Physical processes along the southern continental shelf and slope of Western Australia

Transcription

1 Physical processes along the southern continental shelf and slope of Western Australia Mohd Fadzil Mohd Akhir, BEng (Hons) This thesis is presented for the degree of Doctor of Philosophy At The University of Western Australia School of Environmental System Engineering January 2010

2

3 Abstract The circulation along the south coast of Western Australia was examined using field data and numerical modelling. Physical processes in this region, particularly along the continental shelf and slope regions, were poorly understood due to a paucity of field measurements. Data were collected during a research cruise on RV Southern Surveyor (04/2006) during April 2006 consisting of 18 CTD transects from Twilight Cove (126 o E) to Cape Leeuwin (115 o E) and was augmented by shipborne ADCP data. The field data set provided a detailed understanding of three major current systems: Leeuwin Current (LC), Leeuwin Undercurrent (LU) and Flinders Current (FC). The LC along the south coast exhibits different characteristics when compared to that along the west coast. The LC flows into the colder and lower salinity subantarctic environment of the south coast. This is evident in a strong geopotential gradient off the south-west corner of Australia (Cape Leeuwin) resulting in rapid acceleration of the LC as it reaches a maximum velocity in this region. Numerical modelling studies, using the Regional Ocean Modelling System (ROMS) indicated that wind stress is an important component of the dynamics in this region. This was identified when comparing summer and winter conditions when the winds act in opposite directions, from north-westerly to southeasterly respectively. Along the shelf break and slope, the Flinders Current (FC) interacts with LC. As the dominant current, the FC serves both as a surface and as an undercurrent, transporting sub Antarctic mode water (SAMW). This interconnection the FC and LU can be seen clearly from the salinity, temperature and velocities within the depth range m postulating a connection between subsurface waters off Tasmania (origin of the Flinders Current) and the tropical Indian Ocean through the Flinders and Leeuwin Undercurrents. 1

4 2

5 Contents Abstract... 1 Acknowledgments... 7 Statement of candidate contribution... 9 List of Figures List of Tables Chapter Introduction Summer physical processes along the continental shelf off southern Western Australia: To define Chapter Local setting Geography Bathymetry Climate The current system along the south coast of Western Australia The Leeuwin current Physical oceanography principles The Leeuwin undercurrent The Flinders current Coastal currents Water mass characteristics The regional ocean modelling system (ROMS)

6 2.5 Conclusion Chapter Introduction Methodology and study domain Study region Results and discussion Surface characteristics Subsurface characteristics Water mass characteristics Alongshore geopotential gradient Influence of wind Submarine canyon upwelling Concluding remarks Chapter Introduction Study region, data and methods Water mass structure The FC and LU water mass structure Figure 4.3. Alongshore distributions of (a) temperature, (b) salinity and (c) dissolved oxygen along the 1000-m isobath for transects R, P, N, M, L and K (the unit of DO is μm/l) Oxygen maximum Circulation Discussion Concluding remark Chapter Introduction Model set-up

7 5.3 Model results Seasonal characteristics The influence of wind stress Discussion The Leeuwin current The Leeuwin undercurrent and the Flinders current The Capes and Cresswell currents Concluding remarks Chapter References

8 6

9 Acknowledgments In the name of God the most gracious and most merciful; All praise is due to Him. Sincere thanks to my supervisor Prof. Chari Pattiaratchi, who trained me to listen, speak, write and understand the charming language of oceanography. The guidance and support is persevering, and my PhD journey has been an incredibly valuable learning experience. Dr Michael Meuleners has contributed a valuable support in the numerical modelling works. He always is so generous in sharing knowledge and so considerate in lending his time. Also, many thanks to the crew members of the Southern Surveyor 04/06 and scientific staff aboard, especially Dr Florence Verspecht and Dr Lai Mun Woo for being a good team member along the 3 weeks journey. Tremendous support from the crew members is highly appreciated especially in sharing their thoughts and concern, when I went through painful moment on the rough Southern Ocean. Special thanks to academic and staff members of School of Environmental Systems Engineering (SESE) who provided marvellous support to me during the past 3 years. This also includes my fellow postgraduates and post-docs, the friendships have made this journey smoother and more enjoyable. Special thanks to Ruth, for correcting my English for most of my work, I m sure was not an easy task. My family back home in Malaysia, especially Bapa, Ibu, K.Long, Abang and K.Lang have always been supportive, also to my in-laws and close family who always wishing me the best of luck. To my fellow Malaysians in Perth, your presence always makes us feel like home. And the most important person in my life, my beloved wife Azza. Sharing, the love, laugh and sadness throughout this journey is the most memorable moment that I will cherish forever. And of course the prince and princess of my heart, Al Fateh and Alwani, 7

10 the charm look of their eyes and the sweetness of their smile always made me miss them, especially when I m in front of the PC completing this work. This work could not have been undertaken without the support of the Ministry of Higher Education of Malaysia (KPT) and Malaysia University of Terengganu (UMT) scholarship. The University of Western Australia and SESE provide financial assistance for my trip to Orlando, Florida, USA to present my scientific findings at the Ocean Science Meeting

11 Statement of candidate contribution I hereby declare that all material presented in this thesis is original except where due acknowledgment is given, and has not been accepted for the award of any other degree or diploma. The content of this thesis is the author s own work under the supervision of Prof. Charitha Pattiaratchi. The work contain herein was undertaken at the University of Western Australia between April 2005 and June The main part of this thesis, Chapters 3 to 5, consists of three journal papers which have been prepared for publication and some repetition of the literature review, study site details and methodology has therefore been necessary. The three papers will be submitted to Continental Shelf Research, Deep Sea Research and Ocean Dynamics respectively. 9

12 10

13 List of Figures Figure 1.1. Schematic of the Australasian ocean currents from Halligan (1921) showing a warm current flowing south along the Western Australian coast and turning east at the south-west corner (Cresswell, 1991) Figure 1.2. The distribution of the temperature and salinity casts as a function of position: the small, black dots represent all casts deployed on the continental shelf and slope; the larger, grey dots represent all casts of > 500-m depth (Ridgway and Condie, 2004) Figure 2.1. Location map and bathymetry profile of the study region offshore of south-west Australia Figure 2.2. Section of submarine canyon topography in the south coast Figure 2.3. Wind roses for summer and winter, 2004, at the two coastal stations: Windy Harbour and Hopetoun North Figure 2.4. a) The summer wind pattern. The prevailing winds are westerly over the Great Australian Bight and the high pressure centre is inland. b) The winter wind pattern. The prevailing winds are south-easterly over the south of Western Australia and the high pressure centre is over the Great Australian Bight (Tapper and Hurry 1996) Figure 2.5. Schematic of surface and subsurface currents along the continental shelf and slope off WA Figure 2.6. Depth average velocity from the numerical model of Middleton and Platov (2003) as driven by summertime mean winds. Sverdrup transport acting towards the land mass Figure 2.7. Sea surface temperature image for 9 April 2006 showing the warm core of the LC flowing south towards the south-west corner. The cooler FC, which was flowing westward in the south, separated from the LC at the shelf slope Figure 3.1. Location map of the study area including the Southern Surveyor 04/06 cruise track through south of Western Australia (WA) Figure 3.2. Sea surface temperature (SST) distribution along cruise track Figure 3.3. Surface salinity distribution along cruise track Figure 3.4. Surface current distribution obtained from the shipborne ADCP data Figure 3.5. Altimter image showing sea level anomaly obtained on 1 May Locations of offshore eddies are shown Figure 3.6. Transect P: Cross section of (a) temperature, (b) salinity and (c) alongshore velocities Figure 3.7. Transect K: Cross section of (a) temperature, (b) salinity and (c) alongshore velocities Figure 3.8. Transect G: Cross section of (a) temperature, (b) salinity and (c) alongshore velocities

14 Figure 3.9. Transect A: Cross section of (a) temperature, (b) salinity and (c) alongshore velocities Figure Velocity profile along the line of maximum of LC flow. The dashed line indicates 0 m/s Figure Temperature/salinity diagram of the surface stations at transects R, K, G and A Figure Geopotential anomaly at 1000m depth (a) above 300m (b) below 300m to 750m Figure Wind stress onboard Southern Surveyor 04/06 for transect K and G Figure Section of submarine canyon topography in the south coast (from Geoscience Australia). The upwelling was observed at Albany Canyon and Bremer Canyon Figure Alongshore transects along the 1000-m contour: (a) temperature, (b) salinity and (c) dissolved oxygen Figure Schematic diagram of the current system off Western Australia Figure 4.1. The study region showing the location of the CTD transects and bathymetry contours. The regions referred to in the text are also shown Figure 4.2. T S diagram of the transects representing the south (transects L and K), west (transects R and P), and south-west corner (transect N) of Western Australia Figure 4.3. Alongshore distributions of (a) temperature, (b) salinity and (c) dissolved oxygen along the 1000-m isobath for transects R, P, N, M, L and K (the unit of DO is μm/l) Figure 4.4. Dissolved oxygen concentration across (a) transect K and (b) transect R Figure 4.5. Velocity transect for transect K on the shelf slope Figure 4.6. Velocity transect for transect L on the shelf slope Figure 4.7. Velocity transect for transect N on the shelf slope Figure 4.8. Velocity transect for transect P on the shelf slope Figure 4.9. Velocity transect for transect R on the shelf slope Figure Illustration of the (a) surface and (b) subsurface circulation off the south-west corner of Western Australia showing the effect of the eddy on the undercurrent circulation Figure Generational circulation of the Flinders current, Leeuwin current, and undercurrent off the south-west corner of Western Australia with the cross-sections of transect R (A) and transect K (B) Figure Schematic showing the inferred pathway of the south Indian Ocean SAMW...58 Figure 5.1. a) Model domain with bathymetry contour lines. b) The study area overlaid with the field transect of Southern Surveyor cruise SS04/06. The current system was observed at two location of the same transect: a) Transect R: Capes region (latitude 34ºS) ; and b) Transect K :Albany (longitude 117.7ºE) Figure 5.2. a) NOAA satellite image of south-west Australia in 9 April The Warm water (shown in red/orange) represents the Leeuwin Current. b) A model image result 12

15 representing autumn circulation, showing eddy offshoot south of Cape Leeuwin while the Leeuwin current is close to the shore Figure 5.3. Model output with wind stress forcing representing summer conditions. a) Sea surface temperature ( C) and current velocites; b) the cross-section across transect R (Capes region) showing the north-south velocity (m/s); c) the cross-section across transect K (Albany) showing the east-west velocity (m/s) Figure 5.4. Model output with wind stress forcing representing autumn conditions. a) Sea surface temperature ( C) and current velocites; b) the cross-section across transect R (Capes region) showing the north-south velocity (m/s); c) the cross-section across transect K (Albany) showing the east-west velocity (m/s) Figure 5.5. Model output with wind stress forcing representing winter conditions. a) Sea surface temperature ( C) and current velocites; b) the cross-section across transect R (Capes region) showing the north-south velocity (m/s); c) the cross-section across transect K (Albany) showing the east-west velocity (m/s) Figure 5.6. Model output without wind stress forcing. a) Sea surface temperature ( C) and current velocites; b) the cross-section across transect R (Capes region) showing the northsouth velocity (m/s); c) the cross-section across transect K (Albany) showing the east-west velocity (m/s) Figure 5.7. The velocity field at the depth of the Flinders current (550 m) for the autumn circulation. The colour bar indicates the velocity in m/s Figure 5.8. a) Sea surface temperature ( C) and Capes current velocity for the Capes region in summer. b) Sea surface temperature ( C) and Cresswell current velocity for the Albany region in summer. c) Transect R cross-section of the Capes current north-south velocity (m/s) in summer. d) Transect K cross-section of the Cresswell current north-south velocity (m/s) in summer. e) The Capes current summer temperature ( C) at transect R (Capes region). f) The Cresswell current summer temperature ( C) at transect K (Albany). The colour bars indicate the current velocity in m/s Figure 5.9. a) Sea surface temperature ( C) and Capes current velocity for the Capes region in autumn. b) Sea surface temperature ( C) and Cresswell current velocity for the Albany region in autumn. c) Transect R (Capes region) cross-section of the Capes current northsouth velocity (m/s) in autumn. d) Transect K (Albany) cross-section of the Cresswell current north-south velocity (m/s) in autumn. The colour bars indicate the velocity in m/s Figure The summer and autumn wind vectors close to the nearshore in the Capes region and at Albany. The wind stress forcing was extracted from the National Oceanic and Atmospheric Administration daily wind stress data Figure Schematic of the summer steady state current regime off Western Australia s south coast

16 List of Tables Table 2.1. The characteristics of the water masses found in the 1-km-deep water column defined. The table combines data from voyages FR10/2000 ( S) and SS09/2003 ( S) Table 3.1 Summary of estimated alongshore geopotential gradients at the surface relative to the 300 db (west coast) and subsurface at 300/750 db...36 Table 5.1. Geopotential anomaly slope for summer, autumn and winter at: a) maximum LC and b) specifically at the southwest corner. Geopotential anomaly was calculated between surface and 300m depth along the LC core Table 5.2. Volume transport of the Leeuwin current, Leeuwin undercurrent and Flinders current in summer and winter across transects R (Capes region) and K (Albany) with and without wind stress (units: Sverdrup)

17 Chapter 1 Introduction The famous journey of Captain Matthew Flinders around Terra Australis, or what is now known as Australia, took place between 1801 and Flinders surveyed the Australian coastline, including New Holland (Western Australia) (Flinders, 1814), aboard the Investigator. During this voyage Flinders (1814) documented an eastward current along the south coast of Western Australia with speeds up to 1¼ knots (0.6 ms -1 ) recorded off Point D Entrecasteaux and Albany and is the earliest recognition of the Leeuwin current in this region. Halligan (1921) produced a more detailed current map, which showed the warm Leeuwin current flowing south along the west Australian coast and turning into the Great Australian Bight (Figure 1.1). Figure 1.1. Schematic of the Australasian ocean currents from Halligan (1921) showing a current flowing south along the Western Australian 15 coast and turning east at the south-west corner (Cresswell, 1991).

18 In recent years, many studies of the ocean circulation system along the Western Australian coast have been undertaken. These include those undertaken by Andrews (1977); Cresswell and Golding (1980); Hamilton (1986); Smith et al. (1991); Pearce and Walker (1991); Cresswell and Peterson (1993); Gersbach et al. (1999); Pearce and Pattiaratchi (1999); Feng et al. (2003, 2005); Morrow et al. (2003); Ridgway and Condie (2005); Woo et al. (2006); Rennie et al. (2007); Meuleners et al. (2007); and, Pattiaratchi and Woo (2009). However, most of these studies were concentrated along the west coast of Western Australia examining the structure and the dynamics of the Leeuwin Current. Pearce (1991) and Smith et al. (1991) described the circulation off the Western Australian coast as different from that of any other western continental margin. In each of the main ocean basins, the surface circulation forms a gyre with a poleward flow along the western boundary and an equatorward flow along the eastern boundary. This gyre which is forced by wind system is also responsible to cause upwelling and regions of high productivity along the eastern margins (e.g. off south America and south Africa). In contrast, off the WA coast, the Leeuwin current transports water of tropical origin from the North-West Shelf poleward to Cape Leeuwin and into the Great Australian Bight (Church et al. 1989; Smith et al. 1991; Ridgway and Condie 2004). The Leeuwin current (LC) controls the climate and marine life in the south-west region. The current affects seagrass and algae distribution; coral spawning and distribution; western rock lobster, coastal scallop, and fin fish stocks distribution; and the life cycle of the southern blue fin tuna (Thunnus maccoyii) (Pearce, 1991). The presence of tropical marine organisms off the west Australian coast (Maxwell and Cresswell 1981) and the higher winter air temperatures and rainfall in the region, compared with similar latitudes elsewhere, may also be attributed to the current (Telcik and Pattiaratchi, 1998). There is a paucity of field data along the south coast of Australia (Figure 1.2 and Cresswell and Domingues, 2009) and in this thesis field data obtained aboard the Southern Surveyor (research voyage SS04/2006) are presented. The 20-day voyage (12 April to 1 May 2006), which departed from Esperance and ended at Fremantle, covered 17 transects and more than 200 CTD stations. The data collected on this research cruise forms a step change in the physical oceanographic data along the southern coast of Western Australia. 16

19 Numerical modelling studies were also conducted to reproduce the surface and subsurface currents to better understand the dynamics of the system. The modelling studies examined the seasonal forcing (which was impossible to examine from the field data), the presence of a coastal current, and the effect of wind forcing on the current systems. The model s temporal resolution indicated the differences in the circulation patterns between the seasons, especially between summer (when the LC was weak) and winter (when the LC was strong). The aim of this study was to examine in detail, the flow characteristics along the southern coast of Western Australia in particular the dynamics of the Leeuwin Current, Leeuwin Undercurrent, Flinders Currents, shelf current systems and their interaction. Numerical modelling was used to identify temporal changes in the system to complement the field data which only provided a snap shot of the processes during April The presence of a number of submarine canyons in the region also provided an opportunity to examine the localized circulation patterns, in particular, evidence of upwelling at the heads of the canyons. Figure 1.2. The distribution of the temperature and salinity casts as a function of position: the small, black dots represent all casts deployed on the continental shelf and slope; the larger, grey dots represent all casts of > 500-m depth (Ridgway and Condie, 2004) 17

20 The main themes are divided into three chapters, with the contents of each chapter as follows: 1. The Leeuwin Current along the south coast of Western Australia during autumn: To define LC s water mass characteristics as the LC flows along the south coast the surface and subsurface currents and their velocity distribution the dynamics of the geopotential gradient along the south coast 2. The interconnection of the Flinders current and Leeuwin undercurrent: To define the undercurrent s water mass structure and its continuation at the southwest corner the currents velocity, continuity, and interconnection at the south-west corner whether the subantarctic mode water flows to the tropics through this interconnection 3. The Leeuwin current and surrounding system: a numerical study: To define the surface and subsurface currents on the continental shelf and slope in the southern region the annual cycle and seasonal variation of the surface and subsurface currents the effect of wind stress and the geopotential gradient on the surface current the summer coastal current (Cresswell current) and the effect of upwelling along the south coast 18

21 Chapter 2 Literature review 2.1 Local setting Geography Western Australia covers a third of the Australian continent and is surrounded by two oceans: the Indian Ocean to the west and the Southern Ocean to the south. Western Australia is unique in that its zonal coastline is the longest in the Southern Hemisphere. Knowledge of the geographic, bathymetric, and climatic conditions offshore of southwest Australia is critical to understanding the oceanographic circulation properties in the region. The study region covered the area between the Capes region (115 E 34 S) and Esperance (126 E 34 S), where the coastline has several main changes. Starting from Cape Leeuwin, the first, and biggest, change is at the corner of Western Australia near Point D Entrecasteaux (Figure 2.1). The region between Cape Leeuwin, Pt D Entrecasteaux, and Albany can be considered the south-west corner. The shelf span in this area is between km width. The coastline after Albany turns into a much wider shelf close to Hopetoun. The additional of nearly km of shelf width after Albany is clear from the bathymetry data. The last change is after Esperance which then forms the Great Australian Bight. The coastline s unique shape affects the oceanic processes, such as eddies, offshoots, and meanders, in the region. The continental shelf contains several submarine canyons. Most of these canyons extend from large rivers on the mainland. Studies conducted in the Perth Canyon on the west coast revealed the canyon attributed to upwelling and enriched the photic zone with nutrients on the shelf beyond its immediate vicinity. The patches of cool water observed 19

22 around the canyon suggested the canyon contributed to the physical processes, such as upwelling, in the region Bathymetry The bathymetry of the south-west Australian continental shelf revealed the shelf was narrower than the continental shelf on the west coast. The shelf width (~25 km) is mostly uniform throughout the region, but is narrowest near Albany. At the farthest CTD station at Twilight Cove, near the Great Australian Bight (GAB), the shelf starts to broaden; here the recorded shelf width is almost three times the average. The depth contour is usually parallel to the coast; beyond the continental shelf break, the depth increases from 400 to 4000 m. The slope is mild from the Capes region to the south-west corner, and sharp from Pt D Entrecasteaux to Esperance. A steep continental slope is thought to produce a strong slope current. Fremantle Figure 2.1. Location map and bathymetry profile of the study region offshore of south-west Australia. 20

23 Figure 2.2. Section of submarine canyon topography in the south coast (from Geoscience Australia). The south coast consists numbers of submarine canyon (Figure 2.2). Previous studies suggest that deep submarine canyons which reach depths ~500m and combined with strong alongshore flow can produce upwelling (Kampf, 2007). Figure 2.2 shows few submarine canyons, Albany Canyon and Bremer Canyon, which is among the deepest in the region. Relatively strong alongshore current in the south coast maybe conducive to upwelling through the mechanisms provided by submarine canyon presence (Kampf, 2007) Climate The south of Western Australia experiences hot, dry summer and cool, wet winters defined as a Mediterranean climate. The mean maximum temperature from February to July ranges from 29 to 16 C. The migration of the subtropical baric ridge (atmospheric barometric pressure), which moves from the south in summer to the north in winter, mainly controls the climate. Cold fronts embedded in westerlies, which can cause heavy rainfall, are common in winter; however, the ocean temperature helps moderate the overnight temperature. Sea breezes from the Southern Ocean, which occur often in 21

24 summer, also provide a mild, moist air mass throughout the year (Tapper and Hurry, 1996). The wind regime varies seasonally: winds are usually north-westerly in winter and southeasterly in summer. These can be seen from the wind roses plotted at two coastal locations: Windy Harbour and Hopetoun North (Figure 2.3). The figure represents the data for year This data is chosen specifically to represent the recent pattern in normal year which is free from any ENSO activity. The dominant wind direction turns almost 180 between winter and summer. The migration of the subtropical ridge (Figure 2.4) between the seasons causes these winds to change so significantly. The wind regime in the southern region is unique, given the fact that, the wind regime is favourable to the movement of LC in winter and opposing it in summer. The effect of this wind composition is still poorly understood, and it is important for this study to address the importance of the wind stress in assisting and also confining the current movement in both winter and summer, accordingly. Summer Winter Summer Winter Figure 2.3. Wind roses for summer and winter, 2004 at two coastal stations: Windy Harbour and Hopetoun North. Data obtained from Bureau of Meteorology 22

25 a) H b) H Figure 2.4. a) The summer wind pattern. The prevailing winds are westerly over the Great Australian Bight and the high pressure centre is inland. b) The winter wind pattern. The prevailing winds are south-easterly over the south of Western Australia and the high pressure centre is over the Great Australian Bight (Tapper and Hurry 1996). 23

26 2.2 The current system along the south coast of Western Australia The Leeuwin current The circulation off Western Australia is anomalous compared with other western continental margins (Godfrey and Ridgway, 1985; Pearce, 1991; Ridgway and Condie, 2004; Schott, 1935; Smith et al., 1991). In all ocean basins, the surface circulation forms a gyre with a poleward flow along the western boundary. This system which driven by favourable wind causes upwelling and regions of high primary productivity (e.g. off south America and south Africa). In contrast, off the Western Australian coast, the Leeuwin current (LC) transports warm, low salinity, nutrient-depleted water of tropical origin from the north-west of Australia south along the Western Australian coast. The current, which is shallow (< 300 m) with a narrow band (< 100 km) and strongest in winter, flows south from Exmouth to Cape Leeuwin then east into the Great Australian Bight (GAB), following the continental shelf break and upper continental slope. The LC signature extends from North West Cape to Tasmania (41 S, 145 E) a stretch of 5500 km making it the longest boundary current in the world (Ridgway and Condie, 2004). The geopotential gradient drives the LC, especially along the west coast, despite a strong opposing wind stress (Batteen and Rutherford, 1990; Godfrey and Ridgway, 1985; Godfrey and Weaver, 1991; Pattiaratchi and Buchan, 1991; Thompson, 1984, 1987; Weaver and Middleton, 1989; Woo and Pattiaratchi., 2008). This geopotential gradient overwhelms upwelling-favourable winds and causes downwelling at the coast. This downwelling is induced by ekman transport, generated by strong current flowing towards east. Onshore geostrophic flow from the central Indian Ocean towards WA occurs between 15 and 35 S. Between 15 and 28 S, geostrophic inflow from the west, augmented by tropical water from the North-West Shelf, forms the LC s warm, low salinity core (Smith et al. 1991; Woo et al. 2006; 2008). The changes in the local wind stress and pressure gradient also reflect the current s seasonal variation. The LC is weakest in summer because of the opposing equatorward wind stress and strongest in winter because of a weaker equatorward wind stress and a higher pressure gradient (Godfrey and Ridgway 1985). 24

27 Cresswell and Golding (1980) found that, in the southern region, the LC turned at Cape Leeuwin and flowed east into the GAB. The current is influenced by the Coriolis force to turn left at Cape Leeuwin and then move eastwards. The first major study of the currents off southern Australia was conducted aboard the RV Sprightly (SP6/82) in Here LC speeds of 1.5 m/s were recorded near Cape Leeuwin (Godfrey et al. 1985). Earlier, a research cruise undertaken as part of the Leeuwin Current Interdisciplinary Experiment in recorded LC speeds of 1.6 m/s off Cliffy Head (Cresswell and Peterson 1993). Data obtained aboard the RV Sprightly showed the geopotential anomaly dropped about 0.21 m from Fremantle to Cape Leeuwin, with a further 0.1-m drop at the south-west corner. Here, a large geopotential gradient near the continental shelf that was not in geostrophic balance accelerated the Leeuwin current (Thompson 1984; Godfrey and Ridgway 1985). The same results were reproduced in process-oriented modelling studies, which showed that a rapid change in the geopotential gradient increased the LC velocity at Cape Leeuwin and in the south-west corner (Batteen and Butler 1998; Batteen et al. 2007). The LC is usually associated with eddies and meanders (Pearce and Griffiths 1991; Cresswell 1996; Fang and Morrow 2003; Feng et al. 2005; Fieux et al. 2005; Rennie 2005; Meuleners et al. 2006). Satellite images of SST and altimeter data confirmed eddies were present in the study region which was confirmed through field measurements (Cresswell and Domingues, 2009). The LC s mixed barotropic and baroclinic instability usually forms large, long-lasting mesoscale eddies off Australia s west coast (Meuleners et al., 2008); however, it is believed that LC interacting with changes in the bathymetry and coastline orientation generates eddies, meanders, and offshoots along the south coast. Eddies and offshoots appear where the coastal topography changes direction abruptly, e.g. to the south-west of Capes Naturaliste and Leeuwin, in the south-west corner, to the south of Albany, and to the south of Esperance (Pattiaratchi 2006). Griffiths and Pearce (1985) suggested that baroclinic instability generated eddies in the south-west corner. The eddy offshoots strengthen the LC while narrowing it, with the resulting instability producing cyclone anticyclone pairs (Batteen et al. 2007). Sea surface temperature images of the Recherché archipelago showed the LC was often unstable in winter and sometimes had eddy offshoots; these offshoots were likely due to the abrupt change in the shelf edge orientation (Cresswell and Griffin 2004). 25

28 Figure 2.5. Schematic of surface and subsurface currents along the continental shelf and slope off WA (Pattiaratchi and Woo, 2009) Physical oceanography principles The current system along the study region encompasses unique principles of ocean physics. Geographically, the southeast Indian Ocean is analogue to the eastern boundary current regions in the Atlantic and Pacific (Benguela, Canary, Peru and California currents). Furthermore, the wind off WA is predominantly equatorward and one might expect to find broad equatorward flow and upwelling along the coast. However, the ocean off WA behaves quite unlike the other eastern boundary regions. The explanation for the anomalous behavior in this region seems to lie in the strength of the poleward pressure gradient. Thompson (1984) and Godfrey and Ridgway (1985) pointed out that the slope of the geopotential anomaly along the WA coast is very large and could overcome the equatorward wind stress. The pressure gradient are derived from the equations of motion assuming the flow has no acceleration, du/dt = dv/dt =dw/dt = 0 (u and v= horizontal velocity, w= vertical velocity 26

29 and t = time); that horizontal velocities are much larger than vertical, w «u, v; that the only external force is gravity; and that friction is small. With these assumptions; ; ; (1) Where f=2 is coriolis parameter. The equation can be written as; ;, (2) (3) where p 0 is atmospheric pressure at z = 0, and ζ is the height of the sea surface. Substitute (2) into (1) gives:, (4), (5) The Leeuwin undercurrent Thompson (1984) documented an undercurrent flowing beneath the Leeuwin current off North West Cape and Shark Bay, in opposite direction of the Leeuwin Current. This undercurrent was between 200 and 400-m water depth and transported 5 Sv of high salinity (35.8 psu), oxygen-rich, nutrient-depleted south Indian central water at a rate of m/s (Godfrey and Ridgway, 1985; Smith et al., 1991). (Batteen et al. 2007) suggested that eastward, subsurface-intensified jets dominating the subtropical gyre formed the LU s upper layer. In the south, the undercurrent was first measured aboard the RV Franklin in A cross-section of the region near Cape Leeuwin showed an undercurrent with speeds of 0.3 m/s centred at 450-m water depth; another cross-section of the region near Cliffy Head showed the undercurrent had speeds of 0.2 m/s (Cresswell and Peterson 1993). These data form the only recordings of the undercurrent in the south thus far. The LU has not been 27

30 studied much because of the small amount of field data available; however, several numerical modelling studies have reproduced the LU s mean flow characteristics. The LU has a maximum speed of 0.35 m/s and a mean speed of 0.1 m/s and consists of meanders and eddies, which the more energetic surface flow provides. Climatology has shown that a mean geopotential gradient of 1.9 x 10 7 acting towards the equator drives the LU (Meuleners et al. 2007; Woo et al., 2008). The undercurrent flows along the continental slope with its core between 500 and 1000-m depth and reaches maximum speeds at the slope. The LU maintains its speed as it flows north, but shallows and covers a much smaller cross-sectional area (Batteen et al. 2007). The LU in the southern region is continuous between the south and west coasts, but it also has a meridional flow, which starts from Cape Leeuwin and moves north. In the south coast, part of Flinders Current flows toward west is position at subsurface, close to the shelf slope and acting like Leeuwin Undercurrent in the west coast The Flinders current The Flinders current (FC), which is thought to be the only northern boundary current in the Southern Hemisphere, dominates the southern region (Bye, 1972; Middleton and Bye 2007). Bye (1972), using wide-ranging hydrographical observations, first recorded this boundary current along the south Australian coast. Observations and ocean model outputs of the current suggested it flowed along the shelf slope from Tasmania to Cape Leeuwin. Wind stress curl is the FC s driving force; positive, year-round measurements suggested it gave rise to an equatorward Sverdrup transport (at 135 E) and deflected the FC to the west (Figure 2.6). This dynamic profile satisfies a classical argument in vorticity dissipation and mass conservation (Cirano and Middleton, 2004; Middleton and Bye, 2007). This westward transport occurs between 37 and 39 S, has a transport range between 8 and 17 Sv, and is strongest in summer. The FC is centred within a permanent thermocline (at 600-m depth), with its core located adjacent to the continental slope at ~400-m depth. Upwelling occurs at and below this depth, whereas above this depth, wind forcing causes the downwelling. 28

31 Principles of Sverdrup balance is derived from Ekman transport, assumes spatially uniform steady wind with no pressure gradient. By differentiating Ekman equation with respect to y and x, to look at wind field: ] (5) ] (6) adding (5) and (6) yield: ; (7) Using as follows; 0, meridional transport of water resulted from the curl of the wind is meridional (8) (9) Satellite images of the surface circulation showed a cool FC at the LC s outer edge (Figure 2.7). The vertical profiles also showed the FC and LC interacting at the shelf break and slope. At the shelf break, part of the FC flowed beneath the eastward flowing LC, similar to the undercurrent observed on the west coast, which suggested the Flinders current fed the Leeuwin undercurrent (Church et al. 1989; Middleton and Cirano 2002). The seasonally averaged model results from Meuleners et al. s (2007) study suggested the FC continued east at the end of the zonal landmass and recirculated at Naturaliste Plateau; part of the current then joined the LU at the shelf slope. Our field data suggested part of the FC split with the main current at the corner and continued as the LU. This theory will be discussed throughout this thesis. The absence of detailed data for the Leeuwin current, Leeuwin undercurrent, and Flinders current means the link between these current systems in the study region remains unknown. 29

32 Figure 2.6. Depth average velocity from the numerical model of Middleton and Platov (2003) as driven by summertime mean winds. Sverdrup transport acting towards the land mass. Leeuwin Current FlindersCurrent Figure 2.7. Sea surface temperature image for 9 April 2006 showing the warm core of the LC flowing south towards the south-west corner. The cooler FC, which was flowing westward in the south, separated from the LC at the shelf slope. 30

33 2.2.5 Coastal currents Field data and satellite images were used in several studies to study the local circulation in the inner shelf region of the continental slope (an area covering < 50-m depth) along the west coast during the summer (Church et al. 1989; Cresswell and Peterson 1993; Pearce and Pattiaratchi 1997; Gersbach et al. 1999; Pearce and Pattiaratchi 1999; Pattiaratchi 2006; Woo et al. 2006). These studies described the cool, northward flowing Capes and Ningaloo currents, with the Leeuwin current located farther offshore. The coastal circulation dynamics in the southern region are largely unknown; however, Hazel, v (2001) observed upwelling of cold, deep water in the Recherché archipelago and adjacent waters. It has been proposed that wind-driven coastal currents, which move west with the south-easterly winds south of WA in the summer, similar to the Capes and Ningaloo currents, cause this upwelling. If so, the LC may move farther offshore in the summer. 2.3 Water mass characteristics The Leeuwin current (LC) has a warm (22 25 C), low salinity (< 35 ) surface water known as the tropical surface water (Table 1). This water is mainly derived from the remote equatorial Indian Ocean, via the south Java current, and the equatorial Pacific Ocean, via the Indonesian Throughflow (Woo et al., 2008). Geostrophic inflow from the West Australia Current bring cooler and high salinity water into the system and lower the temperature of the LC core and increase its salinity as the current moves south (Woo et al. 2006; 2008). Satellite images and field data from the south of WA showed the LC had two main types of water: salty, subtropical water from the west of WA and fresh, tropical water from the north of WA (Cresswell and Peterson 1993). (Andrews 1977) found that a branch of the km-wide of northward flowing West Australian, flowed west between about 29 and 32 S. This branch of water, which turned south, following the LC flow, consisted of salty water from the south Indian central water, entrained into the LC by geostrophic in flow. 31

34 The LC cools and becomes saltier in the south, but maintains warm from it surrounding a change termed the ageing process (Woo et al. 2006). With the autumn to winter flow increase, the LC can be distinguished by the warm, tropical water flowing along the shelf break, especially during May and June. However, the LC drops ~8 C as it flows from the north to the GAB during autumn and winter (Ridgway and Condie 2004). Deep winter convection at S in the zone between the subtropical and subantarctic fronts to the south of Australia forms SAMW (Woo et al. 2008). Comparatively, the SAMW in this region is one of the densest and only the far eastern origin has the same profile (Wong 2005). The SAMW s high oxygen content also ventilates the lower thermocline of the Southern Hemisphere subtropical gyre. Thompson and Edwards (1981) found SAMW with a temperature of about 8.6 C and a salinity of about forming to the south of Tasmania along 145 E. Middleton and Cirano (2002) proposed the FC then transported the SAMW formed in this region to the west, where the LU continued transporting it north. Water mass Temperature Salinity range Dissolved oxygen range ( C) range (μm/l) Tropical surface water (TSW) South Indian central water (SICW) Subantarctic mode water (SAMW) Antarctic intermediate water (AAIW) North-west Indian intermediate water (NWIIW) ~ Table 2.1. The characteristics of the water masses found in the 1-km-deep water column defined. The table combines data from voyages FR10/2000 ( S) and SS09/2003 ( S) (Woo et al. 2008). 32

35 2.4 The regional ocean modelling system (ROMS) Study of continental shelf slope circulation began some decades ago with simple and straight forward analytical models. The model basically consists of square coastline and uniform bathymetry. Later, models include a real coastline and topography with more accurate climatological and forcing data. The models grew in complexity, not only in topography used, but also the complexity in describing the flow. Options included variations in the boundary condition, forcing, the stratification (varying density with depth) and inclusion of non-linear terms in the equation of motion. The development of computers allowed numerical models to be implemented more successfully, growing in complexity and realistic simulation. Regional Ocean Model System (ROMS) will be used. Initially, it was based on the S- Coordinate Rutgers University Model (SCRUM) described by Song and Haidvogel, (1994). It is a free-surface; hydrostatic, primitive equation ocean model that uses stretched, terrain-following coordinates in the vertical and orthogonal curvilinear coordinates in horizontal. In the vertical, the primitive equations are discretized over variable topography using stretched terrain-following coordinates (Song and Haidvogel, 1994). The stretched coordinates allow increased resolution in areas of interest. In the horizontal, the primitive equations are evaluated using boundary fitted, orthogonal curvilinear coordinates, which the general formulation of the curvilinear coordinates includes both constant metrics and variable metrics coordinates. The equations for momentum are solved using a split-explicit time-stepping scheme. ROMS has been applied to variety of situations that indicate its usefulness in the present study. The circulation off the west coast of Iberia is similar to Western Australia. There is poleward boundary current with a weaker equatorward undercurrent (Peliz et al. 2002). The dynamics suggest that the poleward current is generated from pressure gradient. The Iberian current is seasonal, occurring in winter, as in summer, strong equatorward wind dissipate and push it offshore. A simple model of this current shows that it acts similar to Leeuwin current, with upwelling wind causing modification of the current with a shelf current forming in the reverse direction (Peliz et al. 2003). 33

36 In other examples, the circulation of Leeuwin Current in northern section of west coast of Australia where the Leeuwin Current and its associated eddies have been successfully modelled using ROMS (Meuleners et al. 2007; 2008). The model applied the pressure gradient and a wind stress term with real bathymetry. It described a numerical experiment of the mean flow and eddy characteristics using an annually averaged boundary forcing condition. The model was able to mimic the spatial, temporal and migratory scales of the LC s system. Eddy generation sites were consistent with those observed from satellite imagery. Rennie et al. (2007) applied ROMS to the LC system to examine the current behaviour on Perth submarine canyon. The model manages to explore the water columns response to climatological forcing and the meandering current, including upwelling and downwelling effects. Comparison with available hydrographical data demonstrates that there were some differences between simulation and field data, however in overall the model is reliable. 2.5 Conclusion Knowledge of the circulation off south-west Australia is limited because of the lack of field data from the region. The LC s seasonal strength has been documented, but less is known of the seasonal undercurrent strength in the west and south, as most of the field studies were conducted along the west coast and focused on only the surface current. We used field data and numerical models to examine the LC and its instability in the southern region. Our findings will enhance the current knowledge of the LC in the study region. 34

37 Chapter 3 The Leeuwin current along the south coast of Western Australian during autumn Introduction The circulation off Western Australia is anomalous compared with other western continental margins (Godfrey and Ridgway 1985; Smith et al. 1991; Pearce 1991; Ridgway and Condie 2004). This is mainly because of the Leeuwin current (LC), which transports warm, low salinity, nutrient-depleted water of tropical origin from the northwest of Australia, south along the Western Australian coast and into the Great Australian Bight (GAB). The current, which is shallow (< 300 m), narrow (< 100 km), and strong with velocity reaching above 1m/s in winter, controls the region s climate and marine life (Maxwell and Cresswell 1981). An alongshore pressure gradient, which overwhelms the opposing wind stress, especially along the west coast, drives the Leeuwin current (Thompson 1984, 1987; Godfrey and Ridgway 1985; Weaver and Middleton 1989; Batteen and Rutherford 1990; Godfrey and Weaver 1991; Pattiaratchi and Buchan 1991). This phenomenon causes downwelling along the coast. The Leeuwin undercurrent (LU), which flows northwards beneath the LC (Thompson 1984; Thompson 1987; Smith et al. 1991) and is located between the 250 and 450-m depth contours, is associated with subantarctic mode water and is characterised by dissolved oxygen (Pattiaratchi 2006; Woo and Pattiaratchi 2008). Cresswell and Golding (1980) found that the Leeuwin current accelerated around Cape Leeuwin and then flowed eastward at speeds up to 1.8 ms -1 towards the Great Australian Bight (GAB), meandering on and off the continental shelf. The LC off Cape Leeuwin 1 This chapter is to be submitted as a journal paper to Continental Shelf Research 35

38 consists of a combination of lower salinity water of tropical origin and subtropical high salinity waters from the west, which arrive in the austral autumn and flow into the lower salinity subantarctic waters south of Western Australia (Cresswell and Domingues, 2009). Using 5-year monthly mean numerical model output from the POP ocean model, Cresswell and Domingues (2009) showed that from December to March, the Leeuwin current was weak and then strengthened during April to evolve into a vigorous current during the austral winter. They also indicated that the eastward flow of the LC was initially stronger only in the western section but by July/August it accelerated into the GAB. Although the LC is a continuous system along the west and south coasts, the dynamics differ due to the seasonality and changes in the wind stress. During winter, when the LC is strongest, the equatorward wind stress is weak, allowing the LC to accelerate; however, along the south coast the wind stress is a maximum and is in the same direction as the LC. Godfrey and Vaudrey (1985) and Ridgway and Condie (2004) found that although the alongshore pressure gradient was lower along the south coast (compared with that along the west coast), north-westerly winds (i.e. eastward component of stress) enhanced the LC. In summer, the westward wind stress due to the south-easterly winds weakens the current. The temperature and salinity characteristics of the LC are different along the south coast compared with that along the west coast, particularly during summer and autumn. Geostrophic inflow into the LC entrains higher salinity south Indian central water into the LC. This occurs mainly between 29 and 34 S (Hamilton 1986; Cresswell and Peterson 1993; Pattiaratchi and Woo, 2009) and thus the LC becomes more saline along the south coast, but still maintains its warmer signature. Along the south coast of Australia, the Flinders current (FC), a northern boundary current flowing from east to west, dominates the circulation pattern (Middleton and Bye, 2007). A positive wind stress curl causes an equatorward Sverdrup transport (across 135 E), which is deflected to the west because of the presence of the Australian continent. This dynamic profile satisfies a classical argument in vorticity dissipation and mass conservation (Middleton and Cirano 2002). The westward transport is between latitudes 37 and 39 S and has flow rates between 8 and 17 Sv, with higher flows during the austral summer. The FC is centred within a permanent thermocline (600-m depth) with a subsurface maximum around 400-m depth, adjacent to the continental slope. The FC 36

39 interacts with the LC at the shelf break and slope, where it flows westward beneath the eastward flowing LC, similar to the undercurrent observed along the west coast. This behaviour suggests the FC likely feeds the Leeuwin undercurrent (Church et al. 1989). The major features of the LC are eddies, meanders, and offshoots (Pearce and Griffiths 1991; Fang and Morrow 2003; Morrow et al. 2003; Feng et al. 2005; Fieux et al. 2005; Meuleners et al. 2006; Rennie et al. 2006). These eddies, which form at the shelf break and usually break away, can be seen in sea surface temperature imagery and altimetry data (Griffin et al. 2001; Fang and Morrow 2003). The LC interacting with changes in the coastline orientation and a combination of barotropic and baroclinic instabilities are responsible for the generation of these eddies (Meuleners et al. 2008). Along the south coast, the generation of eddies is associated with bathymetry and coastline orientation changes (Figure 3.1). The main eddy regions are at the south-west of Capes Naturaliste and Leeuwin (34.5 S, E), south of Albany (35.5 S, 118 E), and south of Esperance (35 S, 123 S) (Pattiaratchi 2006; Cresswell and Domingues, 2009). Although many aspects of the LC have been studied over the past 20 years, little is known about the dynamics of the LC along the south coast of Western Australia (Cresswell and Domingues, 2009). The presence of the FC as a main current system in the southern region is also important; however, until now, the only known study of the FC was Cirano and Middleton s (2004) numerical modelling study. This chapter discusses the CTD and acoustic Doppler current profiler (ADCP) data (collected in autumn), meteorological data, and satellite images used to examine the surface and subsurface circulation along the south coast of Western Australia, focusing on the continental shelf and slope current systems. The dynamics of the current system and changes in the water mass characteristics associated with the LC ageing process are also presented. 37

40 3.2 Methodology and study domain Data were obtained using the RV Southern Surveyor (voyage no. 04/2006) off the south coast of Western Australia, over the period 12 April to 1 May 2006 (20 days). The voyage departed from Esperance and ended in Fremantle (Figure 3.1). Seventeen cross-shelf transects (Figure 3.1) were undertaken with CTD stations being occupied along each transect, depending on the shelf width. The cross-shore transect extended from the coast (~50-m depth) to the 2000-m contour. As with the earlier cruises (Woo et al. 2006), stations were placed at 50-m-depth intervals in the shallow region and increase as approaching shelf slope (50m, 100m, 150m, 200m, 250m, 300m, 400m, 500m, 750m, 1000m, 1250m, 1500m). In average, time taken to complete a transect and approaching another transect is a day. The stations placed on the continental slope is closer to each other given the slope is relatively steep especially between transect K - P. The distance between transect range between km. In addition to the standard CTD data, shipborne ADCP (Teledyne RDI 150 khz with a range to 800m) and lowered acoustic Doppler current profiler (LADCP) data were also obtained to enable the current structure to be mapped. Surface temperature, salinity, and wind data, covering the whole cruise track, were also collected. The field data were supplemented by altimeter and sea surface temperature satellite images obtained during the voyage. The altimeter images were available from the CSIRO Marine and Atmospheric Research (CMAR) website ( The images were used for identifying the location of eddies and offshoots of the Leeuwin current. 38

41 O Fremantle Cape Leeuwin o Esperance o Albany o Figure 3.1. Location map of the study area including the Southern Surveyor 04/06 transects through south of Western Australia (WA) Study region The bathymetry of the southern Western Australian continental shelf is narrow, especially between transects N and D, where the 200-m isobath extends to < 20 km from the shore (Figure 3.1). It stretches farther offshore in the region offshore of Esperance and the Recherché Archipelago. The depth contours are usually parallel to the coastline and typical of the western shelf. Beyond the continental shelf, the depth increases rapidly to 4000 m. The study will focus on the current systems at the coastal margin, extending from the Capes region (34 S, 115 E) to the western end of the Great Australian Bight (GAB) (32.5 S, 126 E). This region will be defined as south of Western Australia or the southern region. The GAB starts at 124 E, the area which the width of the continental shelf starts to increase. 39

42 3.3 Results and discussion Surface characteristics Surface temperature, salinity and velocity (from shipborne ADCP) were used to define changes in the Leeuwin current (LC) properties along the southern coast, in particular, the entrainment of water into the LC from offshore. The surface temperature and salinity data indicated that the LC in the Capes region consisted of a water temperature of 22 C and a salinity of 35.7 and entered the GAB with a temperature of 18.5 C and a salinity of 35.8, i.e. the LC became cooler and slightly saline through its passage along the south coast. There were two patches of lower temperature and salinity water (Figure 3.2): (1) offshore near Albany (118 E) and (2) south west of Esperance (~121 E). These patches of water were associated with locations of submarine canyons and were most likely due to localised upwelling (see Section 3.5). In the Great Australian Bight (GAB), the waters inshore of the LC were warmer (~21 C) and more saline (~36.6) than the LC water. The wide, shallow topography of the GAB in combination with a high heat input and evaporation results in a higher temperature and salinity on the GAB shelf (Herzfeld, 1997). Figure 3.2. Sea surface temperature (SST) distribution along cruise track. 40

43 Figure 3.3. Surface salinity distribution along cruise track. Figure 3.4. Surface current distribution obtained from the shipborne ADCP data. Surface currents recorded using the shipborne ADCP indicated that the maximum currents, associated with the Leeuwin current, were located along the shelf break (Figure 3.4). Similar observations were made by Woo et al. (2006) in the northern section of the Leeuwin current. Off the Capes region (transect P), the maximum surface velocity was relatively weak at ~0.7 m/s. An eddy, which formed off Cape Leeuwin (Figure 3.5), was thought to extend the LC further offshore, decreasing the speed of the current in the shelf/slope region. However, the LC strengthened (maximum speed of 1.2 m/s) as it 41

44 moved past Cape Leeuwin. Godfrey et al. (1985), using field observations, and Batteen and Butler (1998), using model predictions, reported that the maximum Leeuwin current speeds were found in this region (see also Section 3.3.4). The maximum surface currents decreased along the south coast (~0.9 m/s) and decreased further upon reaching the GAB (Figure 3.4). A similar pattern was reported by Cresswell and Domingues (2009). Figure 3.5. Altimeter image showing sea level anomaly obtained on 1 May Locations of offshore eddies are shown (Source: CSIRO) Subsurface characteristics Transects P, K, G, and A (Figure 3.1) were chosen to examine the sub-surface characteristics between the Capes region and the GAB. These transects represent conditions along the two extremes of the study region (Transects P and A) and in the central section (transects K and G). The data represent autumn conditions, when the LC begins to strengthen along the south-west coast. Previous studies have shown that the LC is strongest as it reaches the south-west in May/June (Ridgway and Condie 2004; Feng et al., 2005). The LC transports warmer, lower salinity water, originating from the tropics, into the study region. At transect P, the LC may be identified through the combination of higher temperature (> 21 C), lower salinity (< 35.7) water in the surface 100 m, which was also associated with the southward currents (Figure 3.6) (positive velocity values refer to 42

45 southward, and negative values refer to northward). Below this layer, a higher salinity water mass, the South Indian Central Water (SICW), with a salinity up to 35.9, can be identified. This water mass also flows southward to a depth of 300 m (Figure 3.6). A similar structure was identified by Woo et. al. (2006), where the Leeuwin current was defined as a southward flowing current extending to 300-m depth and consisting of two water masses: (1) lower salinity tropical water and (2) higher salinity SICW. At Transect K (Figure 3.7) the temperature of the LC core had decreased by 1 C, compared with that at Transect P. However, the most striking feature is the absence of the subsurface salinity maximum (the surface salinity associated with the LC core has a salinity 35.8, which was present as a subsurface maximum at Transect P). Hence, between Transect P and K, the LC lost its lower salinity signature. The ADCP data indicated that the core of the LC flowed along the continental slope, with maximum (eastward) currents at the surface exceeding 0.80 ms -1 and extending to a maximum depth of 300 m (positive velocity values refer to eastward, and negative values refer to westward).. If we define the LC as being the V -shaped section shown by the ADCP data (Figure 3.7c), then the LC in this region has a range of temperatures (12 19 C ) and salinities ( ) Through advection, the SICW is believed to mix gradually along its way with the much cooler and lower salinity water of the southern region. The higher velocity of the undercurrent, reaching speeds up to 0.5 m s -1, is the highest value recorded in the region (Figure 3.7c). The LC is much narrower at transect K which only extend up to 50km away from coast; this change is caused by the strong FC offshore and at the subsurface. However, the LC still extends up to 300-m depth. This suggests that the lower salinity and temperature of the LC at the subsurface is actually part of the LC. Transect G (Figure 3.8) showed a decrease in the surface mixed layer. The temperatures dropped 1.5 C from transect P. These results showed the LC core cooled as the LC moved eastward. However, both the temperature (Figure 3.8a) and salinity (Figure 3.8b) transects indicated downwelling along the shelf break. The temperature and salinity offshore were lower compared with that at the shelf break. The ADCP data (Figure 3.8c) showed that the LC core was narrower than transect K. This indicated that the offshore extent of the LC mixing with the FC resulted in a lower temperature and salinity. The 43

46 shear between the LC and FC at the surface can be seen from Figure 2.7, obtained just prior to the field data collection. Figure 3.6. Transect P: Cross section of (a) temperature, (b) salinity and (c) alongshore velocities. At transect A (Figure 3.9), the presence of higher temperatures and salinity values on the GAB shelf is apparent. The salinity values reach ~36.2 and the difference between the shelf and the slope is ~0.6 (Figure 3.9b). This is reflected in the temperature data, where warmer water was observed on the shelf and decreased offshore (Figure 3.9a). The wider 44

47 and shallower continental shelf in this region, combined with the high heat input, results in an increase in both the temperature and salinity (Herzfeld, 1997). Ridgway and Condie (2004) also reported that an increase up to 2 3 C can occur along the GAB shelf. The widening of the continental shelf also decreased the speed of the LC by making the LC shallower and wider (Figure 3.9c). Woo et al. (2006) also found that the LC velocity decreased when the continental shelf widened. Figure 3.7. Transect K: Cross-section of (a) temperature, (b) salinity and (c) alongshore velocities. 45

48 Figure 3.8. Transect G: Cross-section of (a) temperature, (b) salinity and (c) alongshore velocities. An alongshore transect, corresponding with the core of the LC, of the velocity distribution with depth was constructed using the ADCP data. Here, the maximum surface velocity along a CTD transect was used to obtain the depth distribution of the velocity (Figure 3.10). The data indicated that the LC speed varied through its passage along the south coast. At transects P, M, and J the LC was weaker, whereas at transects N, L and F the LC was stronger (Figure 3.10). Transects P, M and J were associated with offshore eddies (Figure 3.5) and their influence is reflected in the LC strength. The satellite images suggested the offshoots occurred offshore of these transects, but their influence was reflected in the LC. The interaction between the LC, LC eddies and the Flinders current is 46

49 highlighted by the data. When the LC was strong (e.g. at N, L and F), the depth of the LC was located around 250-m depth. In contrast, in the vicinity of the eddies (at P, M and J), the LC extended to > 400 m (Figure 3.10). Farther east, the LC weakened as it entered the GAB (transect D) and crossed the wide section of the continental shelf. The ADCP data were analysed to determine the along shore and cross-shore transports associated with the Leeuwin current. The results indicated that the transport into the study region (across transect P) and out of the region (transect A) were similar (1.1 and 1.0 Sv, respectively). But the influence of the eddies was highlighted in this region, as there was an inflow of 3.52 Sv and an equal amount of offshore transport from the region due to eddy activity. Figure 3.9. Transect A: Cross-section of (a) temperature, (b) salinity and (c) alongshore velocities. 47

50 Figure Velocity profile along the line of maximum of LC flow. The dashed line indicates 0 m/s Water mass characteristics The Leeuwin current initially contained a higher temperature and a lower salinity, but this changed when it entered the southern region. This was mainly due to the entrainment of South Indian Central Water (SICW) into the Leeuwin current along the west coast. The SICW, which consists of higher salinity and a lower temperature, accelerated the LC s ageing process, cooling the LC and increasing its salinity. In its journey along the west coast, the LC water characteristics changed from T = 24 C and S = 34.8 (Woo et al., 2006) in the north to T = 21.5 C and S = 35.7 off Cape Leeuwin; they then changed to T = 19.5 C and S = 35.8 when entering the GAB. Note that only a small variation in salinity was recorded along the south coast. The water mass characteristics of the LC along the 200-m isobath are shown in Figure 3.11 and indicated the different surface water masses as well as the transformation of the LC characteristics along the south coast. In the Capes region, the LC has warmer, lower salinity water as well as higher salinity water originating from South Indian Central Water (SICW). At transect K the waters were cooler and less saline through mixing with offshore waters. Transect G had even cooler and lower salinity water, resulting from upwelling through the head of the submarine canyon. The water at transect A was influenced by the GAB, with a higher temperature and salinity. For complete T/S diagram for this figure, please refer to Figure

51 Figure Temperature/salinity diagram of the surface (above 200m) stations at transects R, K, G and A (Figure 3.1) Alongshore geopotential gradient The alongshore geopotential gradient, the main driving force of the LC, was analysed using data from the 1000-m depth contour stations at each transect. The results indicated a mean surface slope of 1.0 x 10-7 relative to the 300-dB level and 0.5 x 10-7 at the 300/750-db interval (Table 3.1; Figure 3.12). The alongshore geopotential gradient relative to the 300 db was smaller than the previous measurements along the west coast (Table 3.1). For example, Woo and Pattiaratchi (2008) estimated a sea surface slope of 4.0 x 10-7 at the 1000-m isobath at the db level for the northern region of WA. The mean values recorded here were comparable to the 2 x 10-7 value reported by Vaudrey and Godfrey (1985) along the south coast. Closer examination of the alongshore geopotential gradient (Figure 3.12a) indicated that there was a large spatial variability, with the biggest change, 1.7 x 10-7, occurring in the initial 400 km (Cape Leeuwin to 49

52 Point D Entrecasteaux; Figure 3.1), with only a smaller change along the remainder of the transects along the south coast. This explains the high acceleration and strong currents reported in this region. Griffiths and Pearce (1985), using high-resolution infrared images from NOAA-7, found that the maximum currents (~1.8 m/s) along Australia s south coast occurred near Cape Leeuwin. Several numerical models have reproduced the same behaviour (Weaver and Middleton 1989; Batteen and Butler 1998). Thus, the sharp density changes (due mainly to salinity) between the west and south coasts in the vicinity of Cape Leeuwin (Section 3.3.3) resulted in acceleration and strong currents in this region. This was also reflected by the maximum surface currents (1.2 m/s) measured by the shipborne ADCP (Figure 3.4) in this region. Author 0/300 db 300/750 db Geopotential gradient (m 2 /s 2 ) x 10-7 Geopotential gradient (m 2 /s 2 ) x 10-7 Hamon, Thompson, Thompson, Smith et al., Woo and Pattiaratchi This study Table 3.1. Summary of estimated alongshore geopotential gradients at the surface relative to the 300dB (west coast) and subsurface at 300/750dB. The subsurface slope of 0.5 x 10-7 at the 300/750-dB interval (Table 3.1; Figure 3.12) is comparable to other studies reported from the region and reflects the effects of the Flinders current. Although the Flinders current is driven by the wind stress curl, the density changes also appeared to act as a driving force. 50

53 a) b) Figure Geopotential anomaly at 1000-m depth (a) above 300 m (b) between 300m and 750m Influence of wind In the southern region, mariners have long believed that the observed current is a result of wind forcing (British Admiralty, 1937). Observations by Newell (1961) indicated that the current along the eastern portion of the southeast coast was correlated with the wind direction. Godfrey et al. (1985) suggested that prevailing westerlies and possibly an independent thermohaline effect drove the eastward current. From the wind data collected onboard the vessel, the existence of a west-southerly wind was clear (Figure 3.13). Previous studies have suggested that at this time of the year, the wind pattern switches to the winter wind pattern, which is favourable to the slope current (Ridgway and Condie, 2004). 51

54 Figure Wind speed onboard the Southern Surveyor 04/06 for transects K and G. The wind speeds measured through the period of the voyage ranged between about 5 to 7 m/s. These wind speeds were relatively low and too weak to be the main forcing of the LC. However, together with the alongshore geopotential gradient analysis, they could still act as an additional forcing mechanism along the south coast. This is investigated further using numerical modelling (Chapter 5) Submarine canyon upwelling The surface distribution of temperature and salinity (Figures 3.2 and 3.3) showed localised patches of colder, lower salinity water, which were associated with submarine canyons along transects G and K (named Bremer Canyon and Albany Canyon, respectively) (Figure 3.14). These submarine canyons are located along the continental slope in water depths of > 500 m and therefore do not interact with the Leeuwin current. The temperature and salinity difference at the heads of the two canyons, when compared with in situ water, was ~2 C and 0.5 respectively, with a dissolved oxygen concentration of 240 μm/l (Figure 3.15). An alongshore temperature and salinity transect along the 52

55 1000-m depth contour (Figure 3.15) clearly showed the upwelling signal at transects K and G from depths > 200 m. Kaempf (2007) used idealised numerical experiments to demonstrate upwelling due to interaction between a deep, westward flowing current (e.g. Flinders current) and a submarine canyon. The westward flowing current formed an anticlockwise eddy within the canyon and, in combination with the canyon s funnelling effect, caused upwelling from below 300-m depth onto the continental shelf. Kaempf (2007) postulated that this mechanism may be applicable to upwelling off South Australia, where the Flinders current interacting with submarine canyons resulted in upwelling. Thus a similar situation may be occurring at Bremer and Albany canyons. Rennie et al. (2008) also showed that interaction between the Leeuwin undercurrent and the Perth canyon offshore Fremantle along the west coast resulted in upwelling. Figure Section of submarine canyon topography in the south coast (from Geoscience Australia). The upwelling was observed at Albany Canyon and Bremer Canyon. 53

56 (a) (b) (c) Figure Alongshore transects along the 1000-m contour: (a) temperature, (b) salinity and (c) dissolved oxygen. 54

57 3.4 Concluding remarks The oceanographic circulation off Western Australia s south coast is unique and complex. The Leeuwin current flows continuously from its origin in the North West Shelf southwards, and at Cape Leeuwin turns eastward and flows onto the GAB. An undercurrent was present, flowing on both the west and south coasts (Flinders current and Leeuwin undercurrent; Figure 3.16) in an opposite direction and beneath the LC. The Leeuwin undercurrent and Flinders current are distinguished by the dynamics of the system and the water mass distribution. It was apparent that the Flinders current along the south coast was a surface current driven by the wind stress curl (Middleton and Bye (2007). At the continental slope, the LC flows above the FC and thus in this region, the FC acts as an undercurrent. The Flinders current acts as a feeder to the Leeuwin undercurrent, which exists along the west coast. Figure Schematic diagram of the current system off Western Australia. Field data obtained as part of this study indicated that the warmer, lower salinity LC underwent a transformation as it neared the southern coast. Cooler and higher salinity south Indian central water was entrained from offshore (from the West Australian 55

58 current), cooling the LC and increasing its salinity. This enhanced the geopotential gradient in the vicinity of Cape Leeuwin, causing it to accelerate. The LC speeds measured in this region were higher than the LC speeds measured along the west coast because of the steep slope and narrower shelf in the south. Eddies also influenced the LC speed and transport. The inflow and outflow was ~1 Sv as the LC entered and exited the study region. However, a cross-shore exchange of ~3.5 Sv was almost entirely due to eddies. The Flinders current was observed as a surface current offshore and an undercurrent flowing in the opposite direction beneath the LC. In addition to the wind stress curl (Sverdrup balance) driving the current (Middleton and Cirano, 2002), it is likely that the alongshore geopotential gradient was also a contributor (Figure 3.11b). The field data indicated that the core of the current was located between 400 and 500-m depth, which agreed with the model results of Middleton and Cirano (2002). The strongest part of the current was generally located as the undercurrent beneath the LC and can be seen in transects P, K, and G (Figures 3.6 to 3.8). The FC was generally stronger (in terms of transport) than the LU, with maximum speeds of 0.6 m/s (transect L). In contrast, the LU along the west coast had a maximum velocity of only ~0.2 m/s (Woo et al. 2006), whereas Middleton and Cirano s (2002) modelling results predicted a maximum speed of only 0.16 m/s for the FC. The temperature/salinity characteristics for the FC indicated that the current s water mass was part of the sub Antarctic mode water (SAMW); this was also the same water mass that the LU transported north along the west coast. The undercurrent system beneath the LC is responsible for transporting cooler ( ºC), low salinity ( ) SAMW westward along the south coast and northward along the west coast. However, the interconnection between these two current systems is still unclear, although some earlier studies postulated that the FC fed the LU. The next chapter will examine this interconnection in detail. Upwelling in the submarine canyon on the south coast appeared to be energetic. Our findings showed that the upwelling brought the water from 200-m depth to the surface and indicated that the submarine canyons along the south coast can have a major influence on the physical and biological processes in the region. 56

59 Chapter 4 The interconnection between the Flinders current and Leeuwin undercurrent Introduction Two major boundary currents flow along the south and west coasts of Western Australia: the Flinders current (FC), which forms off the coast of Tasmania, and the Leeuwin current (LC), which originates from the North-West Shelf region of Australia. The Leeuwin undercurrent flows beneath the LC along the west coast, and the Flinders current flows beneath the LC along the south coast (Chapter 3). It has been proposed that part of the FC joins the LU (Church et al. 1989; Middleton and Cirano 2002; Middleton and Bye 2006; Cresswell and Domingues, 2009); however, no clear evidence of this is available from either field or numerical model outputs. In this chapter, data obtained from the Southern Surveyor 04/2006 research cruise, in which detailed transects for the study region were conducted, will be analysed to examine the interconnection between the two current systems, the FC and LU. A schematic illustration of the circulation south of Western Australia is shown in Figure The Flinders current, which is the main current in the region, is the only northern boundary current in the southern hemisphere (Bye 1972). A positive wind stress curl drives the FC, and the Sverdrup balance causes the FC to transport water, centered along 135º E, northward; this water is then deflected to the west, satisfying the vorticity dissipation and mass conservation (Middleton and Cirano 2002). The FC flows continuously from east to west, but transports more water during the summer. It occupies almost the whole water column to a maximum depth of 800 m. The FC interacts with the 2 This chapter will be submitted as a journal paper to Deep Sea Research I 57

60 Leeuwin current at the shelf break and slope; during this interaction, part of the FC acts as an undercurrent (Chapter 3). A geopotential gradient located at the depth of the Leeuwin undercurrent (LU) drives the undercurrent. Field studies conducted on the northern shelf suggested an equatorward undercurrent flowed beneath the LC (Thompson 1984; Thompson 1987). The Leeuwin Current Interdisciplinary Experiment (LUCIE) (Smith et al. 1991) described the LU profile as narrow, equatorward, and located adjacent to the continental slope between the 250 and 450-m depth contours. The LU, which is usually stronger in summer, transports (northward) 5 Sv of high salinity (35.8), oxygen-rich, nutrient-depleted south Indian central water at a rate of m/s (Thompson 1984; Smith et al. 1991; Pattiaratchi and Woo, 2009). The water mass characteristics suggest that the LU is closely related to the sub Antarctic mode water (SAMW), as shown by Woo and Pattiaratchi (2008). SAMW is formed from convection in the region south of Australia and thus the higher dissolved oxygen concentration in the water mass reflects the relatively young age of the water mass; hence the LU core can be identified from the dissolved oxygen distribution. Little is known of the circulation along the shelf and slope in the south-west corner of Western Australia. Interaction between two undercurrents, such as that between the FC and LU, is rare. Only the Pacific equatorial undercurrent and the Peru Chile undercurrent interact in the same way (Lukas 1986). As the Pacific equatorial undercurrent nears the west coast of South America and turns east to join the Peru Chile undercurrent, it supplies the Peru Chile undercurrent with nutrient-rich water (Lukas 1986). 58

61 4.2. Study region, data and methods The focus of this chapter is the current system along the Western Australian coast, from Cape Naturaliste on the west coast to Albany on the south coast (Figure 4.1), with the aim of examining water properties and currents in this region to identify the interconnection between the Flinders current and the Leeuwin undercurrent. Vertical profiles of temperature, salinity, dissolved oxygen and velocity were obtained from the Southern Surveyor 04/2006 research cruise. Measurements were obtained at several transects; five transects (transects R, P, N, L and K) were chosen for analysis (Figure 4.1). As explained in Chapter 3, each transect contains between 10 and 15 CTD stations, depending on the shelf width. The cross-shore transects extended from the coast (~50-m depth) to the 1000-m contour, and stations were located at 50-m depth intervals. Figure 4.1. The study region showing the location of the CTD transects and bathymetry contours. The regions referred to in the text are also shown. In this paper, unless stated otherwise, the Flinders current is referred to as the part of the FC that flows beneath the LC, whereas the Leeuwin undercurrent is defined as the northward current along the west coast starting from Cape Leeuwin. Both the FC and LU flow beneath the LC and in the opposite direction to the LC (Figure 3.16). 59

62 4.3. Water mass structure The FC and LU water mass structure The Flinders current and Leeuwin undercurrent are both associated with the sub Antarctic mode water (SAMW); thus it is essential to define the SAMW structure in terms of its origins and pathways with respect to the study region. SAMW is formed in the zone between the subtropical convergence and the subantarctic front to the south of Australia (40 50 S) through deep convection during winter (McCartney, 1977; Wong, 2005, Woo and Pattiaratchi, 2008). It has a temperature range of C and a salinity range of ; its σ T value ranges between 28.9 and It is also characterised by high dissolved oxygen concentrations of µm/l and a core depth occurring at m. Thompson and Edwards (1981) observed SAMW with temperatures close to 8.6 ºC and a salinity of about south of Tasmania along 145º E. Middleton and Cirano (2002) proposed the FC transported the SAMW westward and the LU continued transporting the SAMW northward. The LU along the west coast is characterized by this water mass (Pattiaratchi and Woo, 2009). Data from five CTD transects along the 1000-m isobaths were selected from the Southern Surveyor 04/2006 research cruise to determine the hydrographic structure over the continental shelf and slope. This section covered the region from Albany on the south coast to Cape Leeuwin on the west coast (Figure 4.1). The temperature and salinity diagram for each 1000-m profile are shown in Figure 4.2, and the temperature and salinity alongshore cross-sections are shown in Figure 4.3. The T S diagram (Figure 4.2) shows the upper mixed layer, which is part of the LC water mass, changed from the Capes region through to the south coast. The Capes region water (transects R and P) had a higher temperature and salinity, whereas along the south coast (transects L and K), the waters were cooler and less saline (Figure 4.2). Transect N shows intermediate conditions between the Capes region and the south coast. The changes to the TS characteristics resulted from changes to the LC structure as it entrained cooler, lower salinity water from offshore (Chapter 3). The T S diagram shows that the water at all the transects began to share the same TS characteristics at around T = 12 C and S = This was consistent with the TS characteristics of SAMW, which sits between a 60

63 temperature of C and a salinity of The bottom section of the TS diagram also showed that the water mass changed when reaching 8 C and a salinity of The window of SAMW T S properties fitted perfectly with the SAMW found along the west coast (Woo and Pattiaratchi 2008). This confirmed that SAMW was present in both the FC and LU. The alongshore transect along the 1000-m isobath (Figure 4.3) showed the continuity of the water masses. The LC lay in the top (~300-m) depths, in agreement with the LC along the west coast, as shown by Woo and Pattiaratchi (2008). This was shown clearly by the transition to the SAMW characteristics (T = 12 C; S = 35.2, DO > 240 μm/l) at ~ 300-m depth. Thus we concluded that there was a continuation in terms of water mass composition and it was likely that the Flinders current fed the Leeuwin undercurrent. Figure 4.2. T S diagram of the transects representing the south (transects L and K), west (transects R and P), and south-west corner (transect N) of Western Australia. 61

64 (a) (b) (c) Figure 4.3. Alongshore distributions of (a) temperature, (b) salinity and (c) dissolved oxygen along the 1000-m isobath for transects R, P, N, M, L and K (the unit of DO is μm/l) Oxygen maximum 62

65 A main characteristic of SAMW is its high dissolved oxygen concentration. This is due to the formation region of the SAMW being close to the study region. The presence of high dissolved oxygen was clearly visible from the alongshore transects (Figure 4.3c) and the cross-shelf transects at K and R (Figure 4.6). Transect K (Figure 4.6a) showed an oxygen maximum of 260 μm/l at around 450-m water depths extending offshore, whereas at transect R (Figure 4.6b), the oxygen maximum was slightly lower at 250 μm/l and a decrease in the core size was also noticeable. Similar values of the high dissolved oxygen core associated with the LU were reported by Woo and Pattiaratchi (2008). 63

66 (a) (b) Figure 4.4. Dissolved oxygen concentration across (a) transect K and (b) transect R. 64

67 4.4. Circulation Previous studies of the Leeuwin undercurrent and Flinders current have included field approaches (Godfrey et al. 1985; Cresswell and Peterson 1993; Woo and Pattiaratchi, 2008) and numerical models (Batteen and Butler 1998; Middleton and Cirano 2002; Meuleners et al. 2008; Domingues et al. 2007). From the research cruise data, five transects (K, L, N, P, and R) of the velocity distribution at the corner of Western Australia were examined. The transects covered both the FC and LC as the currents flowed along the south-west coast of Western Australia. The data were restricted to the 1000-m isobath because this was the offshore limit of the transects. Transect K (Figure 4.5), which covered the region offshore Albany, showed an undercurrent, flowing westward, along the continental slope with a maximum speed of 0.4 m/s and centered at ~400-m depth. The LC was present at the surface as a wide current with maximum speeds of 0.60 m/s. The undercurrent was strongest adjacent to the slope. At transect L, the undercurrent could be identified with a maximum speed of 0.25 m/s and situated below 250 m (Figure 4.6). However, it is possible that the undercurrent was stronger farther offshore, as the continental slope along this transect is steep and the transect did not cover a sufficient distance offshore. Figure 4.5. Velocity transect for transect K on the shelf slope. 65

68 Figure 4.6. Velocity transect for transect L on the shelf slope. Figure 4.7. Velocity transect for transect N on the shelf slope. Transect N, which was located close to an offshore eddy (Figure 3.5), showed the undercurrent was weaker, with similar characteristics as at transect L, except that the transect extended farther offshore (Figure 4.7). Transect N revealed near-bed westward transport on the shelf. At transect P (Figure 4.8), the LU flowed north at two instances: 66

69 centered at around 200 m and at > 300 m; however, the transect did not extend offshore enough to define the undercurrent structure properly. Transect R, located off the Capes region, showed the southward flowing LC with the northward flowing LU located slightly offshore (Figure 4.9). Transects K to R (Figures 4.5 to 4.9) all showed the LC flowing either southward (Capes region) or eastward (south coast) with the undercurrent flowing westward (south coast) and northward (Capes region). At the surface, the LC flowed south off the Capes region, turning east at Cape Leeuwin, and flowing onto the GAB. The ADCP transects also showed that the FC, which acted as an undercurrent beneath the LC, flowed west along the south coast and fed the LU along the west coast. Figure 4.8. Velocity transect for transect P on the shelf slope. 67

70 Figure 4.9. Velocity transect for transect R on the shelf slope Discussion The T S characteristics of the study area, together with the current data and the dissolved oxygen concentrations, confirmed that the FC along the south coast and the LU along the west coast were interconnected and transported the same water mass: subantarctic mode water (SAMW). McCartney (1982) and Wong (2005) defined the SAMW as having σ T values between 28.9 and 29.5, a temperature range of C, a salinity range of , and a high dissolved oxygen concentration of µm/l. The data presented here showed that at the undercurrent core depth of ~400-m depth, the salinity and temperature values were within the same range as that of the SAMW (Figure 4.2). The water mass characteristics were similar along the transects, except at the interface between the surface current (LC) and the undercurrent at depths of around m (Figure 4.2). The consistency of the water mass between the transects, particularly along the south and west coasts, is a good example of how this water mass maintains its continuity. The ADCP current data also showed a continuous flow between the south and west coasts, particularly at the core depth of the SAMW. In almost every transect, the FC started at ~200 m with the undercurrent located beneath the LC; the exception was at 68

71 transect R, where the LC was unusually deep (to ~650 m), which displaced the LU farther offshore. Here the undercurrent was detached from the continental slope. This offshore movement may have been due to the influence of an offshore eddy (Figure 3.5) or the LU changing as it flowed past Cape Leeuwin. The highest undercurrent speed (0.5 m/s) was measured off Albany at transect K off Albany. This value was much higher than the ~0.20 m/s that Middleton and Cirano (2002) and Batteen and Millar (2009) predicted using numerical models. However, this was the first time that undercurrents were measured, as previous shipborne ADCPs in Australia were limited to depths of ~300 m. The current speeds (~0.20 m/s) were lower as the current flowed west and were a similar size to those predicted by numerical models. The weakening of the currents to the west of Albany could have been due to the presence of a strong eddy off transect N (Figures 3.5 and 3.10). Cresswell and Domingues (2009) showed the LC eddies extended to depths > 1000 m, and Meuleners et al. (2007) showed through numerical experiments that the LU also consisted of meanders and eddies similar to that observed with the LC at the surface. The eddy offshore transect N most likely caused the FC to deflect water offshore, as shown in Figure Figure 4.10 shows the surface and subsurface current circulation based on the field data. The surface and subsurface currents flowed in different directions, especially in the southwest corner. A strong LC swerving around the south-west corner created an eddy offshoot which causes FC to split at the southwest corner. On the surface, the FC followed the eddy and continued west into the Indian Ocean, with an inner part of the current recirculating close to the shelf as it neared the eddy. The undercurrent, which flowed close to the shelf, split and rejoined the LU after moving through the eddy at the subsurface. Figure 4.11 shows the general circulation pattern of the LC, LU, and FC off the southwest corner of Australia. The cross-sections A and B show the velocity profile of transects R and K. The FC split, with one arm becoming the LU while the other arm flowed into the Indian Ocean. Thompson and Edwards (1981) found SAMW with temperatures close to 8.6 ºC and a salinity of about south of Tasmania, along 145º E. Middleton and Cirano (2002) 69

72 proposed the FC transported the SAMW west while the LU continued transporting the SAMW north. The results from this study supported both these findings. (a) (b) Figure Illustration of the (a) surface and (b) subsurface circulation off the south-west corner of Western Australia showing the effect of the eddy on the undercurrent circulation. 70

73 Figure General circulation of the Flinders current, Leeuwin current, and undercurrent off the south-west corner of Western Australia with the cross-sections of transect R (A) and transect K (B). Ridgway and Condie (2004) defined the LC as the longest boundary current in the world. They showed the LC signature extended from the North-West Cape of Western Australia to Tasmania, a distance of some 5500 km. For undercurrents, a similar claim can be made. FC/LU system may be longer than the 5500-km-long LC. New findings found that the undercurrent also provides a connection between the Pacific Ocean and the tropical Indian Ocean (Ridgway and Dunn, 2007). Here a remnant of the East Australian current, the western boundary current of the south Pacific Ocean, flows west past southern Tasmania, feeding the FC and the LU, which Woo et al. (2006) reported flowed past North-West Cape. McCartney et. al. (2007) also proposed the East Australian current and the FC were connected. Thompson and Edwards s (1981) findings that the FC and LU transported SAMW from the south of Tasmania to the west further supported this theory. Wong (2005) analysed ARGO floats and proposed that the transport pathways of SAMW between south of Australia and Indian Ocean is connected (Figure 4.12). The results of this study confirmed these pathways (i.e. it removes the?? mark from Figure 4.12). 71

74 4.6. Concluding remarks The results obtained from examining the water mass characteristics and velocity distribution of the FC and LU confirmed the interconnection between the FC and LU. The FC core was always attached to the slope and part of it flowed below the LC. Because the FC is the main current in the southern region, only part of the FC interconnected with the LU and continued north. At the end of the FC s westward transport, the main flow of the FC flowed into the Indian Ocean. In the study region, the SAMW core was present at ~400-m depths. The temperature, salinity, and dissolved oxygen corresponded with the positions of the FC and LU. The continuity of these components showed the FC fed the undercurrent in the south-west region and provided a pathway for the transport of SAMW from Tasmania to the tropical Indian Ocean, perhaps forming the longest boundary undercurrent in the world. Figure 4.12 Schematic showing the inferred pathway of the south Indian Ocean SAMW (from Wong, 2005). 72

75 Chapter 5 A numerical model study of the Leeuwin current system along the south coast of Western Australia Introduction The unique current system off Western Australia s south coast consists of three main current systems: (1) the eastward flowing Leeuwin current at the shelf break and slope; (2) the westward flowing Flinders current, flowing offshore and beneath the Leeuwin current; and (3) the wind-driven, westward flowing Cresswell current on the continental shelf during the summer in response to easterly winds. The Leeuwin current (LC) is a continuous boundary current mainly centered on the shelf break, which flows southward along the west coast of Australia and eastward into the Great Australian Bight (GAB). The Flinders current (FC), the only northern boundary current in the Southern Hemisphere, interacts with the LC on the continental shelf and flows on the outer part of the shelf break. The flow is opposite to the direction of LC. Part of the FC flows beneath the LC and feeds the Leeuwin undercurrent (LU), then flowing northwards along the west coast (Figure 4.10; Chapter 3). The LC undergoes an ageing process along Western Australia s west coast (Woo et al. 2006; Chapter 3). Cooler, higher salinity water is entrained from the West Australian current between 29 and 32 S (Cresswell and Peterson 1993) and changes the LC water mass characteristics. 3 This chapter is to be submitted as a journal paper to Ocean Dynamics 73

76 (a) Abrolhos Islands Fremantle Cape Leeuwin Albany Point D Entrecasteaux Model Domain (b) Transect R Transect K Figure 5.1. (a) Model domain with bathymetry contour lines. (b) The study area overlaid with the field transect of the Southern Surveyor cruise SS04/06. The current system was observed at two locations: Transect R Capes region (latitude 34º S) and transect K Albany (longitude 117.7º E). 74

77 In Chapter 3 we explained that the alongshore geopotential gradient drive the LC along the west and south coasts, but an enhanced alongshore geopotential gradient at the southwest corner accelerate the LC locally. High LC speeds and the sudden change in the current s direction generate eddies between 115 and 117 E along the south coast (Cresswell and Domingues, 2009). The LC decelerates after it passes the south-west corner, but maintains speeds up to 1 m/s along the south coast. During a research voyage undertaken in April 2006, the measured speeds of the LC were higher along the south coast than those along the west coast. One of the reasons for the higher velocities is related to the narrow continental shelf and the steep bathymetric slope, which decreases the width of the current and thus increases the speed (Pingree et al. 1999). Woo et al. (2006) also found that the LC had higher current speeds offshore the Abrolhos Islands due to the steep continental shelf slope. Higher LC speeds also generate baroclinic instabilities, which cause the current to meander on and off the shelf throughout most of the year and generate numbers of eddy (Batteen and Butler 1998; Middleton and Cirano 2002; Batteen et al. 2007; Cresswell and Domingues, 2009). The Flinders current (FC) and Leeuwin undercurrent (LU) transport sub Antarctic mode water (SAMW) beneath the LC, with the core of the currents located at a depth of 400 m. The combined FC and LU transport water originates from the Pacific Ocean and flows to the west coast and onwards to the tropics (Chapter 4). The large-scale wind stress curl to the south of Australia drives the westward flowing FC (Middleton and Bye 2007). The Flinders current, which is stronger in summer, extends from the surface to a depth of ~1000 m, with peak currents of ~0.50 m/s at about 400-m depth (Chapter 4). Wind-driven currents, which flow opposite to the LC, are present on the continental shelf during summer. Due to upwelling, these currents are characterized by cooler temperatures. The Capes current (CC) originates from the region between Capes Leeuwin and Naturaliste and flows northward (Gersbach et al. 1999; Pearce and Pattiaratchi 1999). Along the south coast, upwelling of colder water was observed in Recherché Archipelago and adjacent waters (van Hazel 2001). It has been postulated that a wind-driven coastal current, similar to the CC, moves westward in response to the strong, southeasterly summer winds, which are the major driving force of the current and upwelling (Figure 2.3). The current is named the Cresswell current (Pattiaratchi 2006; McClatchie et al., 2006). 75

78 In this study we used numerical model simulations to examine the circulation along the south-west coast of Western Australia. The model was developed to simulate the surface and subsurface current systems, with the objective of resolving the seasonal variability and to ascertain the role of wind stress as a driving mechanism of the LC and coastal currents along the south coast. The broader goal was to examine the seasonal variability from the results of the field study, which was undertaken over a period of three weeks in April Model set-up The three-dimensional model used was the Regional Ocean Modelling System (ROMS) v. 2.1 (Haidvogel et al. 2000; Shchepetkin and McWilliams 2005). ROMS is a hydrostatic, primitive equation ocean model, which solves the Reynolds-averaged form of the Navier Stokes equations on an Arakawa C grid and uses stretched, terrain-following (sigma) coordinates in the vertical. Here the model s horizontal advection scheme was configured using a third-order, upstream-bias G-Scheme (Kantha and Clayson 2000) and a spline advection scheme for the vertical transport. The baroclinic hydrostatic pressure gradient term was examined using Song and Wright's (1998) weighted Jacobian pressure gradient scheme. The parameterization of vertical turbulent mixing for momentum and tracers was based on the Mellor Yamada 2.5-order closure model (Mellor and Yamada 1982). The horizontal eddy coefficient used for momentum and tracers adopted a Laplacian parameterization. The model domain encompassed Western Australia s west and south coasts (26 38 S, W) (Figure 5.1a). It consisted of three open boundaries at the study region s northern, southern, and western extents. A time-varying boundary condition for temperature and salinity was maintained at all the open boundaries to represent the internal forcing. Radiation boundary conditions were prescribed for the baroclinic tracers. The forcing data were sourced from the Bluelink Ocean Data Assimilation System (BODAS), which has a resolution of 10 x 10 km and was designed to forecast and reanalyse the oceans around Australia (Oke et al. 2008). The daily data of Bluelink from year 2003 and 2004 was used to complete the model runs. 76

79 The model was run for 400 days to cover a simulation of a complete year cycle. This is important to assess seasonal variation of the study region through the model. The horizontal resolution was set to a nominal 2.5 km in the east/west and north/south directions. Geoscience Australia supplied the bathymetric data at the nominal resolution of 1 km. The model domain s minimum depth was set to 20 m and the number of vertical sigma layers was set to 25, with increased resolution in the near-surface and near-seabed layers. A quadratic bottom-drag coefficient was used to compute the bottom stress, and the model was set to perform two runs, with and without wind stress forcing. The wind stress forcing was extracted from wind data obtained from the National Oceanic and Atmospheric Administration (NOAA). The data was obtained for particular year of the model run (2004). The wind data provide daily wind stress with spatial resolution of 1. This data then was interpolated to incorporate within the model. The spatial and temporal resolution of the wind data is appropriate to provide enough coverage for seasonal variation. Often the Capes region and south of WA were a boundary for wind direction to change direction (Rennie et. al. 2008), Figure 5.2 provide an averaged wind stress for both winter and summer to clearly distinguish the difference in wind pattern to the system, particularly at Capes and southern WA region. For analysis purpose, temperature and current speed is using snap shots from each season for comparison. This is to avoid some of the eddy features which sometimes are quite dominant. By doing seasonal averaging, it will cancel out some of the important features. Thus, snap-shot is chosen to better represent the system. 5.3 Model results Here we compare the model output with satellite imagery and then discuss the seasonal variability of the current system. Results from the model runs undertaken with and without wind stress forcing will then be compared to identify the role of wind stress as a driving force of the Leeuwin current along the south coast. Two transects, transect R (Capes region) and transect K (Albany) (Figure 5.1b), located along the west and south coast, respectively, were selected to examine the variability with depth. 77

80 A satellite image of the sea surface temperature off south-west Australia on 9 April 2006 (Figure 5.2a) showed a warm LC near the Capes region, typical of the circulation pattern during autumn. The model output of the sea surface temperature during autumn conditions showed that the model reproduced the broad-scale features of the satellite image (Figure 5.2b); the LC was located near the shelf and flowed continuously along the shelf break towards the GAB. Along the south coast, the strong temperature front due to the interaction between the warmer LC and the cooler FC was prominent. In the Capes region, an eddy generated off Cape Leeuwin was almost identical to the eddy in the satellite image; a patch of cold water drawn from the anti-cyclonic eddy could also be seen. 78

81 (a) (b) Figure 5.2. (a) NOAA satellite image of south-west Australia on 9 April The warm water (shown in red/orange) represents the Leeuwin current. (b) A model output of the surface temperature representing circulation in the autumn and showing an eddy offshoot south of Cape Leeuwin whilst the Leeuwin current is located close to the shore. 79

82 5.3.1 Seasonal characteristics Summer The LC in summer had a maximum speed of ~1.0 m/s (Figure 5.3a). It was positioned offshore from the shelf break and advected warm water into the southern region. The presence of coastal currents on the shelf had pushed the LC offshore from the shelf, as described by Pearce and Pattiaratchi (1999). The coastal current along the west coast is known as the Capes current and the coastal current along the south coast is known as the Cresswell current. In summer, the wind stress on the west and south coasts are upwellingfavourable, with southerly winds along the west coast and easterly winds along the south coast. Along both coasts, the wind stress overcomes the alongshore pressure gradient, resulting in the surface layers moving offshore, colder water upwelling onto the continental shelf, and the Leeuwin current migrating offshore (Gersbach et al. 1999). A maximum temperature difference between the shelf waters and the LC of around 5 C was predicted by the model, which agreed with the results obtained with the satellite imagery (Cresswell and Peterson, 1993). Transect R (west coast; Figure 5.1) showed a poleward LC with a maximum speed of ~1.0 m/s located just offshore of the shelf break and an equatorward Capes current with a maximum speed of ~0.2 m/s inshore of the continental shelf (Figure 5.3b). The transect also indicated the presence of the equatorward Leeuwin undercurrent beneath the LC flowing adjacent to the continental slope between 400 and 1000-m depth with speeds of ~0.2 m/s centred at 400-m depth. These observations compared favourably with the observations presented in Chapters 3 and 4. The LC maintained its speed as it flowed into the southern region, but narrowed and covered a deeper cross-sectional area (Figure 5.3c). The Cresswell current was present as an easterly flow along the continental shelf with current speeds similar to that of the Capes current along the west coast. Both undercurrents (FC and LU) were of similar magnitude. Autumn The LC was stronger in autumn than it was in summer. The surface current was strongest after it had passed the south-west corner (at ~116 E) and was on a straight path to the 80

83 south of Albany (Figure 5.4a). The current reached speeds of more than 1.1 m/s, but weakened as it flowed eastward, as described by Cresswell and Domingues (2009). Along the west coast the warmer LC waters flowed onto the shelf, although there was a weak Capes current on the inner shelf (Figure 5.4b). Some upwelling was still present near the Capes region and farther north, but it covered less shelf area than it had in summer. Along the south coast the FC flowed eastward, offshore of the LC, with speeds of ~0.2 m/s and a temperature of ~17 C, similar to the model simulations by Middleton and Cirano (1999). Transect R showed the LC had increased to 400-m depth and also covered more area on the shelf (Figure 5.4b). At Transect K, the LC covered nearly the whole shelf, with its core above the shelf break (Figure 5.4c). The LC depth had increased to 500 m, and the core of the FC core was located offshore of the continental slope. Winter The sea surface temperature in winter was lower compared with both summer and autumn mainly due to heat loss from the surface. The warmer LC core was present on the shelf, with the surface temperature ~2 C less than it was in autumn (Figure 5.5a). The current flowed closer to the coast and with the strongest currents (~1.0 m/s) off the south-west corner. Transect R showed the LC, with its core near the slope, covering almost the whole shelf (Figure 5.5b). The LC depth was 500 m, and the current core had narrowed; the eddy present offshore the transect (Figure 5.5a) likely caused this narrowing. Transect K showed the LC had strengthened and covered a larger cross-section than it had in autumn. The LC core was slightly offshore because of the wider area that the LC covered (Figure 5.5c). The LC depth was ~300 m, and the FC, which was centred below the LC core, flowed westward with a maximum speed of ~0.3 m/s. The coastal current, however, had disappeared, as in winter the winds had changed from predominantly having an easterly component to a westerly component. The Leeuwin Current at Transect K in winter is wider but, nonetheless the weaker summer current is much deeper. And undercurrent is slightly strong in winter. This is similar to what found by Batteen and Miller, Earlier model shows undercurrent along the west coast can be generated due to meridional variability (Batteen et al., 1989) 81

84 and/or to the planetary beta effect (e.g., McCreary et al., 1986; Batteen, 1997). The beta effect allows the existence of freely propagating planetary waves, i.e., Rossby waves (Gill, 1982). The offshore propagation of these waves contributes to the generation of an alongshore pressure gradient field, which can aid the development of subsurface currents along the eastern boundary. In time the surface current can widen and causing the undercurrent to intensify and extend closer to the surface. As a result, the beta effect can change both the vertical and horizontal structure of the surface and subsurface currents (Batteen and Miller, 2009). 82

85 (a) (b) (c) Figure 5.3. Model output with wind stress forcing representing summer conditions. (a) Sea surface temperature ( C) and current velocities; (b) the cross-section across transect R (Capes region) showing the north-south velocity (m/s); (c) the cross-section across transect K (Albany) showing the east-west velocity (m/s). 83

86 (a) (b) (c) Figure 5.4. As for Figure 5.3, but showing autumn conditions. 84

87 (a) (b) (c) Figure 5.5. As for Figure 5.3, but showing winter conditions. 85

88 5.3.2 The influence of wind stress A model run was undertaken without prescribing wind stress to determine its importance as a forcing mechanism of the currents along the south coast. The summer and winter conditions were compared because along the south coast, the winds reverse in direction between these two seasons: in summer, the winds have a dominant easterly component, whereas in winter the winds have a strong westerly component. The model simulations without wind stress indicated a strong Leeuwin current, with the southerly flows extending onto the shelf along both west and south coasts. The crosssectional transects R and K were similar to the model runs with wind stress forcing (Figures 5.5b and 5.5c). At transect R, the LC was centred on the shelf break, extending to 350-m depth, with maximum currents of 1.2 m/s (Figure 5.6b). The almost identical velocity distribution between the model simulations with and without wind stress (cf. Figures 5.5b and 5.6b) indicated that the effects of wind stress were negligible for the Leeuwin current dynamics during winter along the west coast. Along the south coast, transect K indicated that the LC extended farther offshore, with a weakening of the FC at the surface, although the FC beneath the LC reached speeds of ~0.2 m/s (Figure 5.6c). In winter, the Leeuwin current was stronger (Figure 5.5c) than that predicted in the absence of wind forcing. 86

89 (a) (b) (c) Figure 5.6. As for Figure 5.3, but without wind forcing. 87

90 5.4 Discussion The Leeuwin current Field and numerical modelling studies have showed the LC flows southward along the west Australian coast, and after flowing past the south-west corner of Australia, flows eastward (Cresswell and Peterson 1993; Batteen and Butler 1998; Ridgway and Condie 2004; Batteen et al. 2007; Batteen and Millar, 2009; Cresswell and Domingues, 2009); the model simulations presented here have broadly reproduced the observations and results of these investigators. The model results indicated that the LC strength and the surface temperatures varied throughout the year. The highest temperatures were recorded in autumn (April May) when the LC core reached the southern region (Ridgway and Condie, 2004). Cresswell and Domingues (2009) summarised the seasonal behaviour of the LC along the south coast as follows: From December to March, the boundary flow of the Leeuwin current is quiescent. It suddenly pulses in April and evolves into a vigorous current during austral winter. The eastward flow of the Leeuwin current in the Great Australian Bight is initially strongest only on the western part of the basin, but by July/August it has accelerated along the entire basin. The vigorous flow then becomes unstable and dissipates a large amount of its energy by shedding mesoscale eddies. The model simulations accurately reproduced this behaviour, including the generation of eddies off Cape Leeuwin and along the south coast. The field data (Chapter 3) indicated that the LC accelerated along the south-west corner in response to an increase in the geopotential gradient in that region as the LC flows into the cooler, lower salinity sub-antarctic region. The model reproduced this feature for all three seasons and this was also reflected in the increased (model-predicted) geopotential gradient in the region, with the maximum slope located off the south-west corner (Table 5.1). The results also indicated a slight variation in the geopotential gradient between the three seasons, with maxima in winter corresponding to the maximum flow and the minima in summer (Table 5.1) confirming the role of the geopotential gradient as the primary driving force of the Leeuwin current. 88

91 Table 5.1. Geopotential gradient for summer, autumn and winter at the location of the maximum LC current speeds and at the south-west corner. The geopotential anomaly was calculated between the surface and 300dB along the LC core. Slope at maximum LC Slope at SW corner Summer x x 10-7 Autumn x x 10-7 Winter x x 10-7 Table 5.2. Volume transport of the Leeuwin current, Leeuwin undercurrent and Flinders current in summer and winter across transects R (Capes region) and K (Albany), with and without wind stress (units: Sverdrup). Transect K With wind stress Without wind stress LC FC LC FC Summer Winter Transect R With wind stress Without wind stress LC LU LC LU Summer Winter Note: Flinders current transport given for the undercurrent only. The wind stress also modulated the strength of the LC, with differences observed between the south and west coasts. Along the south coast (transect K), the winds were mainly westerly in autumn and winter, i.e. in the same direction of the LC. This resulted in a stronger LC (Table 5.2), which was deeper and flowed closer to the shelf. In summer, the winds along the south coast were mainly easterly (i.e. against the current), which weakened the LC. Along the west coast (Transect R), the LC was much stronger during winter than it was in summer (Table 5.2). Both the FC and the LU were marginally stronger in summer compared to winter, as found by Middleton and Cirano (1999). 89

92 There were few differences between the model runs without wind stress and the models runs with wind stress along the west coast. In summer, the LC was slightly stronger in the models runs without wind stress due to the removal of the southerly wind stress whist the LU was slightly weaker (Table 5.2). In winter, the pattern was reversed, with a slightly weaker LC and a slightly stronger LU in the absence of wind stress (Table 5.2). The biggest difference between the wind and no wind model runs was found along the south coast, where the influence of westerly winds as a driving force of the LC was highlighted. The predicted transport of the LC (Table 5.2) was much higher with the inclusion of wind stress (4.5 Sv) than it was in the absence of wind stress (1.4 Sv). This was also reflected in the cross-sectional transect plots, which indicated a stronger and wider LC along the south coast with the inclusion of wind (cf. Figures 5.5c and 5.6c) The Leeuwin undercurrent and the Flinders current Field data (Woo et al. 2006) and numerical modelling (Meuleners et al. 2006; Rennie et al., 2007) have been used to examine the LU. Numerical modelling has also been used to study the major current along the south coast, the Flinders current, which is a surface current as well as an undercurrent (Middleton and Cirano 2002; Batteen et al. 2007; Middleton and Bye 2007). Chapters 3 and 4 discussed the FC flowing in opposite directions to the LC, with the main current flow located offshore of the LC. Part of the FC flows beneath the LC, in a similar manner to the LU along the west coast, and feeds the LU. The strongest currents associated with the FC were associated with the undercurrent, where maximum currents of 0.50 m/s were recorded. The model also reproduced these features, which were observed in the field data. The velocity field at 550-m depth (Figure 5.7) showed a continuous FC undercurrent flowing on the continental slope along the south coast. At Cape Leeuwin it appeared to bifurcate, with one arm flowing northwards as the LU and the other flowing westward into the Indian Ocean (as postulated in Chapter 4). In the eastern section of the model domain, the FC was not well defined at this depth. The speed of the FC (~0.50 m/s) was greater along the continental slope, i.e. where it was an undercurrent, and similar to that observed in the field measurements (Chapter 3). 90

93 Figure 5.7. The velocity field at a depth of 550 m for the autumn circulation. The colour bar indicates the velocity in m/s. The seasonal variability of the LU revealed that the transport was highest in summer, followed by autumn, and lowest in winter; the strong LC was responsible for pushing the LU deeper and resulted in the transport decreasing in winter. The FC transport in summer and autumn resembled the LU transport for the same seasons. In winter, the surface flow of the FC, offshore of the LC, was weak, with the maximum currents consistently found under the LC. Middleton and Cirano (2002) proposed the stronger wind stress increased the FC transport during the summer months. The results from the model run without wind stress forcing agreed with this finding. In summer, the presence of the surface FC offshore was weak, with the maximum FC currents below the LC. In winter, when the surface currents weakened, the FC covered a larger surface area. The FC transport of ~2 Sv (Table 5.2), which is higher in summer, represents the undercurrent and is considerably smaller than the 8 17 Sv range usually associated with the FC (Section 2.2.4). The model runs without wind stress also indicated the undercurrents (FC and LU) were stronger in winter than in summer. Hence, it appears that the undercurrent transport increased when the LC was weak, and this usually happens in summer when the wind stress has less influence. Woo and Pattiaratchi (2008), through analysis of water masses along the west coast, indicated that under El Niño conditions, the depth range occupied by the sub Antarctic mode water (SAMW), the water mass 91

94 associated with the LU was larger, indicating a higher transport of LU. The interannual variability of the LC is such that it is weaker during El Niño events (Pattiaratchi and Buchan, 1991; Feng et al., 2003). This adds more credence to the observation that when the LC is weaker the undercurrents are stronger and vice versa The Capes and Cresswell currents The Capes current is a cooler continental shelf current, driven by a southerly wind stress, which overcomes the alongshore geopotential gradient, mainly during the summer months. The cooler signature of the current is due to wind-driven upwelling. The current originates from the region between Capes Leeuwin and Naturaliste and moves equatorward (Gersbach et al. 1999; Pearce and Pattiaratchi 1999; Hanson et al., 2005; Pattiaratchi and Woo, 2009). Gersbach et al. (1999) described the dynamics of the Capes current off Cape Mentelle. The continental shelf in Australia s south-west comprises a step structure, with an inner shelf break at 50 m and an outer shelf break at 150 m (Pearce and Pattiaratchi 1999). This bathymetry influences the circulation, especially in the summer. In the summer, the alongshore wind stress overwhelms the alongshore pressure gradient on the inner shelf (depths < 50 m), moving surface layers offshore, upwelling colder water onto the continental shelf, and pushing the Leeuwin current offshore. Here the Capes current is present on the inner shelf and bounded offshore by the Leeuwin current on the lower shelf, with upwelling occurring over the inner shelf break (Gersbach et al. 1999). Numerical modelling results showed a wind speed of 7.5 ms 1 was sufficient to overcome the alongshore pressure gradient on the inner continental shelf (Gersbach et al., 1999). During winter, the Leeuwin current strengthens, and in the absence of wind stress, migrates closer inshore, flooding upper and lower terraces (Pearce and Pattiaratchi 1999). The Cresswell current is similar to the Capes current that flows along the south coast from east to west in response to easterly wind stress during the summer, and has only been postulated and observed in satellite imagery. The dynamics of the current remain largely unknown, but are thought to be similar to that of the Capes current (McClatchie et al., 2006; Batteen and Miller, 2009). Van Hazel (2001) observed upwelling of colder water off the Recherché Archipelago and adjacent waters and attributed it to a coastal current moving from east to west. 92

95 The numerical simulations predicted the presence of both the Capes and Cresswell currents in summer and autumn, but not in winter. Figures 5.8a and 5.8b show the sea surface temperature and velocity on the shelf off the Capes region and Albany, respectively, during summer. The cooler Cresswell current, flowing from east to west on the shelf with speeds of ~ m/s, is clearly present off Albany (Figure 5.8b), with the eastward flowing Leeuwin current farther offshore. At Cape Leeuwin, the current fed the Capes current, as postulated by Gersbach et al. (1999). Transects R (Capes region) and K (Albany) showed the Cresswell current was confined to the upper terrace (< 70 m); however, the Capes current extended farther offshore into deeper water of up to 150-m depth (Figures 5.8c and 5.8d). The Cresswell current was generally stronger than the Capes current, especially between Albany and Point D Entrecasteaux, most likely due to changes in the bathymetry: the shelf width and slope increase abruptly to the west of Point D Entrecasteaux. Thompson (1987) suggested that the balance between the alongshore geopotential gradient, wind stress and bottom friction in the alongshore direction may be expressed as: 0 H 1 P 2 dz + u* ρ y + C D V B V B = 0 (5.1) Here, z is the vertical coordinate; y is the alongshore direction; H the local water depth; P is the alongshore geopotential gradient; ρ is mean density; u* is the friction velocity at y the air sea interface; C D is the bottom drag coefficient; and V B is the alongshore velocity just above the bottom friction layer. In the equation (5.1), term 1 represents the depthintegrated alongshore geopotential gradient; term 2 represents the wind stress; and term 3 represents the bottom friction approximated using a quadratic law. Examining the balance of different terms in equation (5.1) reveals that in shallow water, the depth-integrated geopotential gradient is small compared to wind and bottom stress terms. These results in the depth mean flow in the direction of the wind. However, as the water depth increases, the geopotential gradient becomes more important until a point is reached where the geopotential gradient balances the equatorward wind stress, resulting in zero depth- 93

96 averaged velocity. Thompson (1987) defined this depth as H*, given by: H * 2 = u 1 * P ρ y 1 Thompson (1987) predicted that upwelling would occur if H* exceeded the mixed layer depth, and Gersbach et al. (1999) showed that upwelling may occur off south-west Australia up to water depths of 70 m. Along the south coast, as the shelf is narrower and the depth beyond the shelf break increases rapidly, the dominance of the wind over the alongshore pressure gradient is limited to the upper terrace at around 70 m (Figure 5.8d). Figures 5.8e and 5.8f show that the temperature difference between the Cresswell current at the shelf and the slope was about 2 C and that upwelling occurred from beneath the LC. During autumn, the Cresswell current was present between Point D Entrecasteaux and Cape Leeuwin and feeding the Capes current, but was present only close to the coast near Albany (Figures 5.9a and 5.9b). Transect R showed the Capes current reached speeds of ~0.2 m/s and was on the shelf (Figure 5.9c). The Cresswell current was only present close to the coast at Transect K (Figure 5.9d), with the eastward flowing LC spreading onto the continental shelf. Figure 5.10 shows the wind regime for the south-west of Western Australia, which used the data set from the model forcing and the yearly averaged data provided by NOAA. This data set agreed with the field data presented earlier, where the wind regime compose of almost similar direction (Figure 2.3). The Cresswell current was strong in summer when the winds were usually south-easterly (Figure 5.10). The wind direction in summer was upwelling-favourable for the Cresswell current and the Capes current; however, in autumn, the wind direction changed by ~90 (clockwise), resulting in onshore winds off Albany and southerly wind (upwelling-favourable) for the Capes current along the west coast. The wind stress was also weaker in autumn. The changes in wind stress, the driving force of the shelf current systems, explains why the Capes current was still present in autumn and the Cresswell current was weaker. Data obtained aboard the Southern Surveyor in autumn (presented in Chapters 3 and 4) also showed no evidence of the Cresswell current because of the dominance of south-westerly winds. Similarly, the 94

97 model simulation did not indicate the presence of the Cresswell or Capes currents during winter. (a) (b) (c) (d) (e) (f) Figure 5.8. a) Sea surface temperature ( C) and Capes current velocity for the Capes region in summer. b) Sea surface temperature ( C) and Cresswell current velocity for the Albany region in summer. c) Transect R cross-section of the Capes current north-south velocity (m/s) in summer. d) Transect K cross-section of the Cresswell current north-south velocity (m/s) in summer. e) The Capes current summer temperature ( C) at transect R (Capes region). f) The Cresswell current summer temperature ( C) at transect K (Albany). The colour bars indicate the current velocity in m/s. 95

98 Figure 5.9. a) Sea surface temperature ( C) and Capes current velocity for the Capes region in autumn. b) Sea surface temperature ( C) and Cresswell current velocity for the Albany region in autumn. c) Transect R (Capes region) cross-section of the Capes current north-south velocity (m/s) in autumn. d) Transect K (Albany) cross-section of the Cresswell current north-south velocity (m/s) in autumn. The colour bars indicate the velocity in m/s. 96

99 Figure The summer and autumn wind vectors in the continental shelf regions off the Capes region and Albany. The wind stress was extracted from the National Oceanic and Atmospheric Administration daily wind stress data. Figure Schematic of the summer steady state current regime off Western Australia s south coast. 97