Estuarine, Coastal and Shelf Science 63 (2005) 23–31

www.elsevier.com/locate/ECSS

Winter-Spring flushing of Bass Strait, South-EasternAustralia: a numerical modelling study

Paul A. Sandery), Jochen Kampf

School of Chemistry, Physics and Earth Sciences, Flinders University, GPO Box 2100, Adelaide, SA 5001, Australia

Received 6 June 2004; accepted 22 October 2004

Abstract

This study investigates winter-spring flushing of Bass Strait with a two dimensional non-linear depth-averaged shallow-watermodel. The model is driven with a 180-day wind time-series. An advection-diffusion scheme for several tracers is used to reveal theflushing pattern/timescale of the region. The study considers how external water masses flush strait waters. It shows that shelf-water

entering the strait through the passage between King Island and Cape Grim makes the largest contribution to strait waters. Resultsshow that the central area of the strait is a stagnation-area of weak currents and long flushing times (O160 days). The influence ofexternal water masses on the stagnation-area is estimated. The findings have implications for marine ecosystems, residence times,air-sea modifications of water mass properties and dense water formation in the region.

� 2004 Elsevier Ltd. All rights reserved.

Keywords: Bass Strait; currents; water mass transport; flushing; marine ecosystems

1. Introduction

Bass Strait is an area of shallow continental shelfbetween Victoria and Tasmania connecting the south-east Indian Ocean with the Tasman Sea (Fig. 1). Theregion supports a diverse marine ecosystem with a widerange of habitats. The submerged temperate rocky reefsand canyons contain high species biodiversity witha large proportion being endemic to the area (Neil,2000). Marine activities of environmental significanceinclude fisheries, shipping, oil-drilling/processing andcoastal riverine discharges. All are potential sources ofpollutants and contaminants. In winter and to a largedegree in spring, strait waters are well mixed with littleor no apparent stratification (Baines and Fandry, 1983;Tomczak, 1985; Middleton and Black, 1994). In thepassages strong vertical and horizontal tidal mixing

) Corresponding author.

E-mail address: paul.sandery@flinders.edu.au (P.A. Sandery).

0272-7714/$ - see front matter � 2004 Elsevier Ltd. All rights reserved.

doi:10.1016/j.ecss.2004.10.009

occurs. These areas are always well mixed. The centralregion becomes stratified in summer. The approach ofthe next winter sees the entire strait becoming well mixedagain. Lateral flushing results from inflows of threeprimary water masses (see Fig. 1). These are SouthAustralian Current Water (SACW), East AustralianCurrent Water (EACW) and sub-Antarctic SurfaceWater (SASW) (Newell, 1961). Primary water massrelative contributions have an influence on local marineecosystems owing to their different nutrient contents.During the southern winter SASW is found widelypresent in the strait (Newell, 1961). SASW containshigher nutrient levels (Gibbs et al., 1986). It is thereforeimportant to know how much SASW spreads throughstrait waters.

The flushing times of Bass Strait water are unknown.A zone of long flushing times is a zone where seasonalscale air-sea fluxes influence water mass properties. Sucha zone therefore promotes dense water formation andexport (Tomczak, 1985, 1987).

24 P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

Fig. 1. Bathymetry (depth in m) of Bass Strait and the surrounding region. Dashed lines delimit the initial locations of tracers A, B C and D

representing different water masses. Depth contours are in m.

The density of Bass Strait Water is modified by bothinflows of the primary water masses and local seasonalscale air-sea fluxes (Tomczak, 1985, 1987). Bass StraitWater is advected as far north as the Coral-Sea andseveral hundred kilometers eastward underneath thesurface layer of the Tasman Sea (Luick et al., 1994).Dense water export from the strait occurs at severallocations along the eastern-shelf break with the largestflux of dense Bass Strait Water culminating as the BassStrait Cascade in winter in the vicinity of Bass Canyon(Tomczak, 1985, 1987; Luick et al., 1994). Fieldobservations by Tomczak (1985, 1987) gave first insightinto the seasonal variability of the Bass Strait Cascadeand also provided evidence for circulation in the strait.The response to winds and tides was examined by Jones(1980), who found wind-driven flow was largely alignedwith topographic contours. The annual cycle of thedensity field was investigated by Baines and Fandry(1983), who determined the density field to be controlledmainly by the M2 semi-diurnal tide and local seasonal-scale heat fluxes. Fandry et al. (1985) investigated thestrait’s tides using observations and modelling andshowed the principal harmonic constituents in the straitare the M2, S2, O1 and K1 tides with the M2 tide beingthe dominant mode. In the passages and around thelarge Bass Strait islands, tidal currents were estimated tobe as swift as 3 m s�1.

Elevated concentrations of nutrients have beenobserved at the western and eastern sides of Bass Straitat the shelf breaks. Gibbs et al. (1986) suggested themajor source of nutrients for the strait was upward-mixing of deeper water of sub-Antarctic origin at theshelf-break with mixing occurring across tidal fronts.Elevated local concentrations of nutrients at the shelf-break were found only between autumn and spring but

not in winter. The interior of Bass Strait in winter wasfound to be low in nutrients. Blackman et al. (1987)observed a cold bottom layer temperature anomaly inwestern Bass Strait and suggested upwelling or entrain-ment of deeper water of sub-Antarctic origin onto thewestern shelf was a likely explanation. In a threedimensional modelling study, Evans and Middleton(1998) found that upon relaxation of a constant westerlywind-stress a gyre developed off the western shelf-breakleading to a plume of deeper water upwelled andadvected into the strait.

The aim of this study is to estimate the flushing timesof Bass Strait waters and investigate the mixing ofdifferent water masses within the strait. The next sectioncontains the model description. This is followed bya presentation of results and a discussion.

2. Model description

To investigate the flushing of strait waters, a re-quirement is to establish time dependent tidal andatmospherically forced circulation patterns. In theperiod of winter to spring, this can be achieved witha numerical model using the non-linear depth-averagedshallow-water equations. The processes investigated inthis study are meso-scale phenomena so it is justifiableto ignore the curvature of the Earth’s surface and usea rectangular Cartesian grid. The depth-averagedshallow-water equations are suitable for modellingparticular dynamical processes. They describe baro-tropic motion in a single layer un-stratified ocean.Solutions of these equations are a space-time series ofdepth-averaged current vectors where sea level changesare produced by current convergence or divergence.

25P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

Currents are also influenced by pressure gradients asa result of sea level variations. In rectangular Cartesiancoordinates and in continuous form, these equations arewritten as

vu

vtCu

vu

vxCv

vu

vy� f vZ

� gvh

vxC

ts;x � tb;x

rohCAx

1

h

�v2�hu

�vx2

Cv2�hu

�vy2

�ð1Þ

vv

vtCu

vv

vxCv

vv

vyCf uZ

� gvh

vyC

ts;y � tb;y

rohCAy

1

h

�v2�hv�

vx2C

v2�hv�

vy2

�ð2Þ

vh

vtC

v�hu

�vx

Cv�hv�

vyZ0 ð3Þ

where x and y are rectangular coordinates, u and v aredepth-averaged current velocities in the x and ydirections respectively, h is sea level anomaly, t is timed is an initial undisturbed water depth, h is total waterdepth i.e. hZhCd, f is the Coriolis parameter, g isgravitational acceleration, ro is a reference density forsea water, ts,x and ts,y are the components of surfacewind stress, tb,x and tb,y are the components of bottomstress, and Ax and Ay are lateral eddy viscosity coef-ficients, assumed to be constant. For numerical values ofconstants see Table 1.

The first term on the left side in Eqs. (1) and (2)represents the local rate of change of velocity. Thesecond and third terms are non-linear advection terms.These are important in areas subject to strong currentsand large horizontal velocity gradients such as placesalong coastlines or where there exists large topographicgradients. The fourth terms on the left side in (1) and (2)are the Coriolis accelerations. The first terms on the

Table 1

Parameters and constants used in the model

Quantity Value

Time step 12 s

Grid space step in x and y directions 3710 m

Domain length in north-south ( y) direction 560,210 m

Domain length in east-west (x) direction 797,650 m

Maximum water depth 4000 m

Minimum water depth 25 m

Horizontal eddy viscosity coefficient 1 m2 s�1

Bottom friction coefficient 3!10�4

Gravitational acceleration 9.81 m s�2

Coriolis parameter for 39 �S �0.92!10�4 s�1

Mean sea water density 1026 kg m�3

Mean air density 1.24 kg m�3

Coefficient of drag 1.3!10�3

right side of (1) and (2) are the pressure gradient terms,the second terms are the friction terms. The fourth termson the right side of (1) and (2) are terms for the diffusionof horizontal momentum. Eq. (3) is the continuityequation.

The surface wind stress components ts,x and ts,y thatparameterize tangential friction at the air-sea interfaceare calculated using the quadratic law for surfacefriction

ts;xZracdus

ffiffiffiffiffiffiffiffiffiffiffiffiffiu2sCv2s

q

ts;yZracdvs

ffiffiffiffiffiffiffiffiffiffiffiffiffiu2sCv2s

q

where us and vs are surface wind speed components, cd isa coefficient of drag and ra represents mean air density.

The bottom friction components tb,x and tb,y arefound using the quadratic law for bottom friction thatparameterizes tangential friction at the sea flooraccording to

tb;xZrorbu2

tb;yZrorbv2

where in addition to terms already stated rb is anempirical bottom friction coefficient. Through transferof momentum to the sea floor, bottom friction consti-tutes a loss in energy from the system.

Assumptions regarding flow in Eqs. (1)–(3) are:

� Flow is barotropic and single layered.� Density of the fluid is uniform vertically.� The shallow-water approximation i.e. L[H.� The hydrostatic approximation.� Constant diffusion of horizontal momentum.� The Boussinesq approximation� Atmospheric pressure perturbations at the seasurface are unaccounted for.

In winter months water in most areas of Bass Straitis well mixed and close to being uniformly dense,horizontally and vertically, hence the first two assump-tions. The horizontal length scale (L) of the phenomena(meso-scale currents, winds) is much larger than thedepth scale in the strait (H), which satisfies assumptionthree. Geopotential surfaces are very nearly coincidentwith surfaces of constant height. The equilibriumcondition for this is given by the hydrostatic equation,dp/dzZ�gr where p is pressure, z is depth, g isgravitational acceleration and r is density. This is inassumption four and implied when the shallow-waterapproximation is made. The fifth assumption represents

26 P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

a statistical average of the effects of turbulent diffusionof horizontal momentum. Real eddies have morecomplex internal and boundary structures. The solu-tions obtained from the model will be smoother thanthat which would be observed in the field. In the seasonsfocused on in this study, horizontal density gradients inthe strait are small enough to neglect their effect onhorizontal accelerations in the fluid hence assumptionsix.

Following the advection and diffusion of conservativetracer placed at certain locations reveals the longer termnet transport of the water column. Tracer fluxes can beused to estimate the extent of lateral flushing relative toexternal water masses. The conservation equations fordepth-averaged advection-diffusion of a conservativetracer are

vC

vtCu

vC

vxCv

vC

vy� 1

h

�Kx

v2ðChÞvx2

CKy

v2ðChÞvy2

�Z0 ð4Þ

where in addition to terms already stated, C representsdepth-averaged tracer concentration, and Kx and Ky arehorizontal eddy viscosity coefficients in the x and ydirections respectively, assumed equal.

The first term in (4) represents the local depth-averaged flux rate of tracer concentration, the secondand third terms represent the advection of tracerconcentration in the currents and the remaining termsrepresent turbulent diffusion of tracer concentration.Examples of finite-difference schemes for the depth-averaged momentum equations and the advection-diffusion equations can be found in Kowalik and Murty(1993).

This model is frequently used in coastal regionssubject to strong tidal flows and strong vertical mixing(Tang and Grimshaw, 1999). Limitations of the depth-averaged model include the inability to incorporate air-sea heat and salt/freshwater fluxes and the inability toresolve dispersion due to shear-flow. Within the contextof appropriate time and space scales, the solutionsprovide a way of understanding the flushing of regionalwaters.

An explicit Eulerian forward finite-difference numer-ical scheme is used on an Arakawa C grid (Mesinger andArakawa, 1976). Turbulent horizontal diffusion ofmomentum is parameterized with a constant diffusioncoefficient of 1 m2 s�1. Bathymetric data from ETOPO2(c/- National Geophysical Data Centre), is used ona Cartesian grid with a horizontal resolution of 2nautical miles (w3.71 km) (Fig. 1). The model gridspans 215!150 grid cells. The domain is the extent ofthe area represented in Fig. 1. ETOPO2 is known to beunreliable in shallow waters and in resolving coastlines.This data was processed in two ways to improve itsfunction in the model. Isolated unreasonable depthswere removed and corrections to coastal boundaries

were made manually with reference to AGD84 theAustralian topographic grid datum. Bathymetric datawas smoothed using a numerical diffusion algorithm.Maximum water depth is set to 4000 m. A minimumwater depth of 25 m was chosen based on tidalsensitivity studies. In adjusting the bathymetry for thepurpose of modelling, the presence of the continentalslopes and the bathymetry and coastline of Bass Strait iskept whilst the domain is tailored to minimize numericalinstability.

Open-sea boundary conditions for velocity compo-nents are flow dependent with a zero-gradient conditionfor inward flow and an explicit Orlanski radiationcondition for outward flow (Palma and Matano, 1998).The control volume approach for upstream advection isadopted with a semi-implicit approach for bottomfriction.

Experiments were carried out (not shown) with themodel with wind forcing and open-sea boundary forcingusing M2, S2, O1 and K1 tidal waves following theobservations and predictions by Fandry et al. (1985).Over much of the region, particularly in central BassStrait, residual flow resulting from the four principletidal constituents is up to w80 times smaller than actualtidal currents. The largest tidal residual currents areconfined to a few locations such as in the passagebetween Cape Grim and King Island and Banks Straitwhere the maximum is w20 cm/s. The results in thisstudy for wind and tidal flushing are very similar to theresults of purely wind driven flushing.

The model is forced with an observed 180 day hourly-averaged (derived from minutely data) wind time-series.This data is obtained from Cape Grim Baseline Air-Pollution Station (CGBAPS) located at 40 �40#56$S,144 �41#18$E. The time-series used to force the modelcorresponds to the winter-spring period of 1988.Climatologically averaged winds vary by about 5–10%in strength and direction between Cape Grim, WilsonsPromontory and Low Head during this period. Al-though the wind field used does not exactly represent thespatial distribution of winds over the region it stillprovides a first approximation of currents and flushingduring a winter-spring period.

It is noted that using climatologically averaged windsproduces a similar flushing response at the time scalefocused on in this study. Thirty day averages over a 15year time series (again from the hourly wind statistic atCGBAPS) for the period 1988–2003 are shown in Fig. 2.This result shows inter-annual variation, however theclimatologically averaged winds are always south-westerly with a dominant westerly component.

Tracer concentrations represent the volume fractionof particular tracer in the total volume of the watercolumn. Predefined source regions are initialized withtracers A, B and C at unit concentration. An area in thedomain is delimited to represent the strait interior and

27P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

initialized with tracer D at unit concentration. Bound-aries representing these regions are shown in Fig. 1.Zero gradient open-sea boundary conditions for tracersare adopted. Tracers A, B and C are placed in locationswhere primary source water masses occur (Fig. 1). Afteran elapsed time, tracer concentrations represent thefraction of source water mass in the water columncombined with a boundary source. EACW is disre-garded as a source because typical winter-springseasonal mean winds are westerly. Resulting distribu-tions of tracer concentration are shown after certainelapsed times and indicate the instantaneous fraction of

Fig. 2. Climatologically averaged winds at Cape Grim, Tasmania,

between 1988 and 2003. Averaging periodZ30 days. Winds taken from

hourly averaged wind vector time series courtesy of the Australian

Commonwealth Bureau of Meteorology.

each in the water column. Far field forcing modulatesthe intensity and flow directions of SACW and SASW.This has not been accounted for in the present studywhich only attempts to investigate influences of thesewater masses assuming constant source at the bound-aries.

Flushing times are calculated using tracer D (Fig. 1).An arbitrary minimum tracer concentration is requiredto define the flushing time. Walker (1999) uses the time itwould take for tracer to reach 1/e or w37% of its initialconcentration for estimating flushing times of PortPhillip Bay with respect to Bass Strait. For comparisonthis is adopted. The flushing time is recorded when localconcentrations of tracer D have decreased to w0.37 oftheir initial value. When tracer D concentration reachesthis minimum the remaining volume fraction is w0.63water mass originating outside the predefined bound-aries. At this minimum the local water column flushingtime is recorded.

The model is run for an initial period of 24 hours asthe wind is gradually increased from zero to full strengthto minimize inertial oscillations. Tracer is then intro-duced into the simulation. Parameters and constantsassociated with the experiments are given in Table 1.

3. Results

The residual sea level anomaly (contours in cm) andresidual current vector field for the entire simulation isshown in Fig. 3. This conveniently summarizes the sealevel response and currents during the period. It showsthat the residual sea level and currents are a signature of

Fig. 3. Residual sea level anomaly (cm) and residual current vector field (cm s�1) after 180 days simulation time.

28 P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

the westerly wind-driven response. This is characterizedby a lowering of sea level at the south-west corner anda rising of similar magnitude along much of thesouthern Victorian Coastline east of Port Phillip Baywith a zero sea level contour running from the north-west to the south-east in an almost diagonal fashion.Similar findings can be seen in the studies of Fandry(1981), Middleton and Viera (1991), Middleton andBlack (1994) and Bruce et al. (2001). Currents associatedwith this sea-level set-up are primarily aligned with sea-level contours indicating the westerly wind-drivenresponse is essentially geostrophic in nature. Whenwind-driven currents are super-imposed on residual tidalcurrents, wind-driven currents play the primary role inhorizontal transport processes in the region at timescaleslonger than the order of several days.

Fig. 4 illustrates flushing times after 180 dayssimulation time. A stagnation-area of long (O160 days)flushing times is evident in central Bass Strait. A zone oflong flushing times also extends from the stagnation-area to Bass Canyon. The important conclusion that canbe made from this is that the stagnation-area and thezone about the eastern margin with relatively longflushing times appear to be the areas where the oldestwater is in Bass Strait. This implies that water in thiszone is most affected by local air-sea buoyancy fluxesand this zone is likely to be where dense water formationoccurs. Water from this zone may trigger or be a sourceof the Bass Strait Cascade.

Fig. 4. Flushing times (days).

Concentrations of water masses A and D are shownin Fig. 5 at the end of each month in the 180 daysimulation. Final concentrations of tracers C and B areshown in Fig. 6. All concentrations represent the volumefraction of source water mass in the water column. Theimportance of water mass A in this period is evident.The movement of the tracers reveals that shelf-water Centering the strait from outside the north-western corneris advected eastwards and mostly adheres to theVictorian coastline. A small proportion of this waterbranches off just south of Wilson’s Promontory andflows south-eastwards towards Flinders Island. Shelf-water B moves into Banks Strait and northwards pastFlinders Island but does not enter Bass Strait in anysignificant proportion. Shelf-water A is mostly trans-ported into Bass Strait. This water flows northwardsthrough the passage between King Island and CapeGrim. Some is rapidly advected eastwards along thenorthern Tasmanian Coastline, whereas a large pro-portion is entrained in the residual circulation in thestrait. There is a northern limit at which these watersreturn to flow south-east towards the southern side ofFlinders Island. They may meet the eastward flow alongthe northern Tasmanian coastline or be caught innorthward flow in eastern Bass Strait around FlindersIsland. It is possible for water adhering to the northernTasmanian coastline to be advected through BanksStrait and be mixed with water originally adjacent tonorth-eastern Tasmania. Shelf-water A is most widelydispersed in Bass Strait. There is a position near 40 �S,146 �E where a branching of the flow occurs. Some waterfrom here is eventually advected north-eastwards di-rectly out of the strait and some goes southwards tospend further time in the strait.

Analysis of the fraction of each water mass A, B, andC in the total local mass of water after 180 days yieldsinsight into their respective relative contributions. Themost significant water mass involved in the flushingof strait waters in winter-spring is water mass A.Water mass C is present in the lowest concentrationspresumably resulting from advection out of the north-western boundary. Results also show that the stagna-tion-area contains w50% water mass D and theremaining fraction is water masses A and C. Of thiswater mass A is w90% of the mixture of A and C.Water mass B has a less significant influence on flushingin the strait, however it is significant in flushing parts ofnorth-eastern Tasmanian coastal waters and watersalong the inner side of the eastern shelf-break.

4. Discussion

Additional experiments were carried out with themodel using transient synoptic scale winds. Theseconfirm that flushing is controlled by the mean

29P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

Fig. 5. Model predictions of the advection and diffusion of tracer in Bass Strait. Tracer A (left) represents shelf water originating from north-western

Tasmania and Tracer D (right) represents Bass Strait Water. Concentration is normalized and represents fraction of source water mass in water

column.

30 P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

climatologically averaged winds. The main findings ofthe 180 day simulation suggest winter-spring flushingof Bass Strait waters results from eastward advection ofSACW and SASW. Flushing in the central area dependson longer term mean winds (weeks to months) ratherthan shorter term winds (tens of hours). Time scales forflushing vary according to mean wind strength. Resultsalso suggest strait waters can be replenished to somedegree in most places with SASW (excepting minuteconcentrations in the stagnation-area) in a period ofapproximately 30 days in conditions of strong meanwesterly winds.

There are a number of noteworthy features in thetemperature and salinity distributions by Tomczak(1985) that can be compared to model results. The13.5 � isotherm in the June/July observations showa north-eastward flow from the passage between KingIsland and Cape Grim that turns south-eastward andheads back towards Flinders Island and then makes itsway northwards along the eastern edge of the strait. Thetongue of low salinity (!35.3) indicates intrusion of

Fig. 6. Concentration of tracers B (a) and C (b) after 180 days

simulation time.

water through the south western passage. Contoursof sT indicate that flow from the south-western corner issimilar to that suggested above. Water in central north-eastern Bass Strait appears denser than water elsewhere.The main results in this study support the inferred flowsfrom these temperature and salinity observations.Nutrient distributions by Gibbs et al. (1986) for July1980 provide evidence of lower levels of nitrate, nitrite,phosphate and silicate relative to the eastern shelf break.This could be evidence for a stagnation-area. Tidalcauses for these observations were suggested by Gibbset al. (1986). Tides may be only partially responsible.Results also show water masses A and B dominate themixture along the eastern margin of Bass Strait. Thisstudy shows that meso-scale wind-driven advectionthrough the passage between King Island and CapeGrim overcomes strong tidal fronts and the transport ofwater mass during the winter-spring period is almostalways into Bass Strait. This means that water mass Acontinually replenishes waters in this passage and in thestrait during the winter-spring period.

Sources of error in the model arise from dynamicalapproximations, topographic errors, finite differenceapproximation truncation errors and interpolationerrors in the representation of coastlines and islandson the grid (McIntosh and Bennett, 1984). Despite theimportance of tidal currents that cause strong verticalmixing at the edges of the strait, wind-driven currentsdetermine the overall seasonal-scale circulation andflushing. The scale of residual tidal currents is relativelysmall compared to the scale of wind-driven currents.The symmetric nature of tidal currents means thatresidual flow is dominated by wind-driven processes.Issues of uncertainty in the bathymetric data and in thespatial distribution of winds in the region are the mostimportant sources of uncertainty in determining thewinter-spring flushing of strait waters.

5. Conclusions

The study provides a first approximation of thewinter-spring flushing of Bass Strait in un-stratifiedconditions. It also highlights the dominance of meanwind driven flow over tidal flow at the seasonal scale.Wind-driven depth-averaged currents are largely topo-graphically controlled and geostrophic in nature. Thesecurrents determine meso-scale residual flow in BassStrait in the winter-spring period and the presence of thestagnation-area depends on this. Advection of tracerfrom the three different locations suggests shelf watermass from the south-western corner of the region is themost widely dispersed and rapidly transported watermass in the strait in the winter-spring period. Winterflushing with SASW is a significant inter-annual pro-cess replenishing nutrients and supporting ecosystems.

31P.A. Sandery, J. Kampf / Estuarine, Coastal and Shelf Science 63 (2005) 23–31

Water in the stagnation-area takes the longest time to bereplenished by external water mass and occurs attimescales of the order of O6 months. Water enteringthe strait through the south-western passage is likely tobe caught in eddies in western Bass Strait beforereaching the stagnation-area and being exported fromthe strait. The mean westerly winds in the winter-springperiod bring about net eastward flow in the strait andflushing from the west. Water mass transported from thewest has a greater relative influence than that from theeast. A significant volume of water remains in the straitfor periods of the order of months to seasons. The‘stagnation-area’ in the eastern area is a dynamicalaspect of the dense water formation process. Whethercoastal waters along the north-western Tasmaniancoastline are nutrient rich as a result of upwelling, tidalvertical mixing or riverine inputs, in typical winter andspring conditions, the properties of these waters can bewidely dispersed in Bass Strait.

Acknowledgements

The authors express appreciation to Matthias Tomc-zak of Flinders University and Scott Condie of CSIROMarine Research for reviewing the manuscript. Datasets were obtained from the National Oceanographicand Atmospheric Administration (NOAA), c/- theNational Geophysical Data Centre, USA (ETOPO2),the Australian Bureau of Meteorology and the NationalTidal Facility of Australia. Wind data was provided bythe Cape Grim Baseline Air Pollution Station c/- ArthurDowney of the Australian Bureau of Meteorology.

References

Baines, P.G., Fandry, C.B., 1983. Annual cycle of the density field in

Bass Strait. Australian Journal of Marine and Freshwater Research

34, 143–153.

Blackman, D.R., Hinwood, J.B., Lleonart, G.T., 1987. Temperature

anomaly in Western Bass Strait. Australian Journal of Marine and

Freshwater Research 38, 191–195.

Bruce, B.D., Condie, S.A., Sutton, C.A., 2001. Larval distribution of

blue grenadier (Macruronus novaezelandiae Hector) in south-eastern

Australia: further evidence for a second spawning area. Australian

Journal of Marine and Freshwater Research 52, 603–610.

Evans, S.R., Middleton, J.F., 1998. A regional model of shelf

circulation near Bass Strait: a new upwelling mechanism. Journal

of Physical Oceanography 28, 1439–1457.

Fandry, C.B., 1981. Development of a numerical model of tidal and

wind driven circulation in Bass Strait. Australian Journal of

Marine and Freshwater Research 32, 9–29.

Fandry, C.B., Hubbert, G.D., McIntosh, P.C., 1985. Comparison of

predictions of a numerical model and observations of tides in Bass

Strait. Australian Journal of Marine and Freshwater Research 36,

737–752.

Gibbs, C.F., Tomczak Jr., M., Longmore, A.R., 1986. The nutrient

regime of Bass Strait. Australian Journal of Marine and

Freshwater Research 37, 451–466.

Jones, I.S.F., 1980. Tidal and wind-driven currents in Bass Strait.

Australian Journal of Marine and Freshwater Research 31, 109–

117.

Kowalik, Z., Murty, T.S., 1993. Numerical Modeling of Ocean

Dynamics, Advanced Series on Ocean Engineering, vol. 5. World

Scientific Publishing, 496 pp.

Luick, J.L., Kase, R., Tomczak, M., 1994. On the formation and

spreading of the Bass Strait cascade. Continental Shelf Research

14, 385–399.

Neil, A. (Ed.), 2000. Under Southern Seas: The Ecology of Australia’s

Rocky Reefs, Malabar FLa, Kreiger. UNSW Press, Sydney,

238 pp.

McIntosh, P.C., Bennett, A.F., 1984. Open ocean modelling as an

inverse problem: M2 tides in Bass Strait. Journal of Physical

Oceanography 14, 601–614.

Mesinger, F., Arakawa, A., 1976. Numerical Methods Used in

Atmospheric Models, vol. 1. GARP Publications Series No. 17,

World Meteorological Organization, 64 pp.

Middleton, J.F., Black, K.P., 1994. The low frequency circulation in

and around Bass Strait: a numerical study. Continental Shelf

Research 14, 1495–1521.

Middleton, J.F., Viera, F., 1991. The forcing of low frequency motions

within Bass Strait. Journal of Physical Oceanography 21, 695–708.

Newell, B.S., 1961. Hydrology of south-eastern Australian waters:

Bass Strait and New South Wales tuna fishing area. CSIRO

Australian Division of Fisheries and Oceanography, Technical

Paper No. 10.

Palma, E.D., Matano, R.P., 1998. On the implementation of passive

open boundary conditions for a general circulation model: the

barotropic mode. Journal of Geophysical Research 103, 1319–

1341.

Tang, Y.M., Grimshaw, R., 1999. Recent developments in the theory

and modelling of storm surges. In: Noye, B.J. (Ed.), Modelling

Coastal Sea Processes. World Scientific Publishing, pp. 135–157.

Tomczak Jr., M., 1985. The Bass Strait water cascade during winter

1981. Continental Shelf Research 4, 255–278.

Tomczak Jr., M., 1987. The Bass Strait water cascade during summer

1981–1982. Continental Shelf Research 7, 561–572.

Walker, S.J., 1999. Coupled hydrodynamic and transport models of

Port Phillip Bay, a semi-enclosed bay in south-eastern Australia.

Australian Journal of Marine and Freshwater Research 50,

469–481.