (Fig. 16), are presented. Phys-ical and biological field investigations have beencarried out for many decades in this area. Theresults, from numerous investigations carried outsince the beginning of the twenties, may be summedup as follows (see Werner et al., 1993):

Biology:

– Haddock spawning takes place on Georges Bank,mostly over the Northeast Peak, in the period Feb-ruary-June with peak spawning in May-April(temperature dependent).

– Pelagic juveniles are centred over western GeorgesBank and extending towards Great South Channel.

– Cod spawn throughout most of the year, peak fromJanuary to April, and are present on Georges Bankfrom November through June. Highest concentra-tions over the southern flank.

– Vertical distribution of cod and haddock eggs variesin space and time due to processes like buoyancy ofthe eggs and larvae, stratification and turbulence.

– Swimming capability increases as the larvaedevelop into pelagic juveniles.

Hydrodynamics:

– At any time the currents on Georges Bank may bedivided into four main components: seasonally

104 H. SVENDSEN

FIG. 16. – Map of Georges Bank.

varying mean current, low-frequency current fluctu-ations (wind events and passing eddies), tidal cur-rents (M2) and current fluctuations related to e.g.internal waves.

– Both seasonal variations in the mean local wind-stress, the large-scale coastal current, the persis-tent contribution from tidal rectification andfrom the density field, contribute to the seasonalintensification of the mean current pattern (gyre)on Georges Bank.

– The gyre strength on Georges Bank and the recir-culation in Great South Channel are near theirseasonal minima in March-April.

– There is weak stratification over the Bank in thewinter except in the shelf-slope water.Based on the information listed above, one is

able to form a picture of which physical processesdetermine the observed distribution of eggs and lar-vae on the bank (Fig. 17a). However, to get knowl-edge about the relative importance of the tide, meancirculation and wind-driven circulation on the distri-bution on eggs and early life stages, a comprehen-sive and intensive field experiment is necessary. An

alternative, and less costly method is to use 3-Dmodels.

Werner et al. (1993) did a climatological study ofthe influence of mean advection and behavior on thedistribution of cod and haddock early life stages onGeorges Bank. In the study, using a harmonic finiteelement method, a 3-D non-linear shallow-waterequations model with eddy viscosity closure in thevertical was applied. The forcing was the tides,windstress on the surface and a prescribed densityfield. For details see Lynch et al. (1992) and Lynchand Naimie (1993). Both passive particles (eggs)and active particles (larvae and juveniles, vertical aswell as horizontal movements) were “released” inthe model.

The simulations confirmed the influence of tidalcurrents, residual tidal currents, wind-driven cur-rents and inflow from Scotian Shelf on the distribu-tion of cod and haddock early life stages on GeorgeBank (Fig. 17b). The observed distribution of codand modelled combination of physical and biologi-cal processes gave convincing results as theobserved and modelled distributions showed quali-

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 105

FIG. 17. – Observed distribution of Haddock (a) and Cod (b) eggs on Georges Bank (gridded area) and larvae 20-40 days postspawn (dotted area) and larvae 50-70 days post spawn (hatched lines). Simulated distribution of passive particles (eggs) 31days (c) and 62 days (d) after release in 10 layers between 5 and 50 m in the north-eastern part of Georges Bank. (Redrawn

from Werner et al., 1993).

tative agreement. It was shown that an importantpart of the biological variations after spawning weredetermined by physical processes.

Also when looking at variability in eggs and lar-val distributions on the Bank related to high-fre-quency variations of the wind-field the observed(two seasons) and simulated distributions showedgood agreement (Lough et al., 1994).

Impact of man-made environmental changes on ecosystems

Human activity may alter the physical environ-mental conditions for the living organisms in theocean e.g. release of CO2 and CFC-gasses to theatmosphere, pollution from industrial and agricul-

tural areas supplied by major rivers and pollutionwhich comes from discharges of cities and indus-trial complexes along the coast as well as shippingand hydroelectric power production which altersrunoff patterns to marine ecosystems, and thusphysical properties of the upper layer. This in turnaffects numerous biological properties (e.g. Bug-den et al., 1982; Kaartvedt and Svendsen, 1990a,1990b, 1995; Kaartvedt et al., 1991).

To evaluate the impact of freshwater dischargeon a fjord system in western Norway (Fig. 18), acomprehensive physical-biological field experimentwas carried out. The study may be considered as alarge “laboratory experiment” since the experimentswere carried out before and after an agreed plan forrunning the powerplant during the experiments.

106 H. SVENDSEN

FIG. 18. – a. Map of three of the Ryfylkefjords; b. Mean current (25-h sliding means) from four measuring depths at sta-tion H. (pos. values are up-fjord). The arrows indicate commencement and cessation of discharge. (From Kaartvedt and

Svendsen, 1990a).

FIG. 19. – Chlorophyll a profiles from Hylsfjorden prior to (23 October) (open circles), and during discharge (27October) (full circles). a. Station H1, b. Station H2, c. Station H3 and d. Station H4. (From Kaartvedt and

Svendsen, 1990a).

During running of the powerplant the circulationpattern of the upper layer was totally changed (Fig.18b). Most of the variability of the biomass distrib-ution during all the experiments were explained byphysical processes (Kaartvedt and Svendsen, 1990a;Kaartvedt and Svendsen, 1990b; Kaartvedt et al.,1991). An illustrative example is shown in Fig. 19.A comparison of the biomass distribution prior toand during discharge of freshwater from the power-plant showed large changes caused by the changesof the circulation pattern. As is obvious it was ofvital importance for the interpretation of the biolog-ical data that the physical and biological data weresampled simultaneously since the change wasdirectly related to the advection field.

CONCLUDING REMARKS

In this presentation only a few examples havebeen given of the many of biological processes in theocean which mainly are governed by physicalprocesses. It should be needless to mention that pro-jects designed to study the biological processes in theexamples shown must have physical investigation asa strongly integrated part of the projects. It is there-fore an advantage that the marine biology communi-ty take the consequences of what we know todayabout the close relationship between biology andphysical processes in the planning of the projects.

REFERENCES

Bugden, G.L., B.T. Hargrave, M.M. Sinclair, C.C. Tang, J.C. Ther-riault and P.A. Yeats. – 1982. Freshwater runoff in the marineenvironment: the Gulf of St. Lawrence example. Can. Tech.Rep. Fish. Aquat. Sci., : 1-89.

Butman, C.A. –1986. Larval settlement of soft-sediment inverte-brates. some predictions based on an analysis of near-bottomvelocity profiles. In: J.C.J. Nihoul, (ed.), Marine InterfacesEcohydrodynamics, Oceanography Series, 42, pp. 487-513Elsevier.

Cushman-Roisin, B. – 1994. Introduction to Geophysical FluidDynamic. Prentice-Hall Inc. Englewood Cliffs.

Falkenhaug, T., E. Norby, H. Svendsen and K.Tande. – 1995.Impact of advective processes on displacement of zooplanktonbiomass in a North Norwegian fjord system. Seasonal and ver-tical variations. In: C. Hopkins, K.E. Erikstod and H.P. Leinaas,Ecology of Fjords and Coastal Waters. Elsevier Science B. V.

Garret, C.J.R. and J.W. Loder. – 1981. Dynamical aspects of shal-low-sea fronts. Phil. Trans. Roy. Soc. Lond. A 302: 562-581.

Gill, A.E. – 1982. Atmosphere-Ocean Dynamics. InternationalGeophysics Series 30. Academic Press, New York.

Gross, T.F., F.E. Werner and J.E. Eckman. – 1992. Numerical mod-eling of larval settlement in turbulent bottom boundary layers.J. Mar Res., 50: 611-642.

Häkkinen, S. – 1990. Models and their applications to polaroceanography. In: W.O. Smith, Jr., (ed.), Polar Oceanogra-phy. Part A: Physical Science , pp 335-384, Academic Press,Orlando.

Hill, A.E. – 1991. Vertical migration in tidal currents. Mar. Ecol.Prog. Ser., 75: 39-54.

Kaartvedt, S., T.M. Johnsen, D.L. Aksnes, U. Lie and H. Svendsen.– 1991. Occurrence of the toxic phytoflagellate Prymnesiumparvum and associated fish mortality in a Norwegian fjord sys-tem. Can. J. Fish. Res., 48 (12).

Kaartvedt, S. and H. Svendsen. – 1990a. Impact of FreskwaterRunoff on Physical Oceanography and Plankton Distribution iWestern Norwegian Fjord: An Experiment with a ControlledDischarge from a Hydro-Electric Power Plant. Est. Coast. Mar.Sci., 31: 381-395.

Kaartvedt, S. and H. Svendsen. – 1990b. Advection of euphausiidsin two Fjord Systems Subject to Altered Fresk-Water input byHydro-Electric Power Production. J. Plankton Res.,12(6):1263-1277.

Kaartvedt, S. and H. Svendsen. – 1995. Effect of freshwater dis-charge, intrusion of coastal water, and bathymetry on zoo-plankton distribution in a Norwegian fjord system. J. PlanktonRes., 17(3): 493-511.

Lough, R.G., W.G. Smith, F.E. Werner, J.W. Loder, F.H. Page,C.G. Hanna, C.E. Naimie, R.I. Perry, M.M. Sinclair and D.R.Lynch. – 1994. The influence of wind–driven advection on theinterannual variability in cod egg and larval distributions onGeorges Bank: 1982 vs 1985. ICES Mar. Sci. Symposiums. (inpress).

Lynch, D.R. and C.E. Naimie. – 1993. The M2 tide and its residualon the outer banks of the Gulf of Maine. J. Phys. Oceanogr. 23:222-2253.

Lynch, D.R., F.E .Werner, D.A. Greenberg and J.W. Loder. – 1992.Diagnostic model for baroclinic, wind-driven and tidal circula-tion in shallow seas. Cont. Shelf Res. 12: 37-64.

MacKenzie, B.R. and T. Kiørboe. – 1993. Feeding and swimmingbehaviour of larval cod and herring in calm and turbulent envi-ronments. ICES Symposium on Cod and Climate, Reykjavik 23-27 August 1993. ICES CCC No. 24. 18.

MacKenzie, B.R. and W.C. Leggett. – 1993. Wind-based modelsfor estimating the dissipation rates of turbulent energy in aqat-ic environments: empirical comparisons. Mar. Ecol. Prog. Ser.,94: 207-216.

Martinsen E.A., H. Engedahl, G. Ottersen, B. Ådlandsvik, H.Loeng, and Balino, B. – 1992. Climatological and Hydrogra-phical Data for Hindcast of Ocean Currents. The NorwegianMeteorological Institute. Techn. report No. 100.

Miles, J.W. – 1961. On the stability of heterogeneous shear flows.J. Fluid Mech., 10: 496-508.

Phillips, O.M. – 1966. Dynamics of the upper ocean. CambridgeUniv. Press, London.

Provis, D.G. and G.W. Lennon. – 1983. Eddy viscosity and tidalcycles in a shallow sea. Est. Coast. Shelf Sci., 16: 351-361.

Richards, F. A., ed, – 1981. Coastal Upwelling. Coastal and Estu-arine Sciences, Vol. 1, American Geophysical Union, Wash-ington, D.C..

Rogallo, R.S. – 1981. Numerical experiments in homogeneous tur-bulence. NASA Tech. Memo, 81315.

Rothshild, B.J. and T.R. Osborn. – 1988. Small–scale turbulenceand plankton contact rates. J. Plankton Res., 10: 465-474.

Simpson, J.H. and J.R. Hunter. – 1974. Fronts in the Irish Sea.Nature, 250: 404-406.

Soulsby, R.L. – 1990. Tidal-current boundary layers. In: A.B.LeMehaute and D.M. Hades (eds.),The Sea, 9, part, pp.523–566, J. Wiley and Sons Inc., New York.

Squires, K.D. and E.K. Eaton. – 1990. Particle response and turbu-lence modification in isotropic turbulence. Phys. Fluids, 172:1191-1203.

Sundby, S. and P. Fossum. – 1990. Feeding conditions of Arcto–Norwegian cod larvae compared to the Rothshild-Osborn theo-ry on small-scale turbulence and plankton contact rates. J.Plankton Res., 12: 1153-1162.

Sundby, S., B. Ellertsen and P. Fossum. – 1994. Encounter ratesbetween first–feeding cod larvae and their prey during moder-ate to strong turbulent mixing. ICES mar. Sci. Symp. 198: 393-405.

Svendsen, H., L.C. Asplin, O.K. Leth and K. Finne. – 1995. Phys-ical oceanography of coupled fjord-coast systems in NorthNorway with special focus on frontal dynamics and mixingprocesses. Report. Geophysical Institute, Univ. of Bergen.

Werner, F.E., F.H. Page, D.R. Lynch, J.W. Loder, R.G. Lough, R.I.Perry, D.A. Greenberg and M.M. Sinclair. – 1993. Influences of

PHYSICAL OCEANOGRAPHY AND MARINE ECOSYSTEMS 107

mean advection and simple behaviour on the distribution of codand haddock early life stages on Georges Bank. FishOceanogr., 2: 43-64.

Yamazaki, H., T.R. Osborn and K.D. Squires. – 1991. Directnumerical simulation of planktonic contact in turbulent flow. J.Plankton Res.,13: 629-643.

108 H. SVENDSEN