THE LEEUWIN CURRENT: an influence on the coastal climate and marine life of Western Australia Journal of the Royal Society of Western Australia Volume 74,1991 Proceedings of a Symposium of the Royal Society of Western Australia and the Western Australian Branch of the Australian Marine Sciences Association, held at CSIRO Laboratories, Perth, Western Australia, 16 March 1991 Edited by A F Pearce 1 & D I Walker 2 1 CSIRO Division of Oceanography department of Botany, The University of Western Australia w \ < JOURNAL OF THE ROYAL SOCIETY OF WESTERN AUSTRALIA CONTENTS VOLUME 74,1991 The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia Proceedings of a Symposium of the Royal Society of Western Australia and the Western Australian Branch of the Australian Marine Sciences Association. Perth, Western Australia, March 1991. Edited by A F Pearce and D I Walker Page The Leeuwin Current - observations and recent models G R Cresswell 1 Homologous peri-oceanic west coast climates in the southern hemisphere J Gentilli 15 Eastern boundary currents of the southern hemisphere A F Pearce 35 The Abrolhos carbonate platforms: geological evolution and Leeuwin Current activity L B Collins, K H Wyrwoll & R E France 47 Zoogeographic provinces of the Humboldt, Benguela and Leeuwin Current systems G J Morgan & F E Wells 59 The effect of sea temperature on seagrasses and algae on the Western Australian coastline D I Walker 71 Dispersal of tropical fishes to temperate seas in the southern hemisphere J B Hutchins 79 Mass spawning of corals on Western Australian reefs and comparisons with the Great Barrier Reef C Simpson 85 The Leeuwin Current and larval recruitment to the rock (spiny) lobster fishery off Western Australia B F Phillips, A F Pearce & R T Litchfield 93 The influence of the Leeuwin Current on coastal fisheries in Western Australia RC Lenanton, L Joll, J Penn & K Jones 101 Coral reefs in the Leeuwin Current - an ecological perspective B G Hatcher 115 Seabird abundance, distribution and breeding patterns in relation to the Leeuwin Current R D Wooller, J N Dunlop, N I Klomp, C E Meathrel & B C Wienecke 129 Implications of long-term climate change for the Leeuwin Current C B Pattiaratchi & S J Buchan 133 Preface This symposium was sponsored by the Royal Society of Western Australia, with additional support from the Western Australian Branch of the Australian Marine Sciences Association (AMSA). Its purpose was to bring together, for the first time, the multidisciplinary group of researchers studying various aspects of the Leeuwin Current. It is hoped that as a consequence of the symposium and this volume, further research into this unique and interesting current system will be stimulated. It was appropriate that Dr George Cresswell should deliver the opening paper at the Symposium as his oceanographic researches during the 1970s led to the clear identification and naming of the Leeuwin Current. He can justly be termed the ’’father of the Leeuwin Current". It is equally fitting that Dr Joseph Gentilli, who presented the second paper, be the "grandfather of the Leeuwin Current", as he had shown the existence of tongues of warm water down the west Australian coastline a decade earlier. All the papers in this volume were presented verbally at the symposium. Dr Bruce Hatcher's paper was read by Dr Bob Black. It is inevitable that there is a degree of overlap between some of the papers. We have tried to maintain a balance between minor repetition and the requirement for each paper to stand alone. As we abhor the proliferation of "grey" (un-refereed) conference literature, all the papers in this volume have been refereed critically according to the normal standards of scientific publication. We are grateful to the following reviewers for their efforts to ensure that a high scientific standard has been maintained, and for their prompt cooperation: Dr R Allan (CSIRO Division of Atmospheric Research) Dr D Ayre (University of Wollongong) Dr N Caputi (Western Australian Department of Fisheries) Dr G Cresswell (CSIRO Division of Oceanography) Dr C Crossland (CSIRO Institute of Natural Resources and Environment) Dr S Godfrey (CSIRO Division of Oceanography) Dr A Huyer (Oregon State University, USA) Mr G Kendrick (The University of Western Australia) Dr H Kirkman (CSIRO Division of Fisheries) Prof A McComb (Murdoch University) Dr J Middleton (University of New South Wales) Dr C Pattiaratchi (The University of Western Australia) Mr A Pearce (CSIRO Division of Oceanography) Dr B Phillips (CSIRO Division of Fisheries) Dr P Playford (Geological Survey of Western Australia) Dr D Pollard (Fisheries Research Institute of New South Wales) Dr G Poore (Museum of Victoria) Dr D Saunders (CSIRO Division of Wildlife and Ecology) Dr V Shannon (Sea Fisheries Research Institute, South Africa) Dr J Stoddart (Kinhill Engineers) Dr D Walker (The University of Western Australia) The organising committee comprised the editors, and Nick D'Adamo, Peter Murphy, Lesley Thomas and Valerie Pearce. We thank CSIRO Floreat for allowing us to use their excellent conference facilities as a venue for the symposium, and all who assisted us on the day. The manuscript formatting was carried out at the Botany Department, and artwork was provided by the Centre for Water Research, both within The University of Western Australia. We record our appreciation of the contribution made by Ainsley Calladine with diagram transfers and advice on desktop publishing. We thank Fiona Webb for general assistance with word-processing. AFP D I W Journal of the Royal Society of Western Australia, 74, 1991,1-14 The Leeuwin Current - observations and recent models G R Cresswell CSIRO Division of Oceanography, Hobart, Tas 7001, Australia Abstract The Leeuwin Current carries warm low salinity water from northwestern Australia into the prevailing equatorward wind to Cape Leeuwin and then across the Great Australian Bight. Current speeds to the west of the continent can exceed 0.5 ms" 1 , while to the south they can exceed 1.5 ms' 1 . In both regions the maximum speeds are encountered just beyond the continental shelf edge. Although the current is a low salinity feature to the west of the continent, once it rounds Cape Leeuwin it enters a regime of cold, low salinity waters so that it is then relatively high in salinity. The current frequently meanders and breaks out to sea forming both cyclonic and anticyclonic eddies. On its shoreward side it spreads across the continental shelf, commonly reaching the very near shore south of Western Australia. In this review a description of the Leeuwin Current is given by making use of observations ranging back to those of Flinders in 1803. Recent models of the current are found to be quite successful in describing many of its features. Historical If we define the Leeuwin Current (Church et al. 198 9, Smith et al. 199 1) as a stream of warm, low salinity water that flows at the surface from near NW Cape down to Cape Leeuwin and thence towards the Great Australian Bight (as suggested by the satellite image in Fig. 1), then we Find evidence for it as early as the start of the last century. In May 1803 Flinders (1814) was set to the east between Cape Leeuwin and Albany at a little over one knot. Between Albany and the Recherche Archipelago (near Esperance) he reported the current to increase from the coast seaward. And "in coasting qll around the Great Bight" he had no measurable current, suggesting, perhaps, that the current did not spread coastward across the wide continental shelf. Along the west coast of Australia the earliest evidence for a current of tropical origin came from observations of warm waters and tropical marine flora and fauna around the Abrolhos Islands (~29°S) by naturalist Saville-Kent (1897). Also, early reports by fishermen described a southwards current between Geraldton and the Abrolhos Islands in winter which increased when northerly winds blew (Dakin 1919). Halligan (1921) presented a chart of the currents around Australia (Fig. 2) and described how the warm comparatively light waters of the Indian Ocean would have to discharge to the south, since there is no northern outlet. With, as we will see, some precision, he described how a cold and heavy Southern Ocean current approached the south-west coast of Australia and dipped beneath the warm southerly drift. From the vicinity of Cape Leeuwin he reported "a warm southerly and easterly surface current" with speeds of 0.3 - 0.4 knots. The August quarterly sea surface temperature map presented by Schott (1935) clearly showed warm waters (>16°C) to have been carried around Cape Leeuwin and eastward. This is reinforced by his winter (August- September) current chart (Fig. 3), which shows the Leeuwin Current more or less as we know it. Southward flow of low salinity waters in autumn and winter and northward flow of high salinity waters in summer were seen from drift bottle measurements and hydrological observations by Rochford (1969). Historical bathythermograph data were interpreted by Gentilli (1972) to show that the throughflow from the Pacific to the Indian Ocean in autumn and winter was isolated by a reversal of the flow in spring. This water then achieved thermal homogeneity over the summer to become a 'raft' of warm water, which spread southward during the following autumn and winter (Fig. 4). The 'raft' description was coined in the 1950's by Dr D L Serventy during discussions with fishermen who mentioned the warm water and the tropical species in it (Gentilli pers. comm. 1991). 1 Journal of the Royal Society of Western Australia, 74, 1991 Figure 1 The temperature distribution off western Australia as determined from an infrared satellite image from 15 June 1984. The Leeuwin Current starts as a broad fan of warm water (>24°C) off the northwest and progresses southward to Cape Leeuwin and then eastward to the Great Australian Bight. There are eddies, meanders and offshoots associated with the Leeuwin Current. Mixing, radiation and evaporation change its water properties as it progresses. Some fifteen years ago the response of satellite- tracked drifters (Cresswell & Golding 1980) to the flow of water of tropical origin was quite dramatic (Fig. 5), showing a drift to Cape Leeuwin and then eastward, as well as the interaction between this drift and the eddies offshore from it. Cresswell & Golding called the drift the Leeuwin Current, after the Leeuwin (Lioness in English), a Dutch ship that explored eastward towards the Bight in 1622. The Leeuwin Current flows principally in autumn and winter. It is unusual in that it flows southward and into the wind. Other current systems on the eastern sides of oceans - the Benguela, Canary, Peru, and California current systems of the Atlantic and Pacific Oceans - flow equatorward. Further evidence for the Leeuwin Current from ship drift observations, research vessel surveys, and biological data sets is reviewed by Church et al. (1989) and Batteen & Rutherford (1990). In addition, analyses of the data from the Leeuwin Current Interdisciplinary Experiment (LUCIE) in 1986/87 are proceeding with some of the first findings being reported by Smith et oi.(1991). In the following sections we outline the features of the Leeuwin Current and the results of several conceptual and numerical models. Features of the Leeuwin Current Oceanic scale It is instructive to take a large scale view of the salinity west of Australia in March along a line several hundred kilometres offshore running from Java to Antarctica (Fig. 6). Data collected by Deacon on RRS Discovery in 1936 and by Rochford on HMAS Gascoyne in 1963 have been combined - both were presented by Wyrtki (1971). The section shows a number of interesting features: • Near Java, there is low salinity water from river runoff and throughflow from the Pacific Ocean that tapers away to the south. Beneath the surface plume and down to the depth of the sill in the Timor Trench (about 1400m) are the near-constant salinity waters of the Banda Sea where they have had a residence time of some tens of years. • Near Western Australia the excess of evaporation over precipitation produces dense salty South Indian Central Water that sinks and slowly moves northward. • In the Southern Ocean, precipitation and ice melting result in a sinking plume of cold, fresh Antarctic Intermediate Water that also flows with a northward component. • On the Antarctic continental shelf, where winter freezing excludes salt, there is a cold salty plume that sinks and moves northward as Antarctic Bottom Water. In summer, the surface water near Antarctica is warmed and it is diluted by ice melting. This produces the Antarctic Surface Water, which also moves northward. • The southward-flowing Deep Water, which has its origins in the South Atlantic, replaces those northward flowing waters. So then, where is the Leeuwin Current? To see the current one must move closer to the continent - in autumn and winter when it flows strongest. The next diagram (Fig. 7) concentrates on the region from just north of NW Cape down to Cape Leeuwin - a voyage by HMAS Diamantina in August 2 Journal of the Royal Society of Western Australia, 74,1991 Figure 2 A current chart from Halligan (1921) showing a warm surface current flowing down the western Australian coast and a cold northward flowing current which sinks beneath it. Figure 3 A current chart for the Australian region in winter (August-September) taken from a global chart by Schott (1935). The Leeuwin Current is readily apparent. 3 Journal of the Royal Society of Western Australia, 74, 1991 Figure 4 The position occupied by the raft' of warm water, originally from the throughflow, in February, followed by its progression to the south as the Leeuwin Current (from Gentilli 1972). Figure 5 The tracks of the six drifters that were influenced by the Leeuwin Current or its associated cyclonic eddies (numbers 1-5) or the anticyclonic eddy (number 7) during the period March 23 to June 3, 1976 (from Cresswell & Golding 1980). 10°S 20°S 30°S 40°S 50°S 60°S Figure 6 A salinity section looking towards Australia from several hundred kilometres offshore between Java and Antarctica. The data were collected by RRS Discovery in 1936 and HMAS Gascoyne in 1963 and presented by Wyrtki (1971). They were composited for this diagram. 4 Depth (m) Journal of the Royal Society of Western Australia, 74, 1991 Figure 7 The salinity structure at the shelfedge and down the continental slope from NW Cape to Cape Leeuwin showing the low salinity Leeuwin Current and the Undercurrent, which carries South Indian Central Water and Subtropical Oxygen Maximum waters northward (from a cruise by HMAS Diamantina in August, 1971). Current vatoetty (cm •“') Current velocity (cm !“’) -20 -10 0 10 20 -20 -10 _ 0 _ 10 20 -20 -10 0 10 20 -20 -10 0 10 _ 20 0 0 0 0 _ 40< E. \ \ \ \ 40 ( ' s \ - 40- £ V / □ 40- M □a £ a 2 80 V / / / 4 80 \ 1 1 1 £ a 2 80- \ / SO /! 120 / > _ / S. Mar. 82 _ 120 / Apr. 82 120- / / / * Sapt.82 □ 120- /< _tl_ Oct 82 a - -20 -10 0 10 20 -20 -10 _ 0 _ 10 20 .20 -10 0 10 20 -20 -10 _ 0 10 _ 20 0 0 0 0 O a- □ • i 40 / 40 \ \ \ \ 40 \ S 40- \ \ 0 ""a □ • 1 a □ a 80 \ / ✓ / / • 80 / a / / / 80 J \ \ 80 \ j 120 \/ / / May 82 120-1 1 / Jun« 82 120- _J Nov. 82 120 l •-• OnaXore □-Q AK in p«Xor« Figure 8 Vertical profiles of monthly average onshore and longshore currents measured at North Rankin (19° 35'S, 116° 05'E; water depth 123 m) on the southern NW Shelf in 1982. Negative alongshore currents represent southward flow and hence contribute to the Leeuwin Current (from Holloway & Nye 1985). 5 Journal of the Royal Society of Western Australia, 74, 1991 Figure 9 Satellite tracked drifters near and offshore from the Holloway and Nye (1985) moorings in 1982 and 1983 (see Figure 8). 6 Journal of the Royal Society of Western Australia, 74,1991 1971. The diagram shows the salinity structure at the shelf edge and down the continental slope. The Leeuwin Current is the low salinity, warm wedge that extends southward to 30°S. Beneath it is the undercurrent sliding northward carrying high salinity South Indian Central Water as well as oxygen rich waters (not shown in this diagram). As the Leeuwin Current progresses southward it cools and becomes saltier because of evaporation and mixing with the South Indian Central Water. Source What is it that takes place at and south of the source area off NW Australia in autumn? Current meter data for 1982/83 (Fig. 8) revealed a contribution to the Leeuwin Current by NW Shelf waters (Holloway & Nye 1985). It was strongest from February to June reaching 0.2 ms’ 1 in March-April-May in 1982. Satellite tracked drifters near and offshore from the moorings in 1982 and 83 (Fig. 9) revealed an interesting seasonal behaviour: From November to late March the situation could best be described as quiescent, with the drifters moving slowly (-0.1 ms' 1 ) and following no consistent path. At the start of April, however, there was a dramatic move poleward at speeds of about 0.2 ms' 1 on the shelf and up to 0.3 ms' 1 off the shelf. A drifter that rounded NW Cape in late April accelerated to 0.5 ms' 1 . May through August saw predominantly poleward flow with an increasing tendency for the region to feed the South Equatorial Current. West coast Off NW Cape the warm low salinity source of the Leeuwin Current is broad and shallow (400 km by 50 m), but in running southward it tapers to less than 100 km and deepens to more than 100 m, while having speeds that can exceed 1 knot. The warm low salinity waters carried by the Current (Fig. 10) are commonly encountered just beyond the continental shelf edge, but they can spread half way to the coast, except in summer when a wind-driven high salinity northward flow occupies most of the shelf. Incidentally, in summer there is a region of strong shear on the outer shelf between the northgoing shelf waters and the southgoing waters further out to sea (Cresswell & Golding 1980 - their Fig. 7). A number of current meters moored mid-shelf off Dongara and Rottnest Island (near Perth) in the mid 70s (Fig. 11) indicate the influence of the Leeuwin Current flowing southward at ~0.2 ms' 1 (Cresswell et al. 1989). At all times of the year the mid-shelf currents were strongly influenced by passing weather patterns (Fig. 12), which resulted in current variations of up to 0.5 ms' 1 and sea level changes of about 30 cm. In summer, atmospheric troughs from the north interrupted the northward wind stress and allowed the shelf waters to move south. In winter, the passage of lows near and south of Cape Leeuwin gave rise to strong northwesterly winds that augmented the southward Leeuwin Current flow on the shelf. South coast The Current appears to take on a new character once it rounds Cape Leeuwin. It enters a regime where its salinity is higher than ambient, rather than the reverse as is the case west of the continent where it flows through high salinity South Indian Central Water. In the second (fresher) regime the Leeuwin Current has been observed to carry with it a sheath of salty South Indian Central Water. The sheath is then slowly lost downstream through energetic mixing with the fresher offshore waters (Cresswell & Peterson unpubl.). It is just beyond the shelfedge between Cape Leeuwin and the Bight that the Leeuwin Current reaches its greatest speeds of more than 3 knots, or 1.5 ms -1 (Fig. 13). It is quite narrow, a band less than 20 km wide contains the speeds exceeding 1 knot (0.5 ms -1 ). Across the shelf the currents range down from 0.5 ms' 1 . The offshore edge is marked by a temperature front of several degrees Celsius. Temperature 26°S Salinity 26°S Figure 10 A section at 26°S made by HMAS Diamantina in August 1971 showing a) temperature and b) salinity. 7 Journal of the Royal Society of Western Australia, 74,1991 A B & H D E & I R & J N Figure 11 Current meter measurements from mid-shelf sites near the Abrolhos Islands (except for R which was near Rottnest Island off Perth) arranged according to site and time. The stick vectors are ten-day averages. The stippled period, from March to August, shows the period of strongest southward flow, probably the result of the Leeuwin Current spreading onto the shelf (from Cresswell et al. 1989). There are places where the current breaks out to sea (Cresswell & Golding 1980, Griffiths & Pearce 1985) and in June 1987 it was found that one of the offshoots had a trough-shaped cross section 50 km wide by 150 m deep with southgoing flow on the western side and northgoing flow on the eastern side (Cresswell & Peterson unpubl.). Many features of the offshoots have been reproduced in rotating laboratory tanks (Condie & Ivey 1988). Some attempts to explain the current Thompson (1984, 1987) concluded that the Leeuwin Current is driven into the prevailing wind by a longshore sea-level gradient and further that the wind forcing effects are diminished by deep mixed layers. In other words, the force of the northward wind is distributed over considerable depth and is therefore less effective in retarding the southward flow. Godfrey & Ridgway (1985) examined the annual cycles of sea level and wind stress (Fig. 14). The annual average sea level shows a southward flow component near the coast, as indicated by the orientation of the sea level contours, which increases in strength southward from NW Cape. The annual average wind stress is strongly northward between Cape Leeuwin and NW Cape. (Further north, the wind has a strong component to the west and this may explain why drifters in the near-surface layer moved off in that direction to join the South Equatorial Current.) However, the wind stress eases between March and October to reinforce the effects of the peak in sea level difference at the shelf edge between NW Cape and Cape Leeuwin from February to August (Fig. 15). Incidentally, during LUCIE in 1986/87 the seasonal variation in the strength of the Leeuwin Current seemed to be the result of variations in the wind stress and not in the alongshore pressure gradient, which had little seasonal dependence (Smith et al. 1991). 8 1 030 Journal of the Royal Society of Western Australia, 74, 1991 (qw) ajnssajd ouaqdsoiuiv ( j.SUl) pUjM 8J0LjS6u0|V (quj) ajnssajd ouaqdsoiuiv ( j.suj) pu|M ajoqs6uo|V 9 Journal of the Royal Society of Western Australia, 74, 1991 100 - Q- d) Q 150 - 200 -I 20 km Figure 13 A section (bottom panel) showing the current structure measured by RV Franklin along a line out from the coast west of Albany in June 1987. Current speeds greater than 1.0, 2.0 and 3.0 knots have progressively darker shading. The isotherms measured by expendable temperature probes are shown. Note the surface temperature and salinity front (top panel) on the offshore edge of the Leeuwin Current where the speed dramatically decreases and where considerable overturn and mixing take place. Weaver & Middleton (1989) used the Bryan-Cox General Ocean Circulation Model with seven vertical levels and initial conditions of a latitudinal variation of temperature and salinity. In addition, taking a lead from Gentilli’s (1972) suggestion of a raft of warm water off NW Australia, they took all the water in the triangle east and north of NW Cape to have warm, less saline NW Shelf waters (28.3°C and 34.3ppt at the surface). They also allowed for a shelf that tapered from north to south along the WA coast. After running for 30 days, their model (Fig. 16) reproduced both the Leeuwin Current and the Undercurrent quite well -even to the extent of rounding Cape Leeuwin and flowing to the east. Batteen & Rutherford (1990) developed a ten-layer model of the ocean between NW Cape and Cape Leeuwin using a climatological mean density and (in an alternative approach) included an input of water from the NW Shelf. In the latter case, as time passed the nature of the ocean temperature and sea level evolved from their initial setup in response to earth's rotation, eddy viscosity and bottom stress. After a time step of only ten days a Leeuwin Current had been established (Fig. 17). After much longer, an anticyclonic eddy formed off Perth, while north of it was a weaker cyclonic eddy. Both are reminiscent of the features revealed by the drifter tracks and ship data (Andrews 1977, Cress well 1977). 10 Journal of the Royal Society of Western Australia, 74, 1991 1 10°E. 120° 130° 1 10°E. 120° 130° Figure 14 a) Contours of annual average steric sea level relative to 1300 db. The numbers along the coast give mean sea level at tide gauge locations, b) Annual average wind stress and wind stress curl (from Godfrey & Ridgway 1985). Figure 15 The annual cycle of the sum of the forces which drive and retard the Leeuwin Current (full line). These are the sea level gradient (dashed line) and the wind stress (dotted line) (from Godfrey & Ridgway 1985). 04343-3 11 Journal of the Royal Society of Western Australia, 74, 1991 b b 350 m current after 30 days Surface temperature after 30 days Figure 16 The output from the model of Weaver & Middleton (1989) after it had run for 30 days. The Leeuwin Current can be seen rounding Cape Leeuwin in the representation of both the surface current (left panel) and surface temperature (right panel). The undercurrent can be seen at 350 m (middle panel). Surface temperature and current Figure 17a The initial conditions and selected results, at days 10 and 120, for surface temperature and surface currents from the model of Batteen & Rutherford (1990). The current speeds in the anticyclonic eddy off Perth are about 0.5 ms' 1 . 12 Journal of the Royal Society of Western Australia / 74,1991 Dynamic height in cm Figure 17b The initial conditions and selected results, at days 10 and 120, for dynamic height from the model of Batteen & Rutherford (1990). The current speeds in the anticyclonic eddy off Perth are about 0.5 ms' 1 . Concluding comments Models now reproduce many of the features of the Leeuwin Current, such as rounding Cape Leeuwin, having an undercurrent and generating eddies. However, the challenge of developing a model which will produce the annual and interannual variations of the Leeuwin Current remains. Also, such a model will ideally need to mesh in with models of continental shelf circulation. The immediate future will see the launching of satellites such as ERS-1 and TOPEX/POSEIDON which are capable of measuring sea surface elevation, roughness and winds. This information will be used along with ship, drifter and tidal data to constantly update models, rather than have them rely on climatological means. References Andrews J C 1977 Eddy structure and the West Australian Current. Deep-Sea Res 24:1133-48. Batteen M L & Rutherford M J 1990 Modeling studies of eddies in the Leeuwin Current: The role of thermal forcing. J Phys Oceanogr 20:1484-1520. Church J A, Cresswell G R & Godfrey J S 1989 The Leeuwin Current. In: Poleward Rows along Eastern Ocean Boundaries, (eds S Neshyba, C N K Mooers & R L Smith). Springer-Verlag New York Inc 374 pp. Condie S A & Ivey G N 1988 Convectively driven coastal currents in a rotating basin. J Mar Res 46: 473-494. Cresswell GR 1977 The trapping of two drifting buoys by an ocean eddy. Deep-Sea Res 24:1203-09. Cresswell G R & Golding T J 1980 Observations of a south¬ flowing current in the southeastern Indian Ocean. Deep-Sea Res 27A: 449-466. Cresswell G R, Boland F M, Peterson J L & Wells G S 1989 Continental shelf currents near the Abrolhos Islands Western Australia. Aust J Mar Freshw Res 40: 113-128. Dakin W J 1919 The Percy Sladen Trust Expeditions to the Abrolhos Islands (Indian Ocean). Report I. J Linn Soc 34:127-180. Hinders M 1814 A voyage to Terra Australis. Vol I: 269pp, Vol II: 613pp Nicol London (1966 Facsimile Edition) Library Board of South Australia. Gentilli J 1972 Thermal anomalies in the Eastern Indian Ocean. Nature 238: Physical Sciences 93-95. Godfrey J S & Ridgway K R 1985 The large-scale environment of the poleward-flowing Leeuwin Current, Western Australia: longshore steric height gradients, wind stresses and geostrophic flow. J Phys Oceanogr 15:481-495. 7 6 Griffiths R W & Pearce A F 1985 Instability and eddy pairs on the Leeuwin Current south of Australia. Deep-Sea Res 32:1511-34. 13 Journal of the Royal Society of Western Australia, 74, 1991 Halligan G H 1921 The ocean currents around Australia. J Roy SocNSW 55:188-195. Holloway P E & Nye H C 1985 Leeuwin Current and wind distributions on the southern part of the Australian North West Shelf between January 1982 and July 1983. Aust J Mar Fresh w Res 36: 123-37. Rochford DJ 1969 Seasonal interchange of high and low salinity surface waters off South-west Australia. CSIRO Division of Fisheries & Oceanography Technical Paper No 29. Saville Kent W 1897 The naturalist in Australia. Chapman & Hall London. Schott G 1935 Geographic des Indischen und Stillen Ozeans. VerlagvonC Boysen Hamburg: 413pp. Smith R L, Huyer A, Godfrey J S & Church J A 1991 The Leeuwin Current off Western Australia, 1986-1987. J Phys Oceanogr 21: 323-345. Thompson RORY 1984 Observations of the Leeuwin Current off Western Australia. J Phys Oceanogr 14: 623- 628. Thompson RORY 1987 Continental shelf-scale model of the Leeuwin Current. J Mar Res 45:813-827. Weaver A ] & Middleton J H 1989 On the dynamics of the Leeuwin Current. J Phys Oceanogr 19:626-648. Wyrtki K 1971 Oceanographic atlas of the International Indian Ocean Expedition. US Government Printing Office Washington 531pp. 14 Journal of the Royal Society of Western Australia, 74, 1991,15-33 Homologous peri-oceanic west coast climates in the southern hemisphere J Gentilli Department of Geography, The University of Western Australia, Nedlands, WA 6009, Australia Abstract Comparisons and analyses of similarities and differences between these climates in the three southern continents have revealed further aspects of the uniqueness of the Western Australian peri- oceanic and coastal climate. In 1934 Meinardus published his rainfall map of the world, which clearly showed that the driest areas above the oceans were slanted SE-NW between 20° and 10° S off South America and Africa; at their origin, they encroached over the continental shores. The climatically corresponding area off Australia was (a) further south, between latitudes 25° and 20°S, (b) not slanted, (c) slightly offshore, clearly away from the coast, and (d) much smaller. In winter, low cloud, steady rain and passing showers about 30°S are frequent on the Western Australian coast and practically absent from the western coasts of Namibia and Chile (McDonald 1938). The annual rainfall off the west coast at 32°S is nearly 750 mm in Western Australia, under 300 off Chile and under 200 off Namibia. Near 24°S it decreases to under 200 mm off Western Australia, nearly nil off Chile and perhaps 25 mm off Namibia (Gentilli 1952a). Biel (1929) had already shown that rainfall also varied to a different extent and over very different areas in the three southern continents, the Australian region of high variability being the smallest and more clearly restricted in latitude, whereas the South American area of high variability had great meridional extension and came fairly close to the Equator. Ship observations collated by McDonald (1938) recorded about the same prevalence of southerly winds along the western subtropical coasts of the three continents, but their force was greatest near Western Australia and lowest near Chile, where calms were most frequent. Haze was more frequent near the African coast. World maps of surface air temperatures for January and July (eg in Bliithgen 1966 or Newton 1972) and for each month (NOAA 1984 ff.) show equatorward bending of isotherms along the western coasts of all continents except Australia. Atlases and books contain small rainfall maps, but in general the class intervals between isohyets do not allow adequate intercontinental comparisons. It was well known that dry climates continued equatorwards in South America and Africa, but since the Western Australian coast slants north¬ eastwards and eventually eastwards (Cape Londonderry, its northernmost point, is at 14°S) the pursuit of these and similar comparisons just did not eventuate. Progress in meteorological observations and recordings has made it possible to analyse more climatic data in more sophisticated ways, thus isolating additional significant climatic information. Dynamic climatology, particularly with regard to the frequency, morphology and behaviour of tropical cyclones and the subtropical jet stream, has added to the knowledge and understanding of climatic differences between these homologous coastlines. The relevant factors and events may also originate at great distance, as is the case with El Nino, but are nevertheless very significant - their significance being subject to a considerable time lag. A detailed comparison of the climates of the southern continents' western coasts shows differences which can be traced back to latitudinal extent and vertical configuration. Upper air analysis shows promise and some evidence of meaningful relationships as well as significant regional differences. Most important regional climatic differences however arc due to ocean water circulation and particularly differential surface temperatures. Journal of the Royal Society of Western Australia, 74, 1991 Introduction The term homologous defines the latitudinal limits of this study: west coasts occur in all three continents in the Southern Hemisphere, but the Western Australian coast has the shortest span of latitude (only between 14 and 35°S) and is therefore the maximum common factor. Africa and South America continue further north towards the equator, and South America continues further south into the higher latitudes. The term peri- oceanic has been added in order to include the climate of the air above the coastal waters, distinct from the climate of coastal and interior localities and more closely associated with ocean surface temperatures. Such concepts are illustrated by Fig. 1, which appears after the historical section. The position of the three regions on the globe places each one of them under the influence of the eastern margin of the quasi-stationary anticyclone which dominates the ocean offshore (Alissov 1954), but the degree of this influence is already a valid differential criterion, with western South America rigidly controlled by the Andean barrier, southwestern Africa occasionally swept by continental influences, and Western Australia being more continental than oceanic alternately in the north and in the south at the opposite seasons of the year. Such differences are shown, for instance, in Fig. 6. Early observations: Western Australia Early, isolated observations of these climates did not allow any comparative analysis, but on the other hand were not hidebound by official rules and allowed more scope for keen individual observers. Captain King (1827) made careful weather observations along the Western Australian coast between 1818 and 1822. In 1827 Captain Stirling observed "that the sea-breeze on this coast is usually at S.S.W., and is therefore charged with moisture and very cool; this moderates the action of the sun in summer... It is also remarkable that the [early morning]* land wind blowing from these mountains [presumed higher and snow-capped further inland!] is a cold wind... and from the alternate operation of these two winds, which seldom leave an intervening calm, the air, notwithstanding the sun's great heat, is cool and agreeable, except in spots which are sheltered from the breeze, or during calms" (in Cooke 1905). Some of the earliest climatic observations at Perth, published by Irwin in 1835, were quoted by Miihry (1862): "there are two seasons, the wet and the dry', the former, winter-like, lasts from March until November, with heavy rains only in August and September (should be June and July!]; the height of the rainless summer time is in January... Abundant dew is seen every night." Miihry adds that "the rains must come with northwesterly winds... Summer dew [at times!] condenses notwithstanding the continental easterlies because of the proximity of the sea as a * Square brackets [] indicate words or sentences inserted to clarify the original text source of moisture and the nocturnal cooling under a clear sky." Of 838 mm annual rainfall at Fremantle, 4^^ come in the three winter months (279 in July alone) an^ only 38 in the three summer months. Cooke (1901, 1905) pointed out that the winter "set s in, as a rule, rather abruptly" and gave the dates of th G first heavy winter rains at Perth from 1880 to 190^ "Owing to this tendency for the rain to fall principally i^ heavy showers and at night... the general impression Qf the Perth winter is that of a succession of fine, bright calm days, varied occasionally by a severe but bri^f storm. The weather is, on the whole, delightful, but may perhaps be too mild... The summer does not quif c set in quite so abruptly as the winter. With a n occasional hot day in October, it commences generally in November... On a normal hot summer day a se a breeze always sets in about noon on the coast, an1 mm have been shown for the coastal stations, although exact comparisons are not always possible because of the different definitions of wet day (with precipitation = £1 mm metric, > 0.01 inch earlier British and U.S.A., with no possibility of conversion, affecting Australian and South African data). Simultaneous observations at sea at Greenwich 24 Journal of the Royal Society of Western Australia, 74,1991 noon lead to some underestimate in the African region. However, the contrast between the three continents is most remarkable: Arica (Chile) has an average of 0.1 wet days per year, Swakopmund and Ludcritz (Namibia) 1 day, Mardie and Onslow (Western Australia) 19 and 23 days - mostly supporting Korzun's interpretation of the rainfall pattern. Over the area of coastal waters shown, the number of wet days per year rises only to about 20 off Chile, 30 off South Africa, and 45 off Western Australia. Furthermore, the number of wet days on the coast near 35°S is about 1.5 times its offshore counterpart in Chile, twice its counterpart off South Africa and about 3 times its counterpart in Western Australia. An important regional difference is due to the ’sweep’ of rain-bearing systems. Each kind of system - monsoon, tropical cyclone, epitropical disturbance, extratropical front - has a characteristic way of advancing, while at the same time being partly steered by adjoining pressure patterns. In Western Australia they arrive unaffected by colder ocean surfaces to the west or, if coming from the north-west, strengthened by warm water surfaces, and proceed inland almost unimpeded by orographic factors. In Peru and Chile the Andes intensify any fronts from the Pacific at first impact, but eventually fracture and often destroy them; on the Andean foothills Darwin (1839) observed the extreme localisation of a violent rain storm. Shannon et al. (1989) write that ’'rainfall in 'average' years in the Namib normally occurs as 40-50-km-wide storm cells, and not over broader fronts. Consequently, rainfall events are extremely patchy in their dispersion both in space and time." A good comparison of seasonal rainfall patterns in these regions may be made from Fig. 6, which at first sight appears somewhat forbidding. The average monthly rainfall is shown for three stations in each continent: Perth, Cape Town and Valparaiso in the outer subtropical belt, Vlaming Head, Swakopmund (Namibia) and Antofagasta (Chile) near the tropic, and Cape Leveque, Mossamedes (Angola) and San Juan (Peru) in the intertropical belt. The rainfall is shown on a logarithmic scale to do justice to the highest and lowest amounts as well, and to show rates of variations in their true proportions. In the outer subtropical belt the three cities receive similarly plentiful amounts of rain from May to October, but Valparaiso has a much drier summer. In the epitropical situation, the rainfall of Vlaming Head shows two peaks, due to tropical cyclones in January- March and to epitropical storms in May-June, as well as to fronts of well developed mid-latitude depressions in June and July. By September the anticyclones dominate the weather and rains are negligible. At Swakopmund the only average rains worth mentioning range from 1 mm in November and January to 2 mm in February, March and April. Antofagasta fares even worse, with 2 mm in June and less than 3 mm in July. The most succinct statement of these peculiarities is probably still Hann's (1896): "Remarkable is the scarcity of rain on the western coasts of continents in the subtropical, in part even in the tropical latitudes,... from central Chile and as far as Ecuador, the western coast of southern Africa from the River Orange to as far as Benguela... The dryness of these regions depends on the atmospheric pressure... which brings dominant cool winds from the higher latitudes, and similarly also cool ocean currents, flowing in the same direction, which cool the coast... The stronger and more constant these air and ocean currents..., the more extreme the scarcity of rains..." The role of cool upwelling water was not yet known. In the main, the statistical variability of rainfall is a function of its scarcity, or rather, of the likelihood of rain-bearing systems penetrating a given region; the drier the region, the less often is this likely to happen, hence a higher variability. Biel’s 1929 map is highly generalised; it clearly shows the coastal areas of highest (>30%) variability as extending from about 10 to 35°S in Peru and Chile (La Serena 59%) and from about 12 to 28°S in Angola and Namibia (Walvis Bay 33%). A biologist (Seely 1989) writes that "rain is unpredictable" in the Namib desert. On the Western Australian coast the belt with such high variability runs only from 20°S to the tropic, much of it due to tropical cyclones which cross the coast, Mardie being situated on "cyclone alley". Just outside this limited belt, Onslow has a 26% variability, contrasting with reliable Geraldton's 14% and very reliable Perth’s 8%. Other forms of precipitation and turbidity Important differences between the three regions emerge from an examination of the different types of precipitation and related or in some way similar phenomena. A problem is the counting of days with drizzle, specifically recorded only in northern and central Chile. There arc also vocabulary problems affecting the description of fog-related phenomena in some languages (eg Meteorological Office 1939 and Zimmerschied et al. 1962). Some climatologists (Koppen 1931, Conrad 1936) also refer to possible observer's subjective influences on visibility records during fog, mist and haze. Drizzle (llovizna in Spanish, the drizzling fog popularly known as 'Scotch mist' [but see below under fog]) consists of droplets between 0.2 and 0.5 mm in diameter, ie too small to be counted as rain. The table shows that in northern Chile, where its occurrence is recorded even though it cannot be measured, drizzle (loosely called garua in the Andean region) is by far the most frequent form of precipitation; it is least common at Antofagasta (23°39'S) with only 0.6 days per year. From there it increases slightly northwards, and much more markedly southwards, until south of Coquimbo it gives way to definite and measurable rains. Fog and mist are often confused and the distinction is rather conventional. Mist (with visibility over 1 km) is called neblitia, also bruma , especially at sea, and at times niebla or calina in Spanish, Nebel or feuchter 25 Journal of the Royal Society of Western Australia / 74, 1991 MEAN MONTHLY RAINFALL Figure 6 Mean monthly rainfall at Cape Leveque, Vlaming Head and Perth (Western Australia), Mossamedes (Angola), Swakopmund (Namibia) and Cape Town (South Africa), and San Juan (Peru), Antofagasta and Valparaiso (Chile). Note that the scale of monthly rainfall (in mm) is logarithmic, with the base line set at 1. Dunst in German, nevel in Dutch. It is not recorded separately on the land, where it may be counted as fog with better visibility, eg in the Australian records analyzed by Maine (1968). It is strongly affected by the time of observations. Fog , with visibility under 1 km, is formally called niebla in Spanish, Nebel in German, mist in Dutch. The estimated frequencies of fog from marine observations just offshore are shown in Table 1 as Fog I, those at land stations as Fog II. Fog (or low stratus cloud when it is above the ground) is a major element in the climate of Namibia and Peru and central Chile, contributing a very high relative humidity (daily averages of 75 to 90 % every month) and reducing the amount of sunshine received along a coastal strip some 30 km wide in Namibia, narrower and more broken in Peru and Chile. In Namibia it may penetrate further inland: Brain (1988) during six summer weeks recorded three penetrations of fog some 100 km inland, all within five days. Its main cause is the cooling of the air and the condensation of its moisture content over the colder coastal current or even more so over the upwelling ocean water, or of warmer air advectcd over surface air cooled by the underlying water. It is suggested here that the initial fog may be denser and more persistent where upwelling ocean waters are cooler than the already cool surface waters, but in hot climates partial evaporation of water droplets results in the much finer garua. 26 j Journal of the Royal Society of Western Australia, 74,1991 Table 1. Yearly number of days with precipitation and allied phenomena. Lat. Chile Namibia, S Africa Western Australia S Rain Dr'zle Mist Fogl Fogll Haze Rain Mist Fog Haze Rain Misti Mistll Fogl Fogll Haze 15 5 30 14 7 _ 23 6 5 4 38 25 0 3 0 10 9 20 2 1 13 5 0 27 7 6 15 42 23 0 0 1 2 11 25 7 40 13 7 6 25 2 7 19 37 22 8 0 1 0 16 30 15 45 16 9 28 27 15 8 20 30 25 0 1 0 1 10 35 37 n.a. 17 12 54 26 30 9 11 25 50 0 4 3 5 7 Source: Days with rain, mist, fog I and haze calculated as close onshore as possible from McDonald (1938); days with drizzle and fog II for Chile from unpublished official data (Oficina Meteoroldgica de Chile) averaged from the nearest coastal stations; fog off Africa calculated from Deutsche Seewarte (1944); mist II and fog II in Australia calculated from Maine’s (1968) data for Port Hcdland and Onslow, Carnarvon, Geraldton and Perth. Fog II for 15°S (Beagle Bay and Broome) and 35°S (Cape Leeuwin) estimated from Loewe (1944). See text for consequences of standardised time of observations at sea. In western South America there are terminology problems. Bliithgen (1966) quotes Knoch (1930) who describes the garua as "hovering droplets hardly impeding visibility, with a tendency to accumulate locally, fairly evenly distributed throughout the year”. Trewartha (1961) still describes it as a drizzling fog. Conrad (1936) accepts that garua is clearly distinct from ordinary fog because it behaves differently, "being almost colloidal and - unlike fog - distributed unevenly among unsaturated patches of air". Sekiguchi (1986) defines garua in northern Peru as ’dry fog': "We are driving a car, the sun is shining, visibility is good and distant scenery is clearly seen - yet the wind-shield gets wet. A film of water covers it. We have to operate the wipers. The black asphalt on the road becomes shiny. If we stop the car and touch it, it is wet... No large fog particles are floating in the air... The relative humidity exceeds 80% every day. In the morning, it is more than 90% almost throughout the year..." The problem is due to the difference in temperature between the land in the daytime and the surface of the sea. The atmosphere over the sea is almost saturated with water vapour. If the saturation vapour pressure of this air is 12-22 mb [hPal and this air spreads overland, its relative humidity decreases to 50-60%, because the temperature of the coastal area is about 27-28°[C] and its saturation vapour pressure is 37-38 mb (hPaJ. Therefore, water evaporates from the fog particles floating in the atmosphere and the particles get smaller. On the other hand a quotation in Knoch (1930) described the garua at Lima as a fog copious enough to condense in fine droplets, forming a layer some 50 m above the water and 700 to 800 m thick, driven by the southwesterly wind against the seaward slopes. Lima at about 130 m is right in the fog, while its port quarter of Callao remains clear (Fig. 2). By garua weather the air oozes moisture, umbrellas are quite useless, and the sun may not be seen for months. On the other hand at an altitude of some 800 m one is above the stratus layer and in bright sunshine. Sekiguchi (1986) writes that "the weather of Lima..., because it is enveloped by ordinary fog in the morning the year round, is humid and unpleasant. However, most days the sun shines in the afternoon, the fog turns into garua and the humidity comes down to 60-70%. It scarcely rains... To avoid this strange and unpleasant climate, people of upper class families send their children to the winter resort Chosica 50 km east of Lima in the foothills of the Andes Range to give them more sun and light. The season of this winter resort is from May through October, when there are many cloudy days... During this season the schools are open." In southern Peru and northern Chile the dense fog locally known as camanchaca (but recorded as niebla) is very significant particularly over coastal waters. From about 30°S Chile's long coastal strip is subject to very frequent normal fogs, also recorded as niebla (Fog II in Table 1). Dew , mostly nocturnal and often overlooked by observers, may be the only or the main source of surface moisture in dry regions. From a 1925 paper by Hellmann, Conrad (1936) quotes annual totals of 176 days with dew at Coquimbo and 219 at Los Andes, both in Chile but respectively below and above the inversion layer; the Coquimbo record might be slightly inflated by observations of moisture left behind by early-lifted fogs. Haze (calina or bruma seca, popularly also neblina in Spanish, trockener Dunst, often plainly Dunst but at times also Nebel in German, nevel in Dutch), mostly caused by dust or salt particles, is not a form of precipitation. It is normally carried over these regions by dry offshore winds. The table shows that it is frequent off Chile, and even more frequent off Namibia, which both have very arid coasts and practically no plant cover at these latitudes. Its frequency off Western Australia, where aridity is less extreme, is much lower and shows a definite maximum near the tropic. Seasonality of precipitation At the northern end of the three regions (15°S) rain at sea is more frequent in summer, but the frequency 27 Journal of the Royal Society of Western Australia, 74, 1991 ranges from 12% of the observations in Western Australia (with notable contributions from monsoons and tropical cyclones. Fig. 4) to 6.5 in Namibia and only 1% in Chile, where the Andes are a forbidding barrier (Figs. 2 and 6). About 20°S summer rains are still noticeable in Namibia and Western Australia, whereas they are negligible in Chile. From this latitude southwards winter rains become more frequent in Namibia, and even more so in Western Australia, where they become dominant and reach 15% of the observations. At latitudes 25 to 30°S regional differentiation becomes even more pronounced, with moderately frequent rains being concentrated in the winter in Chile, spread to winter and spring in Namibia, and most abundant from late autumn to early spring in Western Australia. The transition to mid-latitude climates with more uniform rains may be seen at the southern end of each of the three regions where the percentage of rainy days in each season, beginning with summer and with the winter percentage in italics, is about 4+7+22+8% of the observations in the four seasons off Chile, 5+8+7 7+7% off South Africa, and 5+10+2 7+15% off Western Australia. These values may be compared with the monthly averages for coastal land stations in Fig. 4. Mist is not frequent, and may be observed in any season off Chile, more in spring off Namibia, and, though uncommon, mostly in winter off the Exmouth Peninsula (mist 1 in Table 1) and on the Western Australian coast (mist II, usually merged with fog data). Although the theory of its formation is very well known, fog is so closely affected by the conditions of the underlying surface (and some subjective judgment by the observer, for example Koppen 1931, Conrad 1936) that observations present many practical difficulties, including the timing of the observations themselves because fog is so readily dissipated by thermal and dynamic factors, being very frequently a nocturnal and early-morning phenomenon which on land would most often escape the Australian 9 am observing time, and certainly in summer even the South African 7 or 8 am observations. At a land station, even a short period of fog a ny time in the 24 hours should result in a "day with fog" being recorded. Noon at Greenwich (the time of observations at sea) corresponds to about 7:30 h off Chile, 13:00 off Namibia, and 20:00 off Western Australia, leading to a fairly adequate assessment of "days with fog" off Peru and Chile (where Bluthgen (1966) shows over 40 days per year between about 18°S and the tropic, fewer further south), a greatly reduced assessment except in winter off Western Australia, and an almost total exclusion off Namibia, where on the other hand Bluthgen (1966) charts over 80 days with fog per year. The glaring differences between sea and land (Fog I and Fog II) observations in Table 1 show just one example, or perhaps one result, of these difficulties. Fog occurs moderately throughout the year, but more often in autumn and winter, off Chile. It forms more readily onshore and is a semi-permanent feature at altitudes of a few hundred metres on the slopes of the Andes (Fig. 2). It is more frequent in autumn and spring off Namibia (Deutsche Seewarte 1944), but after a specialist workshop at the end of 1988 Shannon et al. (1989) could still write that "even the formation of fog was shrouded by ignorance and clouded by speculation." The same workshop stated that "research on fog formation, movement and chemistry in the Namib-Benguela is necessary in view of the important role which fog plays in the system. Even at a very basic level, the contribution that upwelling makes in fog generation and in the maintenance of the aridity of the Namib needs to be established." In Western Australia fog is uncommon, and at sea almost exclusively an autumn phenomenon (summer and early autumn according to the very detailed maps by the Koninklijk Nederlands Meteorologisch Instituut 1949). Onshore the very infrequent fogs occur mostly in winter and early spring (Maine 1968); there is a secondary peak in February at Carnarvon, March at Onslow, Geraldton and Perth, and April at Port Hedland. At Geraldton winter has few days with fog, and the main peak is in September-Octobcr. Haze, generated by dry conditions and therefore affected by the timing of observations in almost the opposite way to dew, is recorded more frequently in the summer off Chile and Western Australia and practically throughout the year off Namibia, with a maximum in the autumn. Shannon et al. (1989) advocate research into its contribution of minerals to offshore waters. Off Chile the distribution, apart from the summer, is relatively even throughout the year, whereas in Western Australia the frequency and abundance of rains prevent the mobilization of dust once the summer drought is over. However, Keough (1951) gives an annual average of about 3 dust storms for Geraldton (adding that "dusty conditions are more frequent in late summer with NE to SE winds") and about 5 for Port Hedland. At Carnarvon "dust storms occur occasionally and dust often accompanies the onset of SSW breeze". At Kalgoorlie "dust storms occur on 13 days per year." The recording of the far more frequent dust haze is obviously inadequate, as is that of smoke haze in the South West. Upper air correlations A pilot study of some upper air comparisons, based on NOAA's Climatic Data for the World, was shown in Fig. 3. A very significant factor in the formation of rain is the dew point, of which upper-air cross-sections are given in Fig. 7. It should be noted that Australian observations do not extend above 500 hPa. Regional differences stand out very clearly. Above Perth there is a gradual transition from 2°C at the surface in June and early July to 15° or 20°C at 700 hPa (about 3000 m). Above Cape Town a surface average of 28 Journal of the Royal Society of Western Australia, 74,1991 hPa A small part of the results is given in Fig. 8, which shows correlations (R 2 ) between each station's actual monthly rainfall and the upper air conditions or movements for the same month, for the month before, and for two months before. To the left are the correlations with upper air temperature, to the right those with geostrophic wind speed at the various levels. CocbUR*) with Hindgwad u (mca) from 5° to 10° and at times 15°C above Cape Town, and rapid change from 2° to 15°C (from below to above the inversion) above Valparaiso. less than 2°C lasts from March to early September, but in altitude temperatures above 20° are far more prevalent, both vertically and seasonally, than above Perth. Above Valparaiso a marked atmospheric stratification is evident, with isotherms running parallel or nearly so at most heights: the surface dew point is below 2°C from February to mid-October, while dew points above 20° are mostly confined to the 700 hPa band. VALPARAISO (Quintero eoundin®*) Figure 8 Correlation (R 2 ) between average monthly rainfall and upper air temperatures (diagrams to the left) and wind speed (diagrams to the right) above Perth, Cape Town and Valparaiso. The most significant difference between the three stations is in the stratification below 1000 m (well below 850 hPa): as the figure shows, this stratification is non¬ existent above Perth, continuous but with a range of some 5°C above Cape Town, and even more solid, with a range of some 10° to 13°C, above Valparaiso. This is evidence of almost permanent atmospheric inversions, particularly above Valparaiso, which prevent convective activity - thunderstorms and hail have hardly ever occurred at some Chilean stations. Upper air and rainfall relationships A preliminary survey was undertaken of correlations between conditions and movements of the upper air and monthly average rainfall at the underlying stations. At substratospheric heights (200 hPa) above Perth warmer air is closely and positively associated (R 2 >0.7) with the rainfall of the same month; this association may be forecast to some extent (R 2 >0.3) on the basis of the previous month's 200 hPa temperature. A similar association is much weaker above Cape Town, and non¬ existent above Valparaiso. The three regions show a similar but negative relationship (the colder the air above, the greater the rainfall below) at lower levels, but the actual levels of greatest significance differ for each station. Above Perth the closest association (R 2 >0.8) is found at 500 hPa; it persists downwards, only slightly reduced, and was also noticeable a month earlier. Above Cape Town 29 Journal of the Royal Society of Western Australia / 74, 1991 a negative correlation between temperature and underlying rainfall (R^ > 0.7) is found from 500 to 300 hPa, a higher and thicker band; it also extends downwards with only a slight loss of closeness (R 2 > 0.6), and is also noticeable the previous month. Above Valparaiso the correlation is still negative, but conditions are quite different: R 2 >0.6 is found at 200 hPa but the closeness of the correlation decreases steadily downwards, to become negligible as it reaches the impenetrable inversion below 850 hPa. Other, related differences are noticeable in the relationship between upper air winds and rainfall at the surface. Above Perth between 100 and 300 hPa (16 500 and 9500 m) runs an enormous band of strong westerlies (Fig. 3) which are closely and positively associated with the rainfall below (R 2 >0.6). Above Cape Town the association is slightly weaker and limited to 50 to 100 hPa. Above Valparaiso the correlation, possibly just significant (R 2 >0.3), is restricted to 50 hPa (20 500 m) but begins already in the preceding month; the mechanism of any relationship is worth further investigation. However, no relationship between wind speed and surface rainfall seems to exist at any level below 50 hPa. Regional thermal differences World maps of January and July temperatures, from the masterly pioneer maps by van Bebber & Koppen in Hann (1896) to Bliithgcn (1966), clearly show regional differences, particularly above offshore waters. In January wedges of air below 20°C reach northward beyond the tropic along the coast of Chile (and along the foot of the Andes) and just off Namibia above the upwelling of cold water from Port Nolloth to Walvis Bay (Hoflich 1984). In Chile a core wedge of air below 15°C normally occurs as far as 32°S; there is no counterpart off Namibia. No cool air wedge is shown off Western Australia, on the contrary, surface offshore air above 30° reaches about 20°S (mention of tropical cyclones above). In Angola inland air above 30° just reaches the coast around 18°S; coastal and offshore air off Peru and north Chile stays below 25°C. Differences are simpler and sharper in July: offshore air below 20°C reaches northward to about 18°S in Western Australia, 8°S off Angola and about 3°S off South America. In general, maximum and minimum screen temperatures are 6° or 7°C higher at Cape Leveque (16°24’S, Western Australia) than at Mossamedes (l5°12 f S, Angola) and 7° to 9° higher than at San Juan (15°22'S, Peru), as may be expected from the differences in the adjoining ocean surface temperatures. At Exmouth (24°49’S, Western Australia) mean maximum summer temperatures rise over 16°C above those at Swakopmund (22°4rS, Namibia) and 12° above those at Antofagasta (23°26’S, Chile), but by winter both are reduced to some 4°C. Summer minima are about 6° higher at Exmouth than at Antofagasta, where in turn they are some 2° higher than at Swakopmund. Similar differences are found in winter minima. Further south, Perth summer maxima are only some 3° above those at Cape Town, but about 8° above those at Valparaiso; summer minima are 2° higher than Cape Town's and over 5° higher than Valparaiso's. In winter travelling anticyclones bring frequent, if transient, inversions over Perth, reducing the differences from the permanent inversions of southwest Africa and Chile. Winter maxima at Perth and Cape Town are very similar, and those at Valparaiso only 2° lower. Winter minima are similar at Perth and Cape Town, and about 1° lower at Valparaiso. In these regions water masses (and very maritime air) can hold the mean daily range at 5 to 7°C, land masses (and very continental air) can send it soaring to 15 or 18°C. Air circulation therefore determines the season or the month when the greatest or smallest range occurs. Thus at Cape Leveque the narrowest range occurs during the strongest onshore air flow in January-February and the widest one during the anticyclonic offshore flow in July-August, whereas at Vlaming Head the narrowest one comes in June (northwesterly epitropical disturbances and extreme reach of southwesterly mid-latitude fronts) and the widest one in September and also in November- December (probably due to phases in the transit of anticyclones). Loewe (1948) showed that in Australia wider diurnal ranges of temperature as well as higher interdiurnal variability of temperature occur during the drier seasons, ie in winter north of the tropic and in summer south of it. On the other hand in western Namibia and South Africa the greater variability occurs in winter and early spring, when the hot dry berg winds descend from the plateau often enough to raise daily maximum temperatures. "In the Namib Desert... the mean monthly temperatures vary little but daily temperatures vary widely and relatively unpredictably..." (Seely 1989). The influences of cool ocean currents, continental geometry and landforms combine in producing another clear regional differentiation: in tropical Western Australia the month with the greatest mean daily range of temperature, 10 to 15°C, progresses gradually from July at Cape Leveque to October beyond Onslow. In the other regions there is no regular time sequence and the range remains smaller: 10°C in May at Mossamedes, 9.5° in April at Lima and 7.8° in January to March at Arica. Oceanic influences and onshore air flow combine in giving the southern reaches of all three regions mean daily ranges of 5 or 6°C in June-July, but only along the South American coast where the flow of cool water along the shore is at its strongest does this pattern continue very far towards the equator. On the southwest African coast, where local cool water upwelling is very effective and the cool current is weaker, the months with the narrowest range vary with each station; in Western Australia the time remains unchanged as far as Shark Bay but daily ranges below 30 Journal of the Royal Society of Western Australia, 74, 1991 6°C are only found on islands, thus confirming the fact that plain oceanic influence is the prime factor in holding thermal ranges small. The delayed annual thermal maximum in March (in April for mean daily maxima) which occurs around Broome in Western Australia, is also found at Mossamedes in Angola, and is probably due to the time of the second transit of the zenithal sun, the position of travelling anticyclones, and the end of the rainy season. On the Peruvian and Chilean coast the warmest time is more diffuse, probably because of the strong ocean current along the shore, but a faint maximum is usually found in February, as is normal and more definite further south in all three regions. Conclusions Evidence so far does not clearly prove any effect of the Leeuwin Current on coastal climates, partly because of deficiencies in the available records of the pairs Abrolhos-Geraldton and Rottnest-Frcmantle. At the ocean-air interface, higher temperatures and levels of humidity may be noticed physiologically. As a stream, the Leeuwin Current is only about 30 km wide (Godfrey & Ridgway 1985), although some of its eddies may reach much further to the seaward. Mean minimum temperatures are higher at Rottncst Island than at Fremantle by 1°C in March, 1.4° in April, 1.8° in May: the April-May rise may be partly due to the arrival of warmer water. This suggestion may be supported by the fact that 9 am relative humidity, relatively stable, shows its greatest monthly increase from March to April: 8% at Rottnest, 16% at Fremantle and 9% at Perth, but on the other hand these changes are even more conspicuous further inland, where no significant influence of the Leeuwin Current could be postulated. Karelsky (1961), having defined ’cyclonicity' as the number of cyclonic centres in a particular area during the month concerned, shows an average cyclonicity of 0.6 in March, 1.3 in April, 1.5 in May and 1.9 in June between 30 and 35°S and 110 and 115°E. On the other hand, most rain is brought to south-western Australia by mid-latitude fronts linked with mid-latitude depressions which form over the southern Indian Ocean south of 45°S and west of 75° or 80°E. A count for 1988-1990 shows rain brought to Perth by nearly 100 weather systems a year, some 84 of them meridional fronts formed far to the south-west and crossing over Perth or further south, but including also a dozen ’subtropical' lows formed hundreds of kilometres due west of Perth. The monthly count of rain-bearing fronts agrees with the average of 9.5 fronts (including non¬ rain-bearing ones) in February and 14 in August shown by Gorshkov el al. (1974). Some rain was brought by coastal troughs and by two tropical cyclones. Cloud formations were already conspicuous in satellite images days before they reached the vicinity of the Leeuwin Current. It is possible that some 'subtropical' lows (more frequent in the autumn) and some troughs, slow-travelling or even briefly stationary, could gather additional vapour from the Leeuwin Current, but their formation and main moisture loading took place hundreds of kilometres away. The most significant and economically important inter-regional climatic differences, particularly in the amount of rain received, are due to the absence of colder water near the Western Australian shore, not to the presence of warmer water: Gorshkov el al. (1974) show an average of 14 fronts reaching the coast near Perth in August, against 4 reaching Chile at the same latitude. As to nor'westerly origin of epitropical storms or cloud bands, the water vapour generated by 2.5 million km 2 of warmer water in the Arafura-Cocos region far outweighs what can be generated by the 25 000 km 2 or so of the Leeuwin Current, even when its surface is greatly increased by meanders, swirls and gyres. Air transiting across the current may be made more unstable, but most of the winter thunderstorms seem to be due to the sudden uplift of the fast windstream by the Darling Scarp (Gentilli 1979b). Immediately above the 'climatic cuticle' affected by the Leeuwin Current is likely to be spreading the mild cooling effect of the much broader West Australian Current, carried by the peripheral winds of the Indian Ocean anticyclone, and partly showing cyclonic divergence long before it meets the western margin of the Leeuwin Current. In principle, it seems desirable to establish regular climatic observations at the Abrolhos and to extend the scope of the Rottnest ones, particularly in the afternoons. It should also be recommended that detailed studies of sea breezes, land breezes and southerly coastal wind be supported as long-term research projects. Evidence of contemporary climatic fluctuations should be monitored regularly. Gentilli (1952b, 1971) showed definite patterns (decrease in the Kimberleys, increase in the South-West) for the period 1881-1940. Pittock (1975) found a reversal of this trend after 1941. Vogel's (1988) research seems to show that in 160 years of Cape Town records there is no indication of climatic change, but his graphs reveal definite and fairly regular fluctuations of floods and droughts. Perhaps there is a need for some well-publicised agreement on definitions of climatic singularities, variations, periodicities and fluctuations, and long-term climatic changes, while trying to avoid the stereotyped rigidity that made the concept of 'climatic normal' almost meaningless to research. References Alissov B P 1954 Die Klimate der Erde. German transl of Klimaticheskie oblasti zarubezhnikh stran VEB Deutscher Verlag der Wissenschaften Berlin. Biel E 1929 Die Veranderlichkeit der Jahressumme des Niederschlages auf der Erde. Geogr Jahresber a Osterr 14/15:151-180. Bliithgen J 1966 Allgemcine Klimageographie. 2nd ed De Gruyter Berlin. Brain C 1988 Water gathering by baboons in the Namib Desert. S Afr ] Sd 84:590-591. 31 04343-4 Journal of the Royal Society of Western Australia, 74, 1991 Breuer A 1974 Die Bewolkerungsverhaltnisse des sudhemispherischen Sudamerika und ihre klimageographischen Aussagemoglichkeiten. Geographisches Institut der Universitat Bonn. Bureau of Meteorology 1956 Qimatic averages Australia. Meteor Summary. Chun C 1903 A us den Tiefcn des Weltmeeres. Fischer Jena. Coleman F 1971 Frequencies, tracks and intensities of tropical cyclones in the Australian Region 1909-1969. Bureau of Meteorology Melbourne. Conrad V 1936 Die klimatologichen Elementen und ihre Abhangigheit von terrestrischen Einflussen. Allgemeine Klimalehre vol 1 Part B (ed W Koppen & R Geiger) Handbuch der Klimatologie Borntraeger Berlin. Cooke W E 1901 The climate of Western Australia from the meteorological observations made during the years 1876-1899. Government Printer Perth. Cooke W E 1905 Climate. In: The Western Australian Official Year Book (ed MAC Fraser) pp 175-208 Government Printer Perth. Council for Scientific and Industrial Research 1933 Meteorological data for certain Australian localities. Pamph 42. Darwin C 1839 Journal of researches into the Natural History and Geology of the countries visited, in: Narrative of the surveying voyages of His Majesty’s Ships Adventure and Beagle between the years 1826 and 1836. (ed Fitzroy R) Calbum London Vol 3. Deutsche Seewarte 1944 Monatskarten fur den Siidatlantischen Ozean. Deutsche Seewarte Hamburg. Downey W K, Tsuchiya T & Schreiner A J1981 Some aspects of a northwestern Australian cloudband. Aust Met Mag 29:99-113. FitzRoy R (ed) 1839 Narrative of the surveying voyages of His Majesty’s Ships Adventure and Beagle between the years 1826 and 1836. 3 vols Calbum London. Gentilli J 1952a A Geography of Climate. University of Western Australia Crawley WA. Gentilli J 1952b Present climatic fluctuations in Western Australia. W Austr Natural 3:155-165. Gentilli J 1969 Some regional aspects of Southerly Buster phenomena. Weather 24:173-180. Gentilli J 1971 Australian climatic factors; Dynamics of the Australian troposphere; The main climatological elements: Qimatic fluctuations. Ch 4-7 (pp 35-211) and climatic tables for Australia (pp 269-371) in: Qimates of Australia and New Zealand (ed J Gentilli) World Survey of Climatology vol 13 Elsevier Amsterdam. Gentilli J 1972 Thermal anomalies in the eastern Indian Ocean. Nature Phys Sd 238:93-95. Gentilli J 1973 Epitropical climatology of the subtropical jet stream. Bonn Meteor Abh 14:31-79. Gentilli J 1979a Epitropical westerly jet advected storms. Queensland Geogr J (3rd ser) 5:1-20. Gentilli J 1979b Scale and some vertical dimensions of the environment W Geogr 4:9-14. Godfrey J S. & Ridgway K R 1985 The large-scale environment of the poleward flowing Leeuwin Current, Western Australia: longshore steric height gradients, wind stresses and geostrophic flow. J Phys Oceanogr 15:481 - 508. Gorshkov C G et al 1974 Atlas okeanov: Tikhii Okean. Ministerstvo Oborony SSSR Moskva. Hann J 1896 Die Erde als Ganzes - ihre Atmosphare uncf Hydrosphare. 5th ed Tempsky Wien. Hann J 1908-11 Handbuch der Klimatologie 3rd ed 3 vol$ Engelhom Stuttgart. Hettner A 1930 Die Klimate der Erde. Teubner Leipzig & Berlin. Hoflich O 1984 Climate of the South Atlantic Ocean, in: Qimates of the Oceans (ed H van Loon ) World Survey of Qimatology vol 15 Elsevier Amsterdam. Johnson A M 1976 The climate of Peru, Bolivia and Ecuador. Ch 4 In: Qimates of Central and South America vol 1^ World Survey of Qimatology (ed W Schwerdtfeger) Elsevier Amsterdam, 147-218 Johnston A K 1848 The physical atlas. Blackwood Edinburgh & London. Kaemtz L F 1843 Cours complet de m6teorologie. Transl by C Martins from the German ed of 1840 Paulin Paris. Karelsky S 1961 Monthly and seasonal anticyclonicity and cydonicity in the Australian region - 15 years (1946. 1960) averages Bur ofMeteorol Meteorol Study 13. Keough J J 1951 Selected Australian climatic data. Commonwealth Experimental Building Station Tech Study 36. King P P 1827 Narrative of a survey of the intertropical and western coasts of Australia performed between the years 1818 and 1822. 2 vols Murray London. Knoch K 1930 Klimakunde von Sudamerika Part G of Amerika vol 2 of W Koppen & R Geiger (ed ) Handbuch der Klimatologie Borntraeger Berlin. Koninklijk Nederlands Meteorologisch Instituut 1949 Sea areas around Australia. K N M I De Bilt. Koppen W 1931 Grundriss der Klimakunde. de Gruyter Berlin & Leipzig. Korzun V I, Sokolov A A, Budyko M I et al. 1974 Atlas mirovogo vodnogo balansa Gidrometeoizdat Moskva and Leningrad. Kuhnel I 1990 Tropical-extratropical cloudbands in the Australian region. Intern J Climat 10:341-364 Loewe F 1944 Coastal fogs in Australia. Commonwealth Meteorological Bureau Bull 31. Loewe F 1948 A contribution to the study of the variability of meteorological elements in the subtropical and temperate latitudes of the Southern Hemisphere. Part 3 of Commonwealth Meteorological Bureau Bull 39. Lorenz J R & Rothe C 1874 Lehrbuch der Klimatologie. Braumuller Wien. Lourensz R S 1977 Tropica! cyclones in the Australian region July 1909 to June 1975. Met Summ [no numberl. Lydolph P E 1957 A comparative analysis of the dry western littorals. Ann Assoc Am Geogr 47:213-230. McDonald W F 1938 Atlas of climatic charts of the oceans. US Dept of Agriculture Weather Bureau. Washington D.C. McGee O S 1988 Precipitable water and surface dewpoint relationships in different airflow conditions at Cape Town. S Afr J Sd 84:64-65 Maine R 1968 Restricted visibility statistics - selected Western Australian stations. Meteor Summary Bureau of Meteorology Mdboume. MatsumotoJ 1989 The seasonal changes of tropical cloud distribution as revealed from 5-day mean outgoing longwave radiation. Jap Progress in Qimatol Dec 1989 pp 18-34. 32 Journal of the Royal Society of Western Australia, 74, 1991 Meinardus W 1934 Einc neuc Niederschlagskarte dcr Erde. Pctcrm Geogr Mitt 80:141-143. Meteorological Office (Great Britain) 1939 Meteorological Glossary 3rd ed HMSO London. Miller A 1976 The climate of Chile. Ch 3 In: Climates of Central and South America vol 12 of World Survey of Climatology (ed W Schwerdtfeger) Elsevier Am s terd am, 113-145. Milton J D 1978 The rainfall from tropical cyclones in Western Australia. Gcowest No 13. Muhry A 1862 Klimatographische Uebersicht der Erde in einer Sammlung authentischer Bcrichte mit hinzugefugten Anmerkungen. Winter Leipzig & Heidelberg. Newton C W ed 1972 Meteorology of the Southern Hemisphere. Meteor Monogr 13 No 35. NOAA (U S National Oceanic and Atmospheric Administration) 1984 ff Monthly climatic data for the World. Pittock A B 1975 Climatic change and the patterns of variation in Australian rainfall. Search 6:498-504. Ramage C S 1984 Climate of the Indian Ocean North of 35°S. In: Climates of the Oceans (ed H van Loon ) World Survey of Climatology vol 15 Elsevier Amsterdam. Samoilenko V S ed 1966 Meteorologicheskie usloviya nad Tikhim Okeanom. With Appendix of 79 climatic maps Nauka Moskva. Schott G 1932 The I lumboldt Current in relation to land and sea conditions on the Peruvian coast. Geography 17:87-98. Schott G 1935 Geographic des Indischen und Stillen Ozeans. Boysen Hamburg. Schott G 1942 Geographic des Atlantischen Ozeans 2nd ed Boysen Hamburg. Schultze B R 1972 South Africa. Ch 15 In: Climates of Africa World Survey of Qimatology vol 10 (ed J F Griffiths) Elsevier Amsterdam, 501-586. Seely M K 1989 Desert invertebrate physiological ecology: is anything special ? S Afr J Sci 85:266-270. Sekiguchi T 1986 Coastal deserts. In: World-wide weather (ed K Takahashi) Balkema Rotterdam. Shannon L V, Siegfried W R & Ward J D 1989 Namib-Benguela interactions. S Afr J Sd 85:277-279. South African Weather Bureau 1957-58 International Geophysical Year (1957-58) World Weather Maps Part III Southern Hemisphere South of 20°S. Weather Bureau Pretoria. Streten N A & Zillman J W 1984 Climate of the South Pacific Ocean. In: Climates of the Oceans (ed H van Loon) World Survey of Climatology vol 15 Elsevier Amsterdam. Tapp R G & Barrell S L 1984 The North-West Australian cloud band: climatology characteristics and factors associated with development. J Climatol 4:411-424. Trewartha G T 1961 The Earth's problem climates. University of Wisconsin Press Madison Wis. Tyson P D 1986 Climatic change and variability in Southern Africa. Oxford University Press Cape Town. van Loon H, Taljaard J J, Jenne R L & Crutcher H L 1971 Climate of the Upper Air: Southern Hemisphere vol II Zonal Geostrophic Winds National Center for Atmospheric Research Boulder Colo and National Climatic Center Asheville N C Vogel C H 1988 160 years of rainfall of the Cape - has there been a change ? S Afr J Sci 84:724-726. Walker N D & Lindesay J A 1989 Preliminary observations of oceanic influences on the February-March 1988 floods in central South Africa. S Afr J Sci 85:164-169. Walpole J 1958 Maximum possible rainfall over Australia (Conference on estimation of extreme precipitation), pp 245-271 Bureau of Meteorology Melbourne. Weischet W 1969 Klimatologische Regeln zur Vertikalverteilung der Niederschlage in Tropengebirgen. Die Erde 100:287-306. Zimmerschied W et al. 1962 Meteorological dictionary in six languages part vii of Meteorologisches Taschenbuch vol 1 (ed F Linke & F Baur) Geest & Portig Leipzig. 33 Journal of the Royal Society of Western Australia, 74,1991, 35-45 Eastern boundary currents of the southern hemisphere A F Pearce CSIRO Division of Oceanography, PO Box 20, North Beach, WA 6020, Australia Abstract Traditional atlases show ocean currents forming anti-clockwise gyres in each of the three southern hemisphere oceans, with northwards currents along the west coasts of the continents (the so-called Eastern Boundary Currents). Associated with the equatorwards currents and winds, cool upwelled water is found along the west coasts of southern Africa and South America. However, sea temperature charts reveal that there is in fact a southwards flow off Western Australia, and, despite net equatorwards winds which favour upwelling, there is no large-scale upwelling off this coast. In this paper, the oceanic circulation in each of the Humboldt (Peru/Chile), Benguela (southern Africa) and Leeuwin (Western Australia) current systems is reviewed. Seasonal sea temperature patterns at the same latitudes clearly show the different thermal regimes operating, with important consequences for the marine biota. The main features of the Leeuwin Current system are summarised, with emphasis on the mesoscalc (10's to 100's km) meander/eddy structure as revealed by satellite imagery. Inter-annual variability in sealevel, which is an indicator of the strength of the Leeuwin Current, is linked with El Nino/ Southern Oscillation (ENSO) phenomena. Introduction The West Wind Drift current, driven eastwards around the globe by the strong westerly winds in the Southern Ocean, links the south Indian, Pacific and Atlantic Oceans. In the tropics, the west ward-flowing South Equatorial Currents are driven by southeast trade winds. In consequence, there is an anti-clockwise gyre in each of the three southern hemisphere oceans, with relatively intense southwards (or polewards) currents along their western boundaries and weaker equatorwards flows in the eastern boundary current (EBC) regions (Wooster & Reid 1963). Such is the picture painted by traditional current atlases (see, for example, Tchernia 1980). In the eastern South Atlantic Ocean, there is the equatorwards flow in the Benguela system, while the Humboldt Current system transports cool temperate water northwards off Chile and Peru. Associated with each of these two major current systems is seasonal upwelling of cooler, nutrient-rich subsurface water onto the continental shelf, leading to highly productive waters and rich fisheries. By contrast, oceanographers have known for decades that the situation off Western Australia is different (Crcsswell 1991), with warm water of tropical origin flowing southwards and a distinctive lack of large-scale and persistent upwelling. This paper briefly reviews the broad-scale circulation off the western coasts of southern Africa and South America, and then focuses on the differences between these two '’typical" eastern boundary currents and the anomalous Leeuwin Current. For further information on EBCs in general, the reader is referred to Wooster & Reid (1963) and Neshyba el al. (1989). Large-scale oceanography The difference between the three southern hemisphere EBCs is effectively illustrated by the mean sea surface temperature (SST) charts for summer and winter (Figs. 1 and 2 respectively). The northwards deflection of the surface isotherms off Chile and Namibia indicate both the equatorwards flow of cool water from the south and the upwelling of cold subsurface water. Upwelling in both areas is most intense during the summer months, shown by the closed isotherms along the Peru coast between the equator and 15°S, and between 20 and 35°S off southern Africa (Fig. 1). In winter, the westwards extension of the isotherms along the equator reflects the flow in the South Equatorial Current systems (Fig. 2). Off Western Australia, by contrast, there is a poleward deflection of the isotherms in both seasons (Figs. 1, 2) indicating the southwards advection of warm water, and there is no evidence of upwelling. 35 Journal of the Royal Society of Western Australia / 74,1991 SUMMER Figure 1 Summer (February) sea-surface temperature charts off the western coasts of South America, southern Africa and Australia (after Reynolds 1982). WINTER Figure 2 Winter (August) sea-surface temperature charts off the western coasts of South America, southern Africa and Australia (after Reynolds 1982). Journal of the Royal Society of Western Australia, 74, 1991 Gentilli (1972) illustrated the different thermal regimes in the three areas using SST data derived from ocean atlases. More recent information using the Combined Ocean/Atmosphere Data Set (COADS) enables the zonal surface temperature to be examined on a monthly basis, and again the appreciable differences between the three EBC regions is apparent (Fig. 3; Table 1). In the three areas, the SST at 10 degrees of longitude (about 1000 km) offshore is not appreciably different: about 21 °C in summer and 16°C in winter. Off Africa and South America, SST falls by about 3°C towards the coast as a result of the cool northwards flow and the upwelling. On approaching the Australian coast, however, SST rises by 1.5°C in February, and 3.2°C in August (when the Leeuwin Current is flowing strongly). Indeed, the waters near Western Australia are warmer in winter than those near Namibia and Chile in summer. Figure 3 Gross surface thermal structure at 31 °S off Western Australia, southern Africa and Chile, derived from COADS monthly data in 2 degree squares. Table 1 Comparison of summer (February) and winter (August) SST's (°C) at latitude 31 °S for the three EBC areas, derived from COADS. EBC region February August Benguela 18.4 14.5 Humboldt 17.0 13.0 Leeuwin 22.3 19.2 The subsurface structure is also grossly different. Godfrey & Ridgway (1985) have shown that the isotherms off Western Australia deflect downwards towards the continental slope, indicative of the southwards current, whereas off the other two southern hemisphere west coasts there is a strong upward deflection associated both with the northward current and the upwelling. Primary production (Fig. 4, from FAO 1981) is strong both off Namibia, where production averages over 500 mg carbon nr 2 d*l in the upwelling region, and off Peru, reaching 500 mg carbon m* 2 d- 1 in an isolated patch but generally in the range 250 to 500 mg carbon nr 2 d _1 . In west Australian waters, primary productivity is generally less than half the above values. The abundance of zooplankton mirrors the geographical distribution of phytoplankton (Fig. 5). The Leeuwin Current is again seen to have a much lower biomass than the other two areas, but it is worth noting that there is still some coastal enrichment despite downwelling. The Benguela system The ocean circulation off the west coast of southern Africa has been described in some detail by Shannon (1985), who proposed a conceptual model of the Benguela system — Fig. 6 is a simplified version of the main features of the circulation. There is a broad (order 200 km wide. Nelson & Hutchings 1983) northwards drift of cool surface waters beyond the shelf with speeds of 10 to 30 cm s' 1 (the classical Benguela Current), forming the eastern branch of the South Atlantic anticyclonic gyre. The volume flow in the Benguela Current is estimated to be between 10 and 16 Sv (1 Sv = 10 6 m 3 s' 1 ; Shannon 1985). A shelf-edge jet, first investigated by Bang & Andrews (1974), has been found to extend from south of Cape Town to a point near 31 °S where it turns westward. This is a permanent baroclinic jet which has a width of some 10 km, extends down to 120 m and attains core velocities of 60 cm s' 1 where the shelf is steepest. Satellite-tracked buoys released off Cape Town during summer moved northwestwards and suggested that topographical steering plays an important role in the trajectory of the Current (Nelson & Hutchings 1983). Current speeds deduced from the buoy tracks were up to 35 cm s' 1 in the shelf-edge jets. A branch of the Benguela Current continues to penetrate northwards along the coast as far as about 12°S, but the main body of water diverges from the coast and joins a warm saline flow from the north (the Angola Current) in moving westwards between latitudes 15 and 20°S. There is a frontal (convergence) zone between the cool waters of the Benguela system and the tropical/equatorial water of the Angola Current. The front varies seasonally in position and strength but generally lies between about 15 and 17°S and has a surface thermal gradient of about 4°C per 1° latitude (Shannon et al 1987). Further offshore between 0 and 5°E longitude, there appears to be a northwards meandering jet over the Walvis Ridge. 37 Journal of the Royal Society of Western Australia, 74,1991 >500 250-500 150-250 100-150 <100 mgC/m2/d PHYTOPLANKTON PRODUCTION 20 ° 10 ° o° 10 ° 20 ° 30° 40° 2 1 Figure 4 Phytoplankton production (mg Cm” d' ) off the western coasts of South America, southern Africa and Australia (after FAO 1981). >500 201-500 51-200 <50 mg/m3 ZOOPLANKTON [ ; ; ; ^ ABUNDANCE - Figure 5 Zooplankton abundance (mg m' 3 ) off the western coasts of South America, southern Africa and Australia (after FAO 1981). 38 Journal of the Royal Society of Western Australia, 74,1991 Below the surface there is a deeper poleward "compensation" countercurrent which reaches as far south as Luderitz at 27°S (Hart & Currie 1960); it is generally characterised by low oxygen content. Nelson (1989) has shown that there is in fact an ambient poleward undercurrent of some 5 cm s" 1 flowing along the whole shelf, shelf-edge and slope regions. This current is modulated by barotropic coastal-trapped waves with two to five day periodicity. On the inner shelf, the flow attains speeds of 40 cm s- 1 over short periods during the poleward phase. The destiny of the shallower poleward flow in the Cape Peninsula area, where the shelf is very narrow and the coastline turns westward, is unknown, but at least a part is known to follow the coast onto the Agulhas Bank. A part may retro fleet into the shelf-edge jet. Figure 6 Conceptual model of the Benguela system, simplified from Shannon (1985). AGC = Agulhas Current, ANC = Angola Current, BCC = Benguela Coastal Current, BD = Benguela Divergence, BMC = Benguela Main Current, DCC = Deep Compensation Current, SEJ = Shelf-edge jet. Regions of locally enhanced upwelling are shaded. Solid arrows are surface currents, dashed are subsurface. The dotted line shows the 200 m contour. The prevailing winds are from the south, and hence upwelling-favourable, although both seasonal and alongshore variations in wind stress occur. There are in fact two distinct regimes in the upwelling system, separated at about 31°S: the southerly region experiences a clear seasonal wind-driven upwelling pattern (upwelling is strongest in spring and summer when the northwards wind stress is strongest), whereas north of 31°S the upwelling is more perennial. The upwelling regime consists of a series of localised upwelling cells along the 1700 km of coastline between 20 and 35°S (the shaded near-coastal features in Fig. 6). The Humboldt system The Humboldt (or Peru-Chile) current system off South America is a classical eastern boundary upwelling region, exhibiting the characteristics of equatorward surface flows associated with wind-driven coastal upwelling, high biological productivity and rich fisheries. The upwelling is most pronounced off the Peru coast (4 to 18°S), which has, as a result, been studied in more detail than the Chilean region, and less information is available for the current regime south of 25°S. This review of the circulation in the upper few hundred metres is taken largely from Gunther (1936), Wyrtki (1963, 1966), Zuta (1988) and Codispoti el al. (1989). Authors differ in their conclusions on some aspects (partly a result of the different seasons in which surveys have been undertaken), but this synthesis attempts to summarise the main features of the circulation. Surface currents off this coast are generally towards the north (Fig. 7) as described earlier. The Peru Oceanic Current (POC), which forms the eastern limb of the anti-cyclonic circulation of the South Pacific Ocean, extends to about 700 m depth. North of about 20°S, it diverges from the coast, flowing westwards south of 10°S and entraining or merging with water from the Peru Countercurrent (PCCC) and the Equatorial Countercurrent (SECC) as it returns westwards to form the South Equatorial Current (SEC). Wyrtki (1963) places the eastern limit of the POC off central Peru at about longitude 82°W. It consists largely of Subtropical Surface Water from the South Pacific Ocean, transporting about 8 Sv at 24°S increasing to 14 Sv as it heads westward (Wyrtki 1966). Nearer the coast lies the northwards Peru Coastal Current (PCC), which is shallower (<200 m) than the POC. Its southern and northern limits vary seasonally, but lie approximately between about 33 & 40°S and 5 & 10°S respectively (Gunther 1936; Wyrtki 1963). Wyrtki (1963) considers that it lies east of about 78°W off central Peru; it transports about 6 Sv. Between the two north-flowing currents, the Peru Countercurrent (PCCC) carries warm equatorial subsurface waters southward, commencing at the coast at about 5°S and then flowing almost due south along 80°W. It draws some of its water from the northern extremity of the Peru Coastal Current. It is about 250 km wide and extends down to about 500 m, but its maximum strength is at about 100 m depth. It is not always evident at the surface because of wind-driven surface currents. The PCCC transports about 11 Sv Journal of the Royal Society of Western Australia, 74, 1991 Figure 7 Simplified diagram of the main features of the ocean circulation off western south America, largely after Wyrtki (1966) and Zuta (1988). EN = El Nino Current, PCC = Peru Coastal Current, PCCC = Peru Countercurrent, POC = Peru Oceanic Current, PUC = Peru Undercurrent, SEC = South Equatorial Current, SECC = South Equatorial Countercurrent. Regions of supposedly enhanced upwelling (Gunther 1936) are shaded to illustrate the alongshore variability of upwelling. Solid arrows are surface currents, dashed are subsurface. southwards at 6°S, and weakens to 6 Sv at 15°S and 2 Sv by the time it reaches 22°S. Below the surface near the coast, there is the southward-flowing Peru Undercurrent (PUC), which is separate from the Peru Countercurrent further offshore (Wyrtki 1963, Brockman et al 1980, Huyer el al. 1991). Poleward undercurrents of this nature, flowing counter to the dominant wind, are persistent and important features of most upwelling regions (Smith 1983). At 15°S, the Peru Undercurrent extends across the shelf under a thin (30 m) equatorward wind-driven surface layer, and down the upper slope to a depth of about 250 m and offshore extent less than 50 km from the shelfbreak. Maximum flow occurs between 50 and 150 m depth; the transport is about 1 Sv at 10°S (Huyer et al 1991). The southern limit of the PUC has been estimated by Silva & Neshyba (1979) to lie at about 48°S. Brattstrom & Johanssen (1983) have suggested that there is a Chile Coastal Countercurrent (which flows southwards) and a north-flowing Chile Coastal Current, both inshore of the Peru Undercurrent. These currents. which are present along sections of the Chilean coastline, have not been included in Fig. 7 for clarity. The South Equatorial Countercurrent (SECC) carries some 11 Sv of low-salinity, warm equatorial surface water into the area from the west, as a surface/subsurface flow between about 4°S and 7°S. It deflects southwards at about 85°W, and then at about 15°S returns westwards and merges with the Peru Oceanic Current (Wyrtki 1963). Wyrtki (1966) has no mention of a countercurrent south of the equator, but shows an eastwards Undercurrent carrying 35 Sv along the equator; some of this returns westwards with the SEC, while the rest runs southward along the shelf as an undercurrent. There is an intermittent poleward intrusion of warm, low salinity equatorial water down the coast of northern Peru (Murphy 1936), associated with interannual variations in the central and western Pacific. This nutrient-poor water generally manifests itself at about Christmas time, and has the traditional name of "El Nifto" (or "the Christ-child"). At intervals of between 2 and 10 years, this intrusion is extensive and devastating, raising the temperature of the water by many degrees, lowering the thcrmocline and thus adversely affecting the upwelling process. Plankton and fish die and decompose, birds starve or leave the area, and the commercial fishery collapses. Such events are now known as "ENSO events", as the El Nino is in fact merely one manifestation of a chain of global oceanic and atmospheric phenomena associated with the Southern Oscillation, which is a major reversal of the atmospheric pressure fields in the Indian and Pacific Oceans (Quinn et al 1978, Cane 1983). In some localised regions, upwelling of nutrient-rich waters onto the continental shelf is stronger in autumn and winter (May to September) than in spring/summer (Zuta 1988). Instead of upwelling occurring simultaneously all along the coast, upwelling cells tend to develop with offshore scales of order 10's of kilometres (Smith 1983). Gunther (1936) describes the existence of locally enhanced upwelling cells which appear to be associated with highly variable anticyclonic eddy-like circulations along the coast, as a result of alongshore variations in both topography and wind stress. His observations agreed substantially with those of some earlier investigators, finding stronger upwelling centred at about 28°S, 23°S, 15°S and 7°S (Fig. 7). As pointed out by Wyrtki (1966), however, localised upwelling centres may in fact occur anywhere along the coast. The upwelling zones north of about 12°S are supplied by high salinity, low oxygen water in the poleward countercurrent PUC at depths of order 100 to 150 m; further south, the source of upwelled water is the lower layer of the less saline Peru Coastal Current (Wyrtki 1966). Both of these water masses are drawn up onto the shelf during upwelling events. The difference in the general location of the strongest upwelling centres between Peru and Namibia 40 Journal of the Royal Society of Western Australia, 74, 1991 reflects a similar variation in the strength of the offshore Ekman transport. The Leeuwin system In contrast with the above two classic EBCs, the Leeuwin Current exhibits many unusual features (Figs. 8,9). The general anticyclonic flow which must complete the circuit of the southern Indian Ocean to maintain continuity of flow, takes about 3 years to complete the cycle; surface drift cards suggest that the mean current speed is about 17 cm s' 1 (Shannon el al. 1973). In the southeastern region of the Indian Ocean, however, there is a large alongshore pressure gradient between the warm (low-density) equatorial waters and the cool (high-density) Southern Ocean (Thompson 1984 and Godfrey & Ridgway 1985). This meridional gradient, which is much stronger than that in the other corresponding eastern boundary current regions (Fig. 10), induces a net eastwards geostrophic flow from the Indian Ocean towards Australia, and the flow is then deflected down the pressure gradient by the continent to form the Leeuwin Current (Fig. 8). Figure 8 Schematic diagram of the main features of the Leeuwin Current system. GI = geostrophic inflow from the open ocean, LC = Leeuwin Current, LCS = Leeuwin Current source area, LUC = Leeuwin Undercurrent, SEC = South Equatorial Current. Solid arrows are surface currents, dashed are subsurface. The dotted line shows the 200 m contour. Godfrey & Weaver (1991) suggest one reason that such a large gradient is found off Western Australia and not off any other eastern ocean boundary has to do with the existence of the open Indonesian passages. These result in vertical temperature profiles off the Northwest Shelf being very similar to those in the (very warm) western equatorial Pacific. The warm water supply leads to a loss of heat to the atmosphere and convective overturn of the water column between about 20 and 36°S, and hence to the unique alongshore pressure gradient of Fig. 10. A summer feature, named the "West Australian Current" by Andrews (1977), meanders eastwards between latitudes of about 29 and 31 °S, and deflects southwards offshore of the Leeuwin Current along the continental margin. This current appears to be a branch of the traditional northwards current of the same name which forms the eastern limb of the south Indian Ocean gyre, and it may perhaps be better to retain the name in its original usage. Andrews' current could be termed the "West Australian Summer Current" (Fig. 9). Figure 9 Schematic diagram of mesoscale features of the Leeuwin Current system, derived largely from satellite imagery. The Leeuwin Current itself is shaded, the solid arrows indicating the flow in the warm surface meanders and jets as well as the currents in the cooler offshore waters. WASC = West Australian Summer Current (modified from Andrews 1977). The dotted line shows the 200 m contour. 41 Journal of the Royal Society of Western Australia, 74,1991 Despite equatorward (upwelling-favourable) winds (Godfrey & Ridgway 1984), there is no upwelling off Western Australia. This lack of upwelling is clearly illustrated by comparing the vertical structure of the upper water column on the continental shelf at about 32°S with that in the Benguela area (Fig. 11). In the Benguela, surface temperatures just beyond the shelf- break exhibit the expected seasonal pattern, with summer values of about 19°C falling to 15°C in winter. Closer inshore, however, in water depths of 80 to 150 m, the surface temperature is about 14°C in all months with minimal seasonal variation. Just 20 m below the surface, the temperature is about the same as at the surface during winter, but in summer it falls by almost 4°C as a result of the upwelling. There is, therefore, a reversed seasonal pattern below the surface off southern Africa, the water being warmer in winter than in summer (Fig. 11). Off Western Australia, by contrast, temperature measurements in 55 m water depth show that the column is well-mixed in all seasons, with mean summer temperatures of about 22°C falling to 19°C in winter. Figure 10 Alongshore sealevel difference on the eastern boundaries of the south Pacific, Indian and Atlantic Oceans (adapted from Godfrey & Ridgway 1985). Thompson (1984) reported the existence of an equatorward undercurrent a few hundred metres below the Leeuwin Current off Shark Bay. This countercurrent, which transports high-salinity South Indian Central Water northward and offshore (Fig. 8), is also reflected in the geopotential topography of the 300 m surface relative to 1000 m (Wyrtki 1971). LUCIE The most comprehensive survey of the Leeuwin Current system to date was undertaken in 1986 and 1987, known as the Leeuwin Current Interdisciplinary Experiment (LUCIE). The following brief review draws from LUCIE papers by Church et al. (1989), Weaver & Middleton (1989) and Smith et al. (1991), as well as the important papers by Thompson (1984), Godfrey & Ridgway (1985) and Batteen & Rutherford (1990). It will concentrate on features of the Current south of Exmouth (22°S); the "source area" to the north has been dealt with by Church et al. (1989) and by Cresswell (1991). It should be pointed out that 1986/87 was an ENSO period, so the Leeuwin Current may not have been "typical" at that time (Pearce & Phillips 1988, Smith et al. 1991). The Leeuwin Current is relatively narrow (200 km in the north, narrowing to 50 to 100 km in the south), and shallow (50 m in the north to 200 m in the south) (Church et al. 1989). It flows more strongly during the autumn, winter and early spring months than in summer. Peak current speeds can exceed 1.5 m s' 1 (or 3 knots). Figure 11 Comparison of the monthly mean thermal structure on the continental shelf at 32°S: Benguela and Leeuwin systems. The Benguela (inner) site was in 100 m water and the outer in 360 m depth (data from Buys 1957). The Leeuwin Current data is from 55m water depth. SST represents sea-surface temperature, T20 is the temperature at 20 m depth. The lower diagram depicts the thermal differential between the water surface and 20 m depth. LUCIE results indicated that, near the Abrolhos Islands during the spring months, there was a net southward flow of 20 cm s' 1 in the current core over the upper slope and a northwards undercurrent of 10 cm s' 1 below about 300 m. In autumn, the southwards jet had strengthened to 55 cm s' 1 , and the 42 Journal of the Royal Society of Western Australia, 74, 1991 undercurrent was very much weaker than in spring. By winter, the alongshore flow to the south had moved further offshore but hydrographic measurements showed that it was still flowing very strongly. The flow in the upper 100 m of water just offshore of the shelfbreak near the Islands was remarkably persistent to the south between February and August. The southwards transport in the Current increased from 1.4 Sv in summer to almost 7 Sv in mid-winter. Along the south coast, the Leeuwin Current was located just beyond the shelf-break (Cresswell & Peterson , unpubl.). A warm offshoot some 50 km wide and 130 m thick was transporting the warm water 200 km southwards into the Southern Ocean, the maximum speed being about 1 m s' 1 . There was a westward undercurrent at a depth of 400 to 700 m, with a core speed of about 20 cm s' 1 . Satellite imagery The pioneering satellite work of Legeckis & Cresswell (1981) directly confirmed earlier concepts of the southward transport of warm water and the seasonal nature of the Leeuwin Current. NOAA AVHRR (Advanced Very High Resolution Radiometer) satellite images received in Perth have subsequently shown the complex nature of the Leeuwin Current. Superimposed on the narrow southwards flow of warm, low-salinity tropical water along the shelf-break is a series of wave-like meanders which transport Leeuwin Current water away from and back towards the coast (Fig. 9). Pearce & Griffiths (1991) have shown that the meanders generally develop into cyclonic- anticyclonic eddy pairs: the anticyclonic wing (in particular) grows offshore to a width of order 200 km and may eventually "pinch-off' to form a free-standing eddy. Batteen & Rutherford (1990) have recently modelled the generation of meanders and eddies through both barotropic and baroclinic instability processes. Between the meanders, the Current tends to flow along the shelf-break and upper slope as a jet-like current towards the south, and cyclonic eddies in the ambient water offshore can be associated with current speeds of over 80 cm s' 1 (over 1.5 knots). The meanders do not appear to propagate along the coast. Similar structures occur along the south coast (Griffiths & Pearce 1985a); on at least one occasion, a warm eddy drifted southwards into the Southern Ocean (Griffiths & Pearce 1985b). The western (offshore) boundary of the Leeuwin Current is generally well defined, with a temperature differential of 2 to 5°C between the warm Current water and that offshore, associated with a strong cyclonic shear zone (Cresswell & Golding 1980). The inshore boundary is less clearly defined as the thermal gradient is weaker. Nevertheless, small-scale (order 20 km) billows indicate zones of current shear and active exchange of water between the Current and the shelf water (Pearce & Griffiths 1991). Interannual variability Pearce & Phillips (1988) have demonstrated that annual mean coastal sealevels (which may be used as one indicator of the strength of the Leeuwin Current) fluctuate with ENSO (El Nirio/Southern Oscillation) events. During ENSO years, relatively low coastal sealevels imply a weaker Leeuwin Current, and conversely, in anti-ENSO years, higher mean sealevels indicate stronger southwards flow. Pattiaratchi & Buchan (1991) have extended the analysis to show that coastal sealevels off Western Australia have been related to ENSO events since the turn of the century. Temperature and salinity measurements along the outer shelf off Perth confirm that, during ENSO periods, the water along the outer shelf is relatively cooler and more saline than in anti-ENSO years, indicative of less tropical water being advected southwards by the Current (Pearce & Phillips 1988). Pearce & Phillips (1988) and Phillips et al. (1991) discuss the implications of this interannual fluctuation in flow for larval recruitment of the western rock lobster. Conclusions The Leeuwin Current has been shown to be quite different from the corresponding EBCs of the other two southern hemisphere oceans. Off Namibia and the Peru-Chile region there are cool northward currents, and the upwclling of nutrient-rich water results in highly productive waters on the continental shelf. Off Western Australia, by contrast, the Leeuwin Current transports warm tropical water southwards and (despite upwelling-favourable winds) there is no upwelling. The difference seems to be largely associated with the flow of warm Pacific Ocean water through the Indonesian Archipelago, leading to a much stronger meridional pressure gradient off Western Australia than exists off Africa or South America. Poleward under/countercurrents exist off southern Africa and the Peru-Chile region, but they are comparatively weak in comparison with the Leeuwin Current. Instead, there is an equatorwards undercurrent beneath the Leeuwin Current — this is not found along any other eastern boundary. Interannual variability of coastal sealevels (which may be used as an indicator of the strength of the alongshore flow) is linked with ENSO events, such that the Leeuwin Current is relatively weak in ENSO years and stronger during anti-ENSO periods. Acknowledgements I am indebted to Dr Vere Shannon and Dr Jane Huyer for helpful suggestions on the Benguela and Humboldt systems respectively, and to Dr George Cresswell, Dr Stuart Godfrey and Dr Chari Pattiaratchi for general comments. Richard Litchfield and Bert de Boer carried out the COADS programming. Ms Liz Jefferson drew the diagrams. 43 Journal of the Royal Society of Western Australia / 74, 1991 References Andrews J C 1977 Eddy structure and the West Australian Current. Deep-Sea Res 24:1133-1148. Bang ND& Andrews W R H 1974 Direct current measurements of an oceanic frontal jet in the upwell regime west of Cape Town. J Mar Res 32:405-417. Batteen M L & Rutherford M J 1990 Modeling studies in the Leeuwin Current: the role of thermal forcing. J Phys Oceanogr 20:1484-1520. Brattstrom H & Johanssen A 1983 Ecological and regional zoogeography of the marine benthic fauna of Chile. Reports of the Lund University Chile Expedition 1948- 49, No 49. Sarsia 68:289-339. Brockman C, Fahrbach E, Huyer A & Smith R L 1980 The poleward undercurrent along the Peru coast: 5 to 15°S. Deep-Sea Res 27:847-856. Buys MEL 1957 Temperature variations in the upper 50 metres in the St Helena Bay area, September 1950 - August 1954. S Afr Div Fish Invest Rept 27:113pp. Cane M A 1983 Oceanographic events during El Nino. Science 222: 1189-1195. Church J A, Cresswell G R & Godfrey J S 1989 The Leeuwin Current. In: Poleward flows along eastern ocean boundaries (eds S J Neshyba, CNK Mooers, R L Smith & R T Barber) Springer-Verlag, New York, 230-252. Codispoti L A, Barber R T & Friederich G E 1989 Do nitrogen transformations in the poleward undercurrent off Peru and Chile have a globally significant influence? In: Poleward flows along eastern ocean boundaries (eds S J Neshyba, CNK Mooers, R L Smith & R T Barber) Springer-Verlag, New York, 281-314. Cresswell G R 1991 The Leeuwin Current - observations and recent models. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. AF Pearce and DI Walker). J Roy Soc WA 74:1-14. Cresswell G R & Golding T J 1980 Observations of a south- flowing current in the southeastern Indian Ocean. Deep-Sea Res 27:449-466. FAO 1981 Atlas of the living resources of the sea. FAO Fisheries Dept, 23pp + charts. Gentilli J 1972 Ocean-atmosphere relations on the Western Australian coast. In: Proposal for marine environmental studies off the western coast of Australia (comp L C Brodie-Ilall) CSIRO, 17-37. Godfrey J S & Ridgway K R 1985 The large-scale environment of the poleward-flowing Leeuwin Current, Western Australia: longshore steric height gradients, wind stresses and geostrophic flow. J Phys Oceanogr 15:481-495. Godfrey J S & Weaver A J 1991 Is the Leeuwin Current driven by Pacific heating and winds? Progr Oceanogr 27:225-272. Griffiths R W & Pearce A F 1985a Instability and eddy pairs on the Leeuwin Current south of Australia. Deep-Sea Res 32:1511-1534. Griffiths R W & Pearce A F 1985b Satellite images of an unstable warm eddy derived from the Leeuwin Current. Deep-Sea Res 32:1371-1380. Gunther E R 1936 A report on oceanographical investigations in the Peru Coastal Current. Discovery Rep 13:107-276. Hart T J &c Currie R I 1960 The Benguela Current. Discovery Rept 31:123-298. Huyer A, Knoll M, Paluszkiewicz T & Smith R L 1991 The Peril Undercurrent: a study in variability. Deep-Sea Res. 3^ Suppl: S247-S271 Legeckis R & Cresswell G R 1981 Satellite observations of sea^ surface temperature fronts off the coast of western an4 southern Australia. Deep-Sea Res 28:297-306. Murphy R 1936 Oceanic birds of South America. 2 vols % Macmillan, New York. Nelson G 1989 Poleward motion in the Benguela area. In^ Poleward flows along eastern ocean boundaries (eds S ) Neshyba, CNK Moocrs, R L Smith & R T Barber) Springer-Verlag, New York, 110-130. Nelson G & Hutchings L 1983 The Benguela upwelling area x Progr Oceanogr 12:333-356. Neshyba S J, Mooers CNK, Smith R L & Barber R T (editors) 1989 Poleward flows along eastern ocean boundaries Springer-Verlag, New York. Pattiaratchi C B &c Buchan S J 1991 Implications of long-tern\ climate change for the Leeuwin Current. In: Th^j Leeuwin Current: an influence on the coastal climat^ and marine life of Western Australia, (eds. A F Pearce and D I Walker) J Roy Soc WA 74:133-140. Pearce A F & Griffiths R W 1991 The mesoscale structure of th<^ Leeuwin Current. ] Geophys Res. C%: 16739-16757 Pearce A F & Phillips B F 1988 ENSO events, the Leeuwii^ Current, and larval recruitment of the western rocl^ lobster. J Cons int Explor Mer 45:13-21. Phillips B F, Pearce A F & Litchfield R 1991 The Leeuwii\ Current and larval recruitment to the rock (spiny) lobster fishery off Western Australia. In: The Leeuwin Current: an influence on the coastal climate and marin 0 life of Western Australia, (eds. A F Pearce and D \ Walker) J Roy Soc WA 74:93-100. Quinn W H, Zopf D O, Short K S & Yang R T W K 197§ Historical trends and statistics of the Southern Oscillation, El Nino and Indonesian droughts. Fish Bull 76(3):663-678. Reynolds R W 1982 A monthly averaged climatology of se* • Chloralgal- • - X • ^ Z | extended ; - 35 _ \ Brvomol J \ Chlorozoan C /2 40 7 • chlorozoan \ : ..?; •. • \ - ■ ..•* - • • • • * • # • \ • •• \ Foramol *• \ \ Chlorozoan - % V 30 - ..•• '* , ***. Vtt \Chloralgal - 35 •* **** ( Bryomol - *•« •• • Foramol \ - - 25 \ 30 _i_i___i_ , .i.i.i . . 0 10 20 30 40 0 10 20 30 MAXIMUM TEMPERATURE C C MINIMUM TEMPERATURE °C Legend ♦ Abrolhos (Easter Group) Platform 28 f 45' S ▲ Abrolhos (Easter Group) Shelf 28 c 29'S • Southern Rottnest Shelf 32% S Foramol (Brvomol) Collins (1988) Figure 4 Salinity - temperature annual range diagram pair for southern Rottnest shelf, Abrolhos shelf and Abrolhos platform. Abrolhos platforms (Easter Group). In contrast to the southern Rottnest Shelf, the Abrolhos data plot in a transitional position between temperate (foramol) grain assemblages and tropical/subtropical (chlorozoan) assemblages. The transitional nature of the Abrolhos platforms, as indicated by the salinity and temperature data, is supported by the data on sediment composition and biotic communities (Fig. 3, Table 1). The Leeuwin Current clearly influences the biotic transition zone in the Abrolhos and the resultant sedimentation. Firstly, the presence of the west coast transition zone at 28 - 29.5°S, compared to the east coast transition at 24°S, is probably a direct result of the Leeuwin Current. Secondly, the presence of platform sediment with chlorozoan affinities, surrounded by shelf sediment of rhodalgal type, is a transition zone relationship for which the Leeuwin Current is probably the driving mechanism. Initiation of reef complexes From our present understanding of the geological history of the Abrolhos region it is clear that during the Tertiary the region saw the development of a seaward thickening carbonate wedge, dominated by bryozoan- mollusc-echinoid skeletal calcarenites and calcilutites, and lacking reef-building corals (Fig. 5). Post-Eocene carbonates were deposited as a thin sheet on a stable shelf (France 1985) and are restricted to the top 130 m of Gun Island No 1 well in the Pelsaert Group. Corak reef-related sediments appear to be confined to the* post-Eocene sequence, and the deepest coral recognised in cuttings is from - 67 m (France 1985). Th^ shallowest of the known Tertiary sediments (Upper Eocene at 130 m) are non-reef calcarenites dominated by foraminifers, bryozoans and molluscs (Hawkins 1969). These limited data suggest that the maximum possible thickness of the reef complex is 130 m, and that the reefs postdate open carbonate shelves of Paleocene to Upper Eocene age. There are no data to suggest that reef localisation and initiation were directly controlled by underlying geologic features. But if this is so, then it needs to be asked, why are there no prominent reef complexes between those of the Abrolhos and Ningaloo Reef to the north? The answer may after all relate to some geological substrate control, which in the case of the Abrolhos, provided a location suitable for reef development. France (1985) suggested that coralline algal biostromes similar to those of the southern Rottnest Shelf (see Collins 1988) may have provided suitable substrates for the initiation of the Abrolhos reefs. Cores up to 3 m thick have been recovered from these biostromes, which probably developed on drowned colianite ridges (Collins 1983 Fig. 13). At the Abrolhos, there is one recorded occurrence of pre-existing colianite topography 52 Journal of the Royal Society of Western Australia, 74,1991 MARINE CARBONATE SEQUENCE U. MIOCENE TO QUATERNARY PRE-MARINE COMPOSITE STRATIGRAPHIC TERRIGENOUS SEQUENCE CROSS-SECTION U. CRETACEOUS OUTER ABROLHOS SHELF ,rVv'Ay L \'' x "'i ."■* MIOCENE EOCENE L. CRETACEOUS M.-U. JURRASSIC PALAEOCENE 500m 400 200 100 1 2 3 4 5km NO RECOVERY UNCERTAIN CONTACT Figure 5 Diagrammatic cross-section of the Abrolhos shelf at 29.5°S. After France (1985). underlying reef limestones of probable Last Interglacial age, at Gun Island in the Pelsaert Group. In terms of sea surface temperature control it is not possible to ascertain whether the original development of the coral reef complex was directly due to the presence of the Leeuwin Current or whether this was linked to wider ocean-scale palaeoccanographic conditions suitable for coral reef development. However, it is clear that the persistence of the reef growth during the Quaternary must be linked to the presence of the Leeuwin Current. There are indications in the geological record that the Leeuwin Current only came into existence during the Early to Middle Pleistocene (Kendrick el al. 1991). In the Perth Basin there was a conspicuous change in sedimentation from a Pliocene - Early Pleistocene, essentially siliciclastic suite of sediments, to a later Quaternary, strongly carbonate style of sedimentation. The shift to a predominantly carbonate environment of deposition was accompanied by higher sea surface temperatures than those which prevailed during the Early Pleistocene. Evidence is found in the Middle Pleistocene mollusc fauna, which indicates greater tropical and subtropical affinities. The details of this are discussed by Kendrick el al. (1991). There is only a weak representation of such elements in the mollusc fauna of the older Plio-Pleistocenc units, which suggests that the Leeuwin Current was then cither of lesser importance than during the Middle Pleistocene, or not active. Late Quaternary stratigraphy, reefs and geological evolution The stratigraphy of the Abrolhos platforms is well exposed in coastal sections of the platform islands (Fig. 2) and a composite summary of the lithostratigraphy is given in Fig. 6. Four unconformity- bounded sequences have been identified, three of probable Late Pleistocene and one of Late Holocene age. At present our chronostratigraphic control on the stratigraphy is limited. A number of U-serics dates are available for the lower bindstonc / framestone/ rudstone unit of Fig. 6. These dates give a Last Interglacial age (ca 125 ka BP - Veeh & France 1988 and our preliminary dates) for this unit. This Last Interglacial unit appears to be widespread and dominates much of the geomorphology of the island groups. This conclusion was anticipated by Teichert (1967) who noted that the 100 ka BP data on the coral reef at Rottnest (subsequently revised to ca 125 ka BP - see below), implies that the fossil reefs of the Abrolhos will prove to be of the same order of age. However, we stress the preliminary nature of our results and would not be surprised if our stratigraphic inferences need to be revised once more numerical dates become available. The only other numerical dates that are available at present are for Holocene storm ridge units in the Pelsaert Group which show that these are Late Holocene in age. In addition, an emergent coral-frame fringe present along some islands in the Easter and Pelsaert Groups was dated at Pelsaert Island by U- series to 4.8 ka BP (Veeh & France 1988) and 14 C to 4.2 ka BP. 53 Journal of the Royal Society of Western Australia, 74,1991 The predominance of emergent Pleistocene reefs at the Abrolhos, and thin, discontinuous veneers of Holocene reef overlying this substrate, is in marked contrast to the Great Barrier Reef. Intensive drilling investigations have shown that though Holocene reefs of the GBR are of variable thickness (4-20 m), there is usually a significant buildup of Holocene reef which overlies a buried Pleistocene substrate (Marshall & Davies 1978). COMPOSITE STRATIGRAPHY C C C CALCRETE PROFILE • • • . CALCRETIZED CLASTS UNCONFORMITY Figure 6 Late Quaternary composite stratigraphy of islands in the Abrolhos platforms. Late Quaternary sea levels Over the duration of the Quaternary, global sealevels have fluctuated from around their present height to more than ca 150 m below present sealevel (Shackleton 1987). The details of these fluctuations are complex and regionally variable. A summary of our present understanding of the Pleistocene sealevel history of the Western Australian margin is given by Kendrick et al. (1991), and mechanisms of Late Quaternary sealevel change are discussed by Lambeck (1987) and Lambeck & Nakada (1990). Clearly the growth history of the Abrolhos over the last few hundred thousand years is closely linked to these sealevel fluctuations. Evidence for sealevel changes is widespread and coral reefs have proven to be especially informative, the most spectacular of these being the coral staircase of the Huon Peninsula of Papua New Guinea (Bloom et al 1974, Chappell 1974). In Western Australia the reef complexes fringing the western flank of the Cape Range are also striking indicators of former sealevel events (Van dc Graaf et al 1976, Veeh et al. 1979, Kendrick et al 1991). Fig. 7 shows global sealevel for the period 135 ka BP to present; this time period represents an interglacial- glacial-interglacial cycle during which sealevels for the most part have been well below their present height. Global climates have been characterized by such glacial-interglacial fluctuations for at least 2.5 million years (Berggren et al 1980). Although it is clear that the periodicity of these events will have changed in that time, it is apparent that sealevel for much of these periods will have been well below its present height. Consequently, much of the geological development of the Abrolhos reef complex was linked to sealevel stands below that of present, and there are clear indications of this in the bathymetry of the region. From our present understanding of Pleistocene sealevel events along the Western Australian coast (Kendrick et al 1991), it is clear that sealevel was close to its present height a number of times since the Early Pleistocene. And provided the tectonic controls allowed this, it is possible that elements of the emerged geomorphology of the reefs dates from these events, of which the Last Interglacial (ca 125 ka BP) was the most recent Pleistocene highstand. The Last Interglacial saw global sealevels probably some 5 m higher than present (eg Bloom et al 1974, Chappell 1974, Ku et al 1974, Szabo 1979). Along the coast of Western Australia the Last Interglacial dominates much of the coastal geomorphology. Consequently, it is not surprising that morphostratigraphic units of this age dominate the geomorphology of the Abrolhos reef complex. Coral growth fluctuations There appears to be a sharp contrast between the pattern of development of Last Interglacial and Late Holocene reefs. Coral framestone to rudstone and coralline algal bindstone facies of probable Last Interglacial age are widespread over the three platforms, as tabular developments of reef complex in excess of 10 m thick. Late Holocene coral reef development is largely restricted to leeward reef slopes, walls of "blue-holes", and leeward, more easterly surfaces of platforms, where emergent reef facies underlie prograding storm ridges, composed of coral rubble, as part of elongate, composite islands. Whilst it is tempting to suggest that the thicker and more widespread Last Interglacial reef facies indicate stronger Lecuwin Current activity and reef growth, the contrasting pattern of Last Interglacial/Late Holocene reef development may also be a function of substrate factors. The importance of antecedent topography in controlling subsequent coral reef growth is a classic theme in studies of reef geomorphology (eg Bloom 1974, Guilcher 1988), and in the case of the Abrolhos, the significantly higher Last Interglacial sealevels were a major determinant of Holocene reef growth patterns. However, there are also clear regional-scale data which point to stronger Leeuwin Current activity during the Last Interglacial. 54 Journal of the Royal Society of Western Australia, 74, 1991 Figure 7 Benthonic oxygen isotope record of East Pacific core V19-30 for the past 140 ka and the sealevel record from the Huon Peninsular, New Guinea (after Chappell & Shackleton 1986). The Last Interglacial was a time of widespread coral reef development along the coastal margin of much of the Perth Basin. The best known example of this is the Acropora -rich Rottnest Limestone (Teichert 1967, Szabo 1979, Playford 1988), 20 km offshore. This has a counterpart on the adjacent mainland in an extensive coralline - algal reef limestone with Acropora spp, and other warm water corals, located at the entrance to Fremantle Harbour (Skwarko 1990). These corals occur some 400 km south of their effective modern range limit at the Houtman Abrolhos. Other occurrences of Acropora spp. from that time extend to Augusta and eastwards to about latitude 119 °E ( Kendrick el al. 1991). There arc also clear indications in the mollusc assemblages of that time of distributional shifts, with the assemblages showing a greater tropical affinity; Kendrick el al. ( 1991) discuss this at some length. From the southward extension of temperature- sensitive coral growth, Kendrick el al. (1991) argue that during the Last Interglacial sea surface temperatures along the inner shelf were higher than today, to the extent that in the Perth region inshore surface temperatures were at least 2°C higher than the 16.5°C minimum of today (Pearce et al. 1989). From available evidence (discussed by Kendrick et al. 1991) it would seem that after the Last Interglacial highstand, sealevels along the coast of Western Australia did not again attain their present height until the Holocene. This conclusion corresponds with the generally accepted view of global sealevel in that time (Fig. 7). It is likely that the linear-ridge structures of the Abrolhos fore-reef may represent drowned reefs corresponding in age to relatively high sealevel during the last 120 ka (Fig. 7). Drowned reefs have been widely recognized in the tropical oceans (see Carter & Johnson 1986), and invariably have proven difficult to date. The most recent work shows that at the height of the Last Glacial Maximum (ca 18 ka BP) global sealevel was some 130 m below present (Fairbanks 1989). Whilst the presence of linear ridge - drowned reef forms west of the Abrolhos at depths of up to 115 m (Harris 1989) implies that at the time of their formation, reef development and coral growth were still possible, the age and composition of the structures is unknown, and it is likely that sea surface temperatures were too cold for coral growth during lowstand conditions. From Indian ocean-scale paleoceanographic reconstructions (Prell et al. 1980) it is thought that at the Last Glacial Maximum sea surface temperatures along the coast 55 Journal of the Royal Society of Western Australia, 74,1991 of Western Australia were significantly lower than present and that the cold Western Australian Boundary Current was stronger. Changes of the order of -4°C are suggested for February and August sea surface temperatures (Pearce 1991). But the cores on which these conclusions were based are few in number and were located in deep water well off the shelf. The lack of geographical resolution in this work makes it impossible to use it to establish the presence or absence of a Leeuwin Current at that time. A firm conclusion that does emerge from this work is that of a much strengthened and persistent cold boundary current off Western Australia. This makes cold water incursions on to the shelf and nearshore zone more likely. Furthermore, the dominance of cold water in the eastern Indian Ocean and a reduced flow of Western Pacific water through the Indonesian Archipelago would seem to make the functioning of the Leeuwin Current at the Last Glacial Maximum much less likely. The full implications of these changes could be firmly evaluated by the numerical models which have been used to explore the controls of Leeuwin Current formation (eg Batteen & Rutherford 1990, Godfrey & Weaver 1991). Conclusions A detailed stratigraphic and associated chronological data base is still being acquired for the Abrolhos carbonate platforms. Despite this limitation some important generalisations can be made concerning the role of the Leeuwin Current in platform and reef development, and the contrasting Late Quaternary stratigraphy of the Abrolhos reefs and eastern Australian reefs. The widespread development of Last Interglacial coral reefs along the western continental margin to latitudes as far south as 34.5°S, and molluscan faunal data, support a period of vigorous Leeuwin Current activity. The emergent coral-algal reefs of the central and western islands of the Abrolhos have been confirmed as an important part of this system. The distribution of known Holocene reefs is, in contrast, far more limited, both in areal extent and latitudinally; the most southerly significant reefs being at the Abrolhos at latitude 28.5°S. These differences are probably largely due to fluctuations in Leeuwin Current activity. The Abrolhos reefs are characterised by widespread emergent Pleistocene substrate, consisting of Last Interglacial reef facies. Holocene corals are present as thin, discontinuous veneers over this substrate. This is in marked contrast to the Great Barrier Reef, where Holocene reefs have buried the Pleistocene substrate. Last Interglacial reef substrates, both at the Abrolhos and elsewhere along the Western Australian coast, are frequently well above MSL and therefore could not be colonised by Holocene corals. The lack of significant thicknesses of Holocene reefs, as yet assessed only in very general ways, may also reflect slow growth rates of corals (Crossland 1981) operating at the southerly limits of their environment, under the influence of the Leeuwin Current. References Batteen M L & Rutherford M J 1990 Modeling studies of eddies in the Leeuwin Current: the role of thermal forcing, t P hys Oceanogr 20:1484-1520. Berggren W A, Burckle L H, Gta M B, Cooke MBS, Funnell 3 M, Gartner S, Hays J D, Kennett J P, Opdyke N ^ Pastouret L, Shackleton N ] & Takayanagi Y 1980 Towards a Quaternary time scale. Quat Res 13: 277. 302. Bloom A L 1974 Geomorphology of reef complexes. In : Reefs ^ time and space (Ed L F Laporte) SEPM Spec Publ 18: Bloom A L, Broeker W S, Chappell J, Matthews R K & Mesolella K J 1974 Quaternary sea level fluctuations 0n a tectonic coast: new Th^^/U^^ dates from Now Guinea. Quat Res 4:185-205. Carannante G, Esteban M & Simone L 1988 Carbonate lithofacies as palaeolatitude indicators: problems and limitations. Sed Geol 60: 333-346. Carter R M & Johnson D P 1986 Sea level controls on the post¬ glacial development of the Great Barrier Re e f Queensland. Mar Geol 77:137-164. Chappell J 1974 Geology of coral terraines. Muon Peninsula New Guinea: a study of Quaternary tectonic movements and sea level changes. Geol Soc Amer Bull 85: 553-570. Chappell J & Shackleton N J 1986 Oxygen isotopes and sea level. Nature 27: 137-140. Collins L B 1983 Postglacial sediments and history, southern Rottnest Shelf, Western Australia. PhD Thesis, Uni of WA (unpublished). Collins L B 1988 Sediments and history of the Rottnest Shelf, southwest Australia: a swell-dominated, non-tropi Ca j carbonate margin. Sed Geol 60:15-49. Crossland C J 1981 Seasonal growth of Acropora cf. formosa and Pocillopora damicornis on a high latitude reef (Houtman Abrolhos, Western Australia). Proc Fourth Int Coral Reef Symp Manila 1: 663-667. Fairbanks R G 1989 A 17,000 year old glacio-eustatic sea level record: influence of glacial melting on the Younger Dryas event and deep ocean circulation. Nature 342* 637-642 Fairbridge R W 1948 Notes on the geomorphology of the Pelsart Group of the Houtman's Abrolhos Islands. J R 0 y Soc W Aust 33:1-43. France R E 1985 The Holocene geology of the Pelsaert reef complex, southern Houtman Abrolhos, Western Australia. PhD Thesis, Uni of WA (unpublished). Godfrey J S & Weaver A G 1991 Is the Leeuwin Current driven by Pacific heating and winds? Progress in Oceanogr 27:225-272 Guilchcr A 1988 Coral Reef Geomorphology. John Wiley & Sons 228pp. Harris P T 1989 Sidescan sonar study of submerged reefs and terraces on the outer shelf and upper slope adjacent to Houtman Abrolhos reefs. Western Australia (unpublished manuscript). Hawkins R D 1969 Gun Island No 1 completion report. BMR Aust Rec 1968/205 (unpublished). Kendrick G W, Wyrwoll K H & Szabo B J 1991 Pliocene- Pleistocene coastal events and history along the western margin of Australia. Quaternary Science Reviews (in press). 56 Journal of the Royal Society of Western Australia, 74, 1991 Ku T L, Kimmel M A, Easton W H & O'Neil T J 1974 Eustatic sea level 120,000 years ago on Oahu, Hawaii. Science 183: 959-967. Lambeck K L 1987 The Perth Basin : a possible framework for its formation and evolution. Explor Geophys 18:124- 128. Lambeck K L & Nakada M 1990 Late Pleistocene and Holocene sea-level change along the Australian coast. Palaeo Palaeo Palaeo (Global and Planetary Change Section) 89:143-176. Lees A 1975 Possible influence of salinity and temperature on modern shelf carbonate sedimentation. Mar Geol 19: 159-198 Lees A & Buller A T1972 Modern temperate - water and warm water shelf carbonate sediments contrasted. Mar Geol 13: M67-M73. Levinton J S 1982 Marine Ecology. Prentice-Hall 526p. Marshall J F & Davies P J 1978 Skeletal carbonate variation on the continental shelf of eastern Australia. BMR J Aust Geol Geophysics 3:85-92. Nelson C S, Keone S L & Head P S 1988 Non-tropical carbonate deposits on the New Zealand continental shelf. Sed Geol 60: 71-94. Pearce A F 1991 Eastern boundary currents of the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. AF Pearce and DI Walker) J Roy Soc WA 74:35-45. Pearce A F, Praia A J & Manning C R 1989 Comparison of NOAA/AVHRR-2 sea surface temperatures with surface measurements in coastal waters. Internat J Remote Sens 10: 37-52. Playford P E 1988 Guidebook to the geology of Rottnest Island. Geol Soc Aust (Western Australian Division) Excursion Guidebook No. 2:67pp. Prell W L,Hutson W H, Williams D F, Be A W H, Geitzenauer K & Molfino B 1980 Surface circulation of the Indian Ocean during the Last Glacial Maximum, approximately 78,000 yr B P. Quat Res 74: 309-336. Purdy E G 1974 Reef configurations : cause and effect. In : Reefs in time and space. (Ed L F Laporte) SEPM Spec Publ 18: 9-76. Shackleton N J 1987 Oxygen isotopes, ice volume and sea level. Quat Sci Rev 6: 783-790. Skwarko S K 1990 On the age and correlations of the fossil reef at North Mole, Fremantle. Geol Surv W Aust, Unpubl Palaeontological Rep 1990/2. Smith G C & Cowley R G 1987 Potential of the northern Abrolhos Sub-Basin, Western Australia. APEA Journal 27:112-136. Stoddart P R 1969 Ecology and morphology of recent coral reefs. Biol Rev 44: 433-490. Szabo B J 1979 Urarium-series age of coral growth on Rottnest Island, Western Australia. Mar Geol 29: M11-M15. Teichert C 1947 Contributions to the geology of the Houtman's Abrolhos, Western Australia . Proc Linn Soc NSW 71 (3, 4): 145-196. Teichert C 1967 Age of coastal limestone. Western Australia. Aust J Sci 30:71. Van deGraaf W J E, Denman PD& Hocking RM 1976 Emerged Pleistocene marine terraces on Cape Range, Western Australia. Geol Surv W Aust Ann Rep 1975:62-69. Veeh H H & France R E 1988 Uranium-series ages of corals and coexisting phosphate deposits on Pelsaert Reef Complex, Houtman Abrolhos Islands, Western Australia. Quat Res 30:204-209. Veeh H H, Schwebel D, Van de Graaf W J E & Denman P D 1979 Uranium-series ages of coralline terrace deposits in Western Australia. J Geol Soc Aust 26: 285-292. Veevers J J 1974 Western continental margin of Australia. In: The Geology of Continental Margins. (Eds C A Burke & C L Drake), Springer-Verlag New York N.Y.: 605-515. Journal of the Royal Society of Western Australia, 74,1991, 59-69 Zoogeographic provinces of the Humboldt, Benguela and Leeuwin Current systems G J Morgan & F E Wells Department of Aquatic Invertebrates, Western Australian Museum, Perth, WA 6000, Australia. Abstract The distributional patterns of inshore marine faunas of southern South America, southern Africa and southern Western Australia are discussed. They are related to the effects of the cold northward flowing Humboldt and Benguela Currents and warm southward flowing Leeuwin Current respectively. The extent and nature of zoogeographic provinces, their faunal affinities and levels of endemicity are reviewed. In South America and southern Africa, cold temperate provinces resulting from the effects of the cold currents and associated upwellings act as barriers to dispersal of warm water faunas especially along the southwestern coasts. A corresponding cold temperate province is absent from southwestern Australia and warm water species are distributed farther south than in South America and Africa. Introduction Surface current patterns in the Atlantic, Pacific and Indian Oceans are characterized by major gyres that flow along the oceanic boundaries in a clockwise direction in the northern hemisphere and counter¬ clockwise in the southern hemisphere. The cold water, northward flowing Humboldt and Benguela Currents act with the prevailing winds to produce upwellings of nutrient-rich subsurface waters along the western coasts of southern South America and Africa. The high nutrient levels provide for increased phytoplankton production and so high production by zooplankton and secondary consumers such as anchovies (Tait 1968, Cushing & Walsh 1976). The presence of a southward flowing current along the western coast of Australia that contacted the Houtman Abrolhos but did not reach the continental mainland was postulated by Saville-Kent (1897). Dakin (1919) compared temperatures at the Houtman Abrolhos with those at Gcraldton, providing evidence of a warm offshore current. The presence of tropical species of marine invertebrates, especially molluscs, at the western end of Rottnest Island led marine biologists in the 1950's to conclude that there must be a current bringing planktonic larvae of tropical species south from areas such as the Houtman Abrolhos. It was not until 1980 that the Leeuwin Current was described (Cresswell & Golding 1980). Current systems have major effects on the distributions of marine biota (Sverdrup et al. 1942, Ekman 1953, Briggs 1974, Cushing & Walsh 1976). Currents significantly influence the dispersal of organisms, especially of larval stages. Also, they determine the ambient conditions along much of the inshore environment, particularly with respect to water temperatures, salinity and nutrients. As a result, they permit survival of species in areas that would otherwise be unsuitable. Conversely, they can act as barriers to settlement by distributing organisms away from suitable habitats or by causing otherwise suitable habitats to become sub-optimal or uninhabitable. The oceanography of the Humboldt, Benguela and Leeuwin Currents is discussed in detail by Pearce (1991) and is only briefly noted here. This paper summarises the effects of these currents on faunal distributions, and hence zoogeographic provinces, of the southern shores of South America, Africa and Australia (Fig.l). The three regions are discussed in turn and subsequently compared. The currents primarily influence southwestern coasts of the three continents but to discuss their effects relative to areas at similar latitudes with differing currents, eastern shores at the same latitudes are mentioned more briefly. The marine biota of the southern oceans is poorly known relative to that of the northern hemisphere. Discussion of faunal distributions must therefore be circumspect. The concept of biogeographic provinces or zones remains a somewhat contentious one and indeed there is no quantitative definition of a province that has enjoyed general support. The broader principles of marine zoogeography have been discussed by many workers, foremost amongst them Ekman (1953) and Briggs (1974). In concentrating on the effects of the major currents, this paper is a simplified discussion of zoogeography in the three systems. Ambient hydrographic conditions including currents are by no means the only determinants of marine faunal occurrences. In particular, geological and long term climatological events associated with the movement of continents have had major influence upon the distributions and affinities of modern faunas. Present conditions, especially currents, act as recent 59 Journal of the Royal Society of Western Australia / 74, 1991 Figure 1 Regions discussed in this paper. Inserts enlarged in figures 2, 4 and 6. modifiers of those longer term processes. The provinces here discussed are taken from the recent literature. We have attempted to incorporate data on a wide variety of taxa but our personal interests result in a bias towards crustaceans and molluscs. The Humboldt, Benguela and Leeuwin Currents are surface currents and have limited direct influence on deeper waters. As such, this discussion is restricted to faunas of the intertidal and subtidal zones to about 200m depth. South America Currents: Extending south to 56°S, the western coast of Peru and Chile is contacted directly by the West Wind Drift that splits into the northward flowing Peru, or Humboldt, Coastal Current and the southward flowing Cape Horn Current that skirts the Cabo de Hornos and flows northward along the east coast as the Falkland Current (Fig.2). The Humboldt Current usually contacts the west coast at 41-46°S but in winter it may develop as far north as 32°S. The northward progression of this current and the prevailing southwesterly winds result in major upwellings of cool waters to the surface along most of the central and northern Chile coast. In apposition to the Humboldt Current, warmer subtropical water is transferred south by the surface Chile Coastal Counter-current. The southern penetration of this current usually reaches to 37°S in summer and only 31 °S in winter. At the Subtropical Convergence (23-25°S in winter, 33-34°S in summer), the Humboldt Current flows under this warmer water. The net effect of water current movements is that there are relatively small differences in water temperatures over large distances of the Chile-Peru coastline (Brattstrom & Johanssen 1983). Provinces: The relatively constant hydrographic conditions along the coast of southwestern South America (Brattstrom & Johanssen 1983, Brattstrom 1990) suggest that biotic distributions should show Figure 2 Zoogeographic provinces of southwestern South America. Cold currents indicated by solid arrows, warm currents by open arrows. (After various authors especially Briggs 1974, Brattstrom & Johanssen 1983). 60 Journal of the Royal Society of Western Australia, 74, 1991 clinal rather than sudden changes in composition. Indeed, most intertidal and shallow-water species have very large ranges along this coast (Brattstrom 1990). However, the distribution patterns for many taxa support the recognition of two zoogeographic provinces, one cold temperate (or antiboreal), the other warm temperate (Fig. 3). lOO-j In addition to the effects of currents at this point, the topography of the coast north and south varies. North of 42°S there are no archipelagos, few islands or sheltered bays and the rocky and sandy beaches are exposed to the open ocean. South of 42°S, the coast is fringed by archipelagos of thousands of islands, is split by narrow sounds and fjords and has more sheltered beaches and mudflats. 83 A V Cl CH Figure 3 Composition of fauna (mixed taxa) along southwestern coast of South America. 0: endemic species; M ‘.species with southern affinities; □ .-species with northern affinities; B :other. A: Arica; V: Valparaiso; Cl: Chiloe Island; CH: Cabo de Homos. (Calculated from data of Brattstrom & Johanssen 1983). South of about 42°S, the northern end of Chiloe Island, the faunal assemblages show cold temperate affinities and this province is usually known as the Patagonian or Magellanic Province. Its northern border corresponds closely with the position of impact of the Humboldt Current on the South American coast. It is recognised that the southeastern coast of South America is also cold temperate but there is some uncertainty whether or not the southeastern and southwestern coasts should be regarded as a single or as two provinces, separated at the Cabo de Homos. The warm temperate province, known variously as the Peruvian, Peruvian-Chilean or Peru-Chilean Province, is recognised as extending from 2-6°S south to somewhere between 30° and 42°S. The close approach of this province to the equator is a result of the upwellings and movement of cold water by the Humboldt Current. The southern boundary of the Peru-Chilean Province is open to some interpretation. There seems little doubt that a change in faunal composition occurs at or near 42°S, evident in the distributions of decapod crustaceans (Rathbun 1910, Haig 1955, Garth 1957), echinoderms (Madsen 1956, Bernasconi 1964), polychactes and ostracod crustaceans (Hartmann- Schroder & Hartmann 1962), molluscs (Dali 1909, Stuardo 1964, Dell 1971, Marincovich 1973), fish (Norman 1937, Mann 1954) and mixed taxa (Semenov 1977, Brattstrom & Johanssen 1983). From the study of Brattstrom & Johanssen (1983), approximately 25% of shelf benthic invertebrates have their northern or southern boundaries at or very near 42°S. Most workers have acknowledged a biogeographic transitional zone extending from 42°S to somewhere between 38 and 30°S (see Brattstrom & Johanssen 1983: fig.18). Many cold temperate species from the Magellanic Province extend into the Peru-Chilean, though few reach farther north than 30°S. This southern component is replenished by dispersals in the cold water of the northward flowing Humboldt Coastal Current. Most species in this region, however, are warm temperate. Some workers have regarded the transition zone as a separate province (eg Knox 1960: Central Chilean Province) but most agree with Dahl’s (1960) observation that characterization of a province by a strong endemic element is lacking. Brattstrom & Johanssen (1983) argued that the transitional area, having a preponderance of warm temperate species, is part of the Peru-Chilean Province. It could also be suggested that it is an overlap zone of the two provinces, similar to that proposed for western Australia. Balech (1954) and Brattstrom & Johanssen (1983) attributed the transitional zone to variations in the relative effects of the cold Humboldt Current and warmer Chile Coastal Counter-current. There is little variation in hydrographic conditions south of 42°S. The proportion of endemic faunas varies considerably between taxa. For the warm temperate Peru-Chilean Province, endemicity of 23-53% has been recorded for littoral molluscs (Dali 1909) and decapod crustaceans (Haig 1955, Garth 1957) and Briggs (1974) suggested that at least 50% of fish may be endemics. For the cold temperate Magellanic Province, including both southwestern and southeastern coasts, endemicity figures are generally higher with estimates of 33-61% for anomuran and isopod crustaceans, echinoderms, bivalve molluscs and fish (Ekman 1953, Haig 1955, Madsen 1956, Soot-Ryen 1959, Menzics 1962, Pawson 1969, Briggs 1974). As was noted above, there is some uncertainty as to the southern boundary of the Magellanic Province. Stuardo (1964) suggested that Cabo de Hornos might be regarded as a boundary between two cold temperate provinces. Other authors have suggested several cold temperate provinces for South America, either with a border at Cabo de Hornos (Balech 1954) or with the southernmost landmass of the continent on both sides of the Cabo de Hornos in one province and a further cold temperate province adjoining to both the east and west (Forbes 1856, Knox 1960). Most authors have treated the southwestern and southeastern coastline as belonging within a single province (Briggs 1974) although the southeastern fauna is much less well known. The northern boundary of this province on the 61 Journal of the Royal Society of Western Australia, 74,1991 east coast is generally regarded as coinciding with the Rio de la Plata at 35°S where the cold Falkland Current diverts to the east as it contacts the warm Brazil Current. An Eastern South American Warm Temperate Province is recognisable from 35°S to approximately Cabo Frio at 23°S but the composition of this fauna is very poorly known (Briggs 1974). Summary: The major provincial boundary of cold temperate faunas of southwestern South America coincides with the normal point of impact of the Humboldt Current. This current flows north, dispersing cold water and southern species while the Cape Horn Current flows south. The point of divergence of currents acts as a fairly effective barrier to dispersal of inshore species. The contrasting effects of the Humboldt Current and Chile Coastal Counter-current result in an overlap zone between cold and warm temperate provinces between 30° and 40°S. The Humboldt Current extends the warm temperate conditions much farther north on the western coast (approximately 3°S) than the Falkland Current does in the east (about 23°S) (Fig.2). Thus tropical faunas range much farther south along the eastern coast than along the western. Southern Africa Currents: At about 35°S, the southern tip of Africa at Cape Agulhas is 20° farther north than Cabo de Homos of South America. Nevertheless, it is affected by the cold water northward flowing Benguela Current (Fig.4). The Benguela Current and the prevailing southeast trade winds result in major upwellings along the southwest coast bringing cold sub-surface waters from 100-300 m to the surface. The Benguela Current flows northward into the Gulf of Guinea, where it encounters the warm Angola Current flowing southeastward, and then continues near the Equator as the now warm South COAST PROVINCE Nslal „ .PROVINCE Durban 4 ii'V Transkri, ~ r T " ‘ -,/w Cape Town SaWanha ***?'i Pon Clurahcth/) 'war M TFMI'ERAIE ' o* V SOUTH SOUTH COASI PROVINCE Figure 4 Zoogeographic provinces of southern Africa. Currents as per Fig. 2. (After Brown & Jarman 1978, Kensley 1981). Equatorial Current. The combined effects of the cold northward flowing current and upwellings result in relatively little seasonal variation in sea surface temperatures along the southwestern coast of Africa. In contrast, the eastern coast of southern Africa is dominated by southward flowing offshoots of the Indian Ocean South Equatorial Current. The warm water Agulhas Current flows southward along the southeastern coast to the vicinity of Cape Agulhas. Some warm water moves west around the Cape but most diverts to the south on contact with the Benguela Current. Provinces: There is a substantial literature relating to the faunal provinces of southern Africa (Briggs 1974, Brown & Jarman 1978). Early workers (Forbes 1856, Woodward 1856, Ortmann 1896) suggested that a single province extends from Angola or South West Africa (Namibia) to the vicinity of Durban. Ekman (1935) regarded this province as warm temperate. Stephenson (1947), Hedgpeth (1957) and Knox (1960) divided the coast into a cold temperate province extending from the tropics of west Africa to Cape Town and a warm temperate province east of Cape Town and best developed between Cape Agulhas and Port Elizabeth. Ekman (1953) and Briggs (1974) regarded the southwestern province as a second warm temperate one; Briggs discussed the Southwestern Africa Province from Mo^amcdes, Angola (15°S) to Cape Peninsula (34°S) and the Agulhas Province from Cape Peninsula east to Port Elizabeth (34°S). Briggs cited the level of endemicity of the Southwestern Province to be considerably lower than that of the Agulhas Province with figures of 17% and 34% respectively (taken from Ekman (1953), based upon Stephenson's works). More recent papers (Brown & Jarman 1978, Kensley 1981, 1983, Kilburn & Rippey 1982) supported the existence of both warm and cold temperate provinces, with a subtropical province on the east coast north of East London. The names and precise boundaries of the provinces differ between workers, the latter at least in part due to differences in distributions between taxa. For the purposes of this discussion, the provinces of Brown & Jarman (1978) and Kensley (1981, 1983) generally will be followed (Fig. 4). This pattern differs slightly from that of Kilburn & Rippey (1982) and where significant, differences are discussed. The Tropical West African Province extends south to 20°S although Kilburn & Rippey (1982) placed its southern limit at 17°S. The fauna of Namibia is poorly known and definition of provincial borders remains speculative. The southern limit is determined by cold water effects of the Benguela Current and upwellings and typical tropical species rarely range south of this limit (Kensley 1981). Although it may be regarded as a tropical province, the influence of the Benguela Current continues a considerable distance farther north. For example, no coral reefs occur south of the equator on the west coast, the southernmost reefs being those of the Gulf of Guinea at 0-5°N. In contrast, on the 62 Journal of the Royal Society of Western Australia, 74,1991 east coast coral reefs are found as far south as southern Mozambique at about 28°S. The Cold Temperate West Coast or Namaqua Province extends from 20°S to Cape Agulhas (35°S). Kilburn & Rippey (1982) recognised an overlap zone between the Namaqua and West African Provinces, extending for 7° of latitude south of 17°S and offsetting the northern boundary of the Namaqua Province proper southwards by 4° to 24°S. The dominant influences in the province are the cold water effects of the Benguela Current and upwellings. The faunas here are more poorly known than are those to the east but are characterised by low species diversities but high populations of species present (eg crustaceans: Kensley 1981, 1983, molluscs: Kilburn & Rippey 1982). Productivity is high due to the nutrients brought to the surface in the upwellings. The fauna comprises few Indo-Wcst Pacific (IWP) species and these are largely confined to the southernmost areas of the province (Kensley 1981, 1983). Atlantic-Mediterranean species proportionally dominate IWP forms amongst crustaceans and molluscs. Interestingly, the number of Atlantic crustacean species is lower in the Namaqua Province, itself situated in the South Atlantic, than in the Warm Temperate South Coast Province. The low number of Atlantic species in the Namaqua Province can be explained in terms of the water temperature regime. Water temperatures are on average warmer both north and south of the major area of upwelling at 25-30°S. The cold water of the upwellings and the northward flow of the Benguela Current act as a cold water barrier to colonisation by the more diverse tropical Atlantic faunas. Endemic species are the dominant faunal element amongst molluscs, accounting for 88% of the Namaqua species (Kilburn & Rippey 1982). Their data were pertinent only to south of the Orange River, information being too limited to speculate on endemicity farther north. The species tend not to be confined to the Namaqua Province but are endemic to southern Africa. Analysis of Kensley’s (1983) data reveals that for crustaceans the endemicity is lower with 35-57% for decapods and 41-85% for peracarids, endemicity decreasing to the north. Ekman (1953) noted 17% endemicity for several taxa on the basis of Stephenson's works. Endemic fishes appear to comprise less than 20% of the fauna (Penrith 1969, Smith 1949, I960, Briggs 1974). Figure 5 Composition of crustacean and molluscan taxa along southern coast of Africa. A, Crustacea (Dccapoda, Amphipoda, Isopoda combined). B3 :endemic species; H Atlantic-Mediterranean species; □ :Indo-West Pacific species; S .other. KR: Kunene River; L: Liideritz; SB: Saldanha Bay; FB: False Bay; PE: Port Elizabeth; EL: East London; D: Durban; II: Inhaca Island. (Data from Kensley 1981,1983); B, Mollusca. As per Fig. 5A except E3 :Cape Endemics and S3 subtropical endemics. Na: Namaqua (southern); FB: False Bay; CA: Cape Agulhas; EC: East Cape; WT: West Transkei; ET: East Transkei; Nt: Natal. (Data from Kilburn & Rippey 1983). 63 04343-6 Journal of the Royal Society of Western Australia, 74, 1991 A south coast warm temperate province has been recognised by most workers (Briggs 1974). The Warm Temperate South Coast Province of Brown & Jarman (1978), or Algoa Province of Kilburn & Rippey (1982), extends in an east-north-east direction from an overlap zone between Cape Town to Cape Agulhas to just north of East London. Species diversity and, not surprisingly, the number of IWP species is higher than in the Namaqua Province. For pcracarid crustaceans, this province is the most diverse in southern Africa. The proportion of IWP molluscs decreases progressively from east to west, reaching their western limits at False Bay/Cape of Good Hope (Kilburn & Rippey 1982) (Fig.5). The numbers of Atlantic species are as high as or higher than in the Namaqua and as proportions of the total number of species, Atlantic representation can be as high or only slightly lower. The increase in Atlantic species to the east of Cape Agulhas, at least in the Crustacea, may be explained by at least two processes. Some species may represent relict populations of more diverse faunas present in the warmer Pleistocene Period. Others are likely to be more recent migrants, distributed by some movement of South Atlantic waters around Cape Agulhas (Shannon 1966) or by the large gyre circulation of Atlantic water into the South Equatorial Current of the Indian Ocean and thence southwest into the Agulhas Current. The warmer conditions of the Warm Temperate South Coast Province are more suitable for recruitment of many Atlantic species. The number of species endemic to southern Africa is very high. Endemics account for 33- 54% of decapod and amphipod crustaceans, 88-92% of isopod crustaceans and 38-70% of molluscs (Kensley 1981, 1983, Kilburn & Rippey 1982, Gosliner 1987), with the proportion of endemics generally decreasing to the east. High endcmicity has also been recorded in this province for hydroids (Millard 1978), soft corals (Williams 1990), ascidians (Millar 1962), fishes (Smith 1949, 1960, Penrith 1969) and mixed taxa (Ekman 1953, Day' et al 1970). Day (1978) and Kensley (1983) suggested that the province has been a centre of evolutionary radiation for crustacean groups with ancestral stock from both Atlantic and IWP origins; Millard (1978) proposed a similar evolutionary centre for hydroids. The Subtropical East Coast Province extends from an area of overlap with the Warm Temperate South Coast Province to Inhambane, Mozambique, close to the Tropic of Capricorn. Kilburn & Rippey's (1982) equivalent Natal Province has its northern boundary 6° farther south, their interpretation being that mollusc faunas north of 29°S are essentially tropical. Coral reefs appear about midway along the north-south extent of this province indicating the warm water effects of the Agulhas Current. The diversity of decapod crustaceans increases markedly but that of pcracarid groups declines (Kensley 1983). The IWP and Atlantic faunal elements continue to increase and decrease respectively in both numbers of species and proportion of total species. Endemicity declines towards the north in numbers of species and proportionally for crustaceans (Kensley 1981, 1983) and molluscs (Kilburn & Rippey 1982, Gosliner 1987) (Fig. 5). In molluscs, the endemics show a shift towards tropical rather than temperate affinities. This province may also be regarded as an overlap or transition zone between tropical and warm temperate provinces (see concluding comparison of systems). The Tropical East Coast Province ranges from the Tropic to north of the equator; Kilburn & Rippey's (1982) Indo-Pacific Province begins 6° farther south. The province is heavily dominated by IWP forms. There are very few Atlantic and few endemic species. Overall affinities of the fauna for the African coastline south of 15°S vary somewhat between taxa. The IWP faunal element is dominant in species diversity for decapod crustaceans (61-66% of species) (Kensley 1981, 1983) while the Cape endemic component is dominant for pcracarid crustaceans (46% of amphipod species, 65% of isopod species) and molluscs (19-89%, depending on locality, Kilburn & Rippey 1982) (Fig.5). The Atlantic-Mediterranean element accounts for only 5-15% of crustacean and 1- 12% of molluscan species. Summary: The Benguela Current affects the distribution of inshore faunas in several ways. It circulates cold water northwards and encourages the upwcllings of sub-surface water along the southwestern coast. The current and upwcllings combine to extend a cold temperate province to within 20° of the equator and form an effective cold water barrier to movement of warm water Atlantic species southwards and to warm water IWP species westwards. There is a distinct containment of tropical faunas to the north of the cold water province. The western coast tropical province has its southern boundary at least 5° farther north than its equivalent province on the eastern coast. Areas on the eastern coast at similar latitudes to that of the cold temperate Namaqua Province arc subtropical or warm temperate. Southwestern Australia Currents: The southwestern coast of Australia extends just south of 35°S near Albany, similar in latitude to the southern tip of Africa and is therefore north of the main flow of the West Wind Drift. The southward flowing Leeuwin Current, originating to the north of North West Cape (21° 47'S), brings warm, low salinity water south around Cape Leeuwin and east into the Great Australian Bight (Cresswcll & Golding 1980, Godfrey & Ridgeway 1984, Pearce & Cresswell 1985, Pearce 1991) (Fig.6). The Leeuwin Current usually peaks in the winter, becoming weak in summer. It flows along the outer continental margin off the coast, intercepting offshore islands but not the inshore continental coastline until the Cape Naturaliste-Cape Leeuwin area. Because of the Leeuwin Current there is no substantial upwelling along the southwestern coast of 64 Journal of the Royal Society of Western Australia, 74, 1991 Figure 6 Zoogeographic provinces of Western Australia. Currents as per Fig. 2. (After Wilson & Allen 1987). Australia. Surface waters are nutrient poor and primary productivity is low. The Leeuwin Current roughly parallels the East Australian Current that brings warm waters southward to about 33°S before diverting as eddies into the Tasman Sea. Thus, unlike South America and Africa, Australia has southward flowing warm currents on both western and eastern coasts. Provinces: Biogeographic reviews of Australian marine inshore fauna, especially intertidal species, from the 1930's into the 1970's recognised five to six provinces, of which two occurred along the western coast (Whitley 1932, Bennett & Pope 1953, Knox 1963, Briggs 1974). A tropical province, the Dampieran, extended from about Shark Bay or Gerald ton north and east to Cape York and a warm temperate province, the Flindersian, extended from the Dampieran south and east to western Victoria. More recently, the marine biogeographic zonation of Australia has been simplified into a recognition of a Northern Australian Tropical and a Southern Australian Warm Temperate Province with broad zones of overlap on both western and eastern coasts that for this discussion will be regarded as the Western Coast and Eastern Coast Overlap Zones (Fig.6) (Wilson & Gillett 1971, Marsh 1976, Wilson & Stevenson 1977, Wells 1980,1986,1990, Wilson & Allen 1987). The northern coast of Western Australia, northeast from North West Cape, has a tropical biota that is continuous with other parts of the Indo-Wcst Pacific. In general, species diversity decreases with increasing latitude but there are no major distributional boundaries, most species reaching as far south as North West Cape. Endemicity is low in the Northern Australian Tropical Province with about 10% of mollusc species (Wilson & Allen 1987), 13% of fishes (Wilson & Allen 1987), 17-22% of brachyuran and anomuran decapod species (Griffin & Yaldwyn 1967, Morgan 1990), essentially no corals (Potts 1985, Veron 1985, Wilson & Allen 1987) and, for the northwestern coast, about 13% of echinoderms (Marsh 1976, Marsh & Marshall 1983) being endemic to Australia (Fig.7). The relationship of the tropical northern coast with other parts of the Indo- West Pacific has been most recently discussed by Wells (1986, 1990) and Wilson & Allen (1987). The southern coast of Western Australia east of Cape Leeuwin (34°22’S) is part of the Southern Australian Warm Temperate Province. In addition to the effects of the weakening Leeuwin Current, the warm temperate province is influenced by the cold West Wind Drift, which has been responsible for the distribution of some widespread or circumpolar elements into the southern Australian fauna (Fell 1962, Knox 1979, Edgar 1986). IWP species representation is low and decreases from west to east. Most of the temperate species that occur along the south coast of Western Australia reach as far west as Cape Leeuwin without major biogeographic discontinuities. Rates of species endemicity arc much higher in this province than in northern waters: approximately 85% for fishes (Wilson & Allen 1987), possibly 95% for molluscs (Wells 1980, Wilson & Allen 1987), 90% for echinoderms (Clark 1946, Rowe & Vail 1982) and 63% for decapod crustaceans (77% endemic to Australia) (Morgan & Jones 1991). Fig. 7 shows the affinities of crustaceans and molluscs of western Australia. Note that for crustaceans, species endemic to Australia are plotted, accounting for a high proportion of the south coast fauna. For molluscs, south coast endemics are incorporated in the category of species with southern affinities and endemics are those confined to the four regions. In many respects, the Western Coast Overlap Zone is a region of transition with gradual replacement of a tropical fauna in the north by a predominantly temperate fauna in the south (Wilson & Allen 1987). There is a small proportion of species endemic to Western Australia and most of these have at least part of their range in the Western Coast Zone and often achieve their greatest numbers there (Wells 1980, Wilson & Allen 1987). The proportion of endemics varies between taxa: 20% for shallow water asteroids (Marsh 1976) and less than 10% for prosobranch molluscs (Wells 1980). Two offshore regions clearly illustrate the effects of the Leeuwin Current on faunal composition: the Houtman Abrolhos (28-29°S) and Rottnest Island (32°S). Although it has substantial numbers of temperate species and Western Australian endemics, the fauna of 65 Journal of the Royal Society of Western Australia, 74, 1991 the Houtman Abrolhos is essentially tropical (Montgomery 1931, Wilson & Marsh 1979, Wells 1980, Veron 1985) and the Abrolhos is generally considered to be the southern limit in Western Australia of the tropical marine biota (Wells 1980, Wilson & Allen 1987). The southern limit of the tropical fauna in eastern Australia is usually regarded as being slightly farther north, somewhere between 26° and 27°S (Endean 1957, Wilson & Gillett 1971). H4 Mollusca 85 K SB RI SC Figure 7 Composition of crustacean and molluscan taxa along western coast of Australia. A, Crustacea (decapods only except Rottnest Island). 0 . Australian endemic species; U rspecies with southern affinities (excluding south coast endemics); □ :lndo-West Pacific species; & .other. K: Kimberleys; SB: Shark Bay; RI: Rottnest Island; SC: South Coast. (Data from Jones 1990; Morgan 1990, Morgan & Jones 1991.) B, Mollusca. :0 species endemic to the region, ■ .species with southern affinities (including Australian south coast endemics); □ :IWP species. (Data in part from Wells 1980) Hodgkin el al. (1959) recorded 18 tropical invertebrate species at Rottnest Island. Hutchins (1979) found that about 26% of the 350 fish species recorded from Rottnest Island were of tropical origin and almost 40% of marine crustaceans known from the island are tropical IWP species (Jones & Morgan unpublished data) (Fig.7). The coral Pocillopora damicornis forms a small reef near Parker Point, the southernmost reef development in the state and one of the most southerly in the world. The fauna associated with the coral is similar to that of many tropical localities elsewhere in the world (Black & Prince 1983). The tropical Echinometra mathaei, the dominant echinoid at the western end of Rottnest Island, shows a continuous reproductive season typical of tropical species (Pearse & Phillips 1968). Black & Johnson (1983) reported that many of the fauna at Rottnest Island are of tropical origin. Tropical marine invertebrates extend farther to the south, well into the Southern Australian Temperate Province. Maxwell & Cress well (1981) have shown that larvae of tropical species can be distributed into the Great Australian Bight by the Leeuwin Current. Wells (1980) showed that 9 of 308 tropical prosobranch gastropod species examined reached Cape Leeuwin and 5 extended onto the south coast. Veron & Marsh (1988) found that 25 of 318 species of hermatypic corals reached as far south as Rottnest Island and 9 species occurred on the south coast. Differences in composition of the fauna between inshore and offshore areas have long been known (Saville-Kent 1897, Dakin 1919). Wells (1985) found that proportions of tropical, temperate and endemic molluscs were almost identical at the eastern end of Rottnest and inshore. The proportion of endemic species at the western end of Rottnest was similar (13- 16%) but the proportion of tropical species was nearly double and temperate species declined from 67% inshore and at the eastern end of Rottnest to 52% at the western end. Similarly, molluscan biomass and densities were dominated by temperate species at inshore sites while tropical species accounted for over half the biomass and density at some western end sites. Summary: The warm southward flowing Leeuwin Current disperses tropical representatives of many taxa to the southwestern and southern coasts of Australia. No cold temperate province exists and tropical and warm temperate provinces grade into each other in a broad overlap zone. The offshore nature of the current results in higher proportions of tropical marine faunas at offshore localities than along the mainland. Due largely to the effects of the Leeuwin and East Australian Currents, the pattern of provinces is similar along the western and eastern coasts of Australia. Comparison of Humboldt, Benguela and Leeuwin Systems There are substantial differences between the hydrography and marine faunas of the southwestern coasts of South America, Africa and Australia. Although South America reaches fully 20° of latitude farther south than Africa and southwestern Australia, there are greater similarities in inshore water circulation between South America and southern Africa. Both show a general pattern of northward moving cool waters along their western coasts that is reinforced by a series of major upwellings. In contrast, the southwestern coast of Australia lacks both a northward flowing cold current and significant upwellings. The dominant inshore current is southward flowing and warm water. Australia is unique amongst . I Journal of the Royal Society of Western Australia, 74, 1991 the mid-latitude southern continents in having a similar pattern of major currents on western and eastern coasts. The Humboldt and Benguela Currents serve to extend temperate waters and their associated faunas much farther north than along the Western Australian coast. In South America, a warm temperate province extends north to about 3°S and in southern Africa, a cold temperate province to within 20°S. The cold waters and northward flow serve as effective barriers to southward movement of tropical west coast faunas that are restricted to only limited penetration south of the equator. The cold water conditions also restrict invasion by eastern coast tropical and warm water species, distributed much farther southward than on the respective western coasts. The Humboldt and Benguela Currents therefore act as barriers to warm water species through distributional and environmental effects. The Leeuwin Current extends warm waters much farther south along the southwestern Australian coast. A tropical north coast province extends to about 22°S inshore and to 29°S at the Houtman Abrolhos, approximately 19° (26°) and 2° (9°) farther south than in western South America and southern Africa respectively. Many tropical species with IWP affinities are distributed southward with a significant tropical inshore faunal element evident as far south as Rottnest Island at 32°S. A number of tropical species range farther south into the Great Australian Bight. There is no distinct barrier to warm water species, tropical faunas instead gradually diminishing with increasing latitude. Coral reefs are richly developed at the Houtman Abrolhos and occur at Rottnest Island but are found no farther south than 1°28'S and 0-5°N off southwestern South America and Africa respectively (Kensley 1981, Wells 1988). A comparison of the marine zoogeographic provinces of the three continents reveals some close similarities between the western coast of Australia and the eastern coast of southern Africa. The Northern Australian Tropical Province and the Tropical East Coast Province of Africa share high species diversity, a high incidence of tropical IWP species and low species endemicity. The Western Coast Overlap Zone of Australia and the Subtropical East Coast Province of Africa have relatively high but decreasing diversity, many IWP species and a recognisable endemic component. The Southern Australian Warm Temperate Province and the Warm Temperate South Coast Province of Africa show decreased diversity, fewer tropical species that diminish with increasing latitude and relatively high numbers of endemics. For many taxa, these provinces appear to have the highest rate of endemicity in western Australia and southern Africa and in the latter region has the highest diversity for some groups (eg peracarid crustaceans). It is more difficult to make valid comparisons between western Australian and eastern South American provinces. The marine fauna of the latter region remains very inadequately described. The Brazilian Province shows high species diversity and low endemicity, broadly similar to the Australian Tropical Province. There is no recognised overlap zone between tropical and temperate provinces of eastern South America but this may reflect more the paucity of faunal distributional data rather than a real disjunction between provinces. There is no southwestern Australian counterpart to the cold temperate Magellanic and Namaqua Provinces of southern South America and southwestern Africa. Consequently there is no corresponding drop in diversity, rise in productivity and retreat of tropical faunal elements associated with a cold water mass and upwellings. There is no distinct cold water barrier to dispersal and recruitment of tropical species. Largely as a result of inshore circulation effects, the marine biogeographic provinces of the western and eastern coasts of South America and Africa are asymmetrical. This is particularly true for southern Africa where the cold temperate province is offset to the southwest. In Australia, the pattern of provinces is essentially symmetrical along the western and eastern coasts with relatively minor differences in the extent of provinces due to variable effects of the Leeuwin and East Australian Currents. It is wise to conclude this review with the note of caution sounded in the Introduction. Prevailing hydrographic conditions are modifiers of faunal distributions established by long term geological and climatological processes. This can be illustrated briefly for Australia. Throughout the mid- to late Tertiary, the southwestern coast supported a high proportion of faunas originating in the late Mesozoic tropical ocean Tethys and derived IWP species. This is in contrast to southeastern Australia which then had many temperate palaeoaustral forms (Knox 1980, Darragh 1985, Wilson & Allen 1987). Most invasions of tropical species along the southern Australian coast have been from the west. The present hydrographic conditions have maintained rather than caused the relatively high tropical influence in southwestern Australia. References Balech E 1954 Divisibn zoogeografica del litoral sudamericano. Revta Biol Mar 4:184-195. Bennett I & Pope E 1953 Intertidal zonation of the exposed rocky shores of Victoria, together with a rearrangement of the biogeographical provinces of temperate Australian shores. Aust J Mar Frcshw Res 4:105-159. Bernasconi I 1964 Distribution geografica de los Equinoideos y Asteroideos de la extremidad austral de Sudamerica. Boln Inst Biol Mar Mar Del Plata 7:43-50. Black R & Johnson M S 1983 Marine biological studies on Rottnest Island. J R Soc W Aust 66:24-28. Black R & Prince J 1983 Fauna associated with the coral Pocillopora damicornis at the southern limit of its distribution in Western Australia. J Biogeogr 10:135- 67 Journal of the Royal Society of Western Australia/ 74,1991 Brattstrom H 1990 Intertidal ecology of the northernmost part of the Chilean Archipelago. Reports of the Lund University Chile Expedition 1948-49. No. 50. Sarsia 75:107-160. Brattstrom H & Johanssen A 1983 Ecological and regional zoogeography of the marine benthic fauna of Chile. Reports of the Lund University Chile Expedition 1948- 49. No. 49. Sarsia 68:289-339. Briggs J C 1974 Marine Zoogeography. McGraw-Hill, New York. Brown A C a a Psammoperca waigiensis a a Epinephelus lanceolatus a Epinephelus rivulatus a Belonepterygion fasciolatum j Plectorhinchus flavomaculatus a a Plectorhinchus schotaf a Parupeneus chrysopleuron j' a Parupeneus signaius j' a j Pempheris oualensis j/ a Pempheris schwenkii j Pla tax leira j' a j a a Chaelodon auriga j' a Chaelodon citrinellus j Chaetodon lineolatus a Chaetodon lunula j/ a Chaetodon plebeius j/ a Abudefduf sexfasciatus Abudefduf sordidus j Abudefduf vaigiensis j ,a Plectroglyphidodon leucozonus j/ a Pomacentrus coelestis j Pomacentrus milleri j/ a Stegastes obreplus Anampses caeruleopunctatus j Anampses geographicus j' a a Cheilio inermis a Cor is aygula j' a Gomphosus varius j ,a Hemigymnus faciatus j Labroides dimidiatus Stethojulis bandanensis j/ a Stethojulis strigiventer j' a Thalassoma amblycephalum j Thalassoma hardwicke j Thalassoma lunare Thalassoma lutescens j/ a Thalassoma purpureum j' a Scarus festivus a Scarus ghobban j/ a Scarus gibbus j Scarus reticulatus j Scarus schlegeli j/ a Scarus sordidus j' a Enlomacrodus strialus j Omobranchus germaini j,a j/ a Petroscirtes breviceps j j/ a Petroscirtes mitratus j Plagiotremus rhinorhynchos j/ a a a Plagiotremus tapeinosoma j/ a Norfolkia brachylepis Amblygobius phalaena a Barbuligobius bohlkei j Gnatholepis inconsequens a Priolepis nuchifasciatus a Ptereleotris evides j Acanthurus nigrofuscus j Acanthurus triostegus j Naso unicornis j Siganus fuscescens j Balistoides viridescens a a Ostacion cubicus j Arothron hispidus j Diodon holacanthus j 81 Journal of the Royal Society of Western Australia, 74, 1991 Table 2 Summary of times of recruitment of 23 species of tropical reef fishes at Rottnest Island Species j F M A M J 1 A s O N D Plotosus lineatus Parupeneus chrysopleuron X X Parupeneus signatus X X X X X Pempheris oualensis X X X Chaetodon plebeius Abudefduf sexfasciatus X X X X X Abudefduf vaigiensis X X X X X Plectroglyphidodon leucozonus Stegastes obreplus X X X Anampses geographicus Labroides dimidiatus X X X X X Stethojulis bandanensis Stethojulis strigiventer X X X X Thalassoma lunare X X X Thalassoma lutescens X X X Thalassoma purpureum X X X X X Scarus ghobban Scarus gibbus X X X X X Ptereleotris evides X Omobranchus germaini Acanthurus nigrofuscus X X X Acanthurus triostegus X X X X Naso unicornis X one species, the blenny Omobranchus germaini, were recruited during the summer months. The times of settlement for some species were difficult to determine. Although species such as the pomacentrids Abudefduf vaigiensis and A. sexfasciatus remain in the open after settling, others such as the labrids Thalassoma purpureum and T. lutescens hide among the rocks and coral and are therefore difficult to detect. Thus, the times of recruitment for some species, as indicated in Table 2, may be affected by this cryptic phase. Many of the small tropical fish which settle at the Island disappear during the subsequent months, perhaps falling prey to predators, or even succumbing to unfavourable environmental conditions. However, some of the hardier species persist right through the winter months, often surviving to adulthood. These either remain in the shallow bays or move out to deeper offshore reefs as they become larger. In contrast to the above pattern of recruitment, Russell etal. (1977), working on the tropical reef fish fauna at Queensland’s One Tree Island (latitude 23°30'S), showed that a peak in recruitment occurs in summer, with most species having a breeding season in the mid spring to early autumn period. They found no evidence of settlement during June and July. This summer peak in recruitment also appears to occur on the coral reefs of Western Australia's northern waters (G R Allen, pers comm). Why is there a difference at Rottnest Island? As noted earlier, the Leeuwin Current flows predominantly during autumn and winter, weakening and moving away from the island in summer. This close relationship between the commencement and subsequent cessation of tropical recruitment at the island and the arrival and departure of the Leeuwin Current indicates that this current is obviously carrying the young of these tropical fish species to Rottnest Island. The two peaks of recruitment possibly involve larvae from the end of one northern breeding season (late summer - early autumn) and the beginning of the next (mid spring). (Evidence provided by satellite imagery does show that in some years the Leeuwin Current is still present in the Rottnest Island area during October). Individuals recruited during the winter months may include late arrivals at the island which have been temporarily trapped in isolated eddies of the current. Furthermore, the lack of new recruits in summer suggests that tropical reef fishes are not breeding at Rottnest Island. As mentioned earlier, the only exception to this is the blenny Omobranchus germaini, a hardy inhabitant of shallow reef flats and intertidal areas which is able to tolerate a considerable range of environmental conditions. It is widespread at the island and is also one of the few tropical reef fishes which is reasonably numerous on the coastal reefs of the nearby mainland. The most likely source of breeding stocks to the north of Rottnest Island is the Houtman Abrolhos, an area also in the path of the Leeuwin Current. Although a considerable distance south of the tropics, many of the tropical reef fishes at the Houtman Abrolhos appear to be breeding successfully (G R Allen, pers comm); this includes most of those tropical species recorded for Rottnest Island. The latter all have pelagic larvae which are capable of surviving in the plankton for moderate periods (eg larval durations of between 39 and 55 days for Thalassoma lunare and 19 to 27 days for Abudefduf vaigiensis are given by Victor 1986 and 82 Journal of the Royal Society of Western Australia, 74, 1991 Thresher et al. 1989, respectively). This should easily be enough time for the planktonic larvae to cover the 350 km between the Houtman Abrolhos and Rottnest Island at a current rate as high as 2-3 knots, even allowing for breaks in southwards movement due to meanders and eddy formation, etc. South of Rottnest Island, records of tropical reef fishes are spasmodic and widespread (Table 1), with two species being found as far east as the Recherche Archipelago. The report of Epinephelus lanceolatus from the Coorong in south-eastern South Australia by Kailola & Jones (1981) indicates the great distances that can be covered by current-borne tropical larvae. Nevertheless recruitment success for tropical reef fishes in these more southern areas appears to be poor. Therefore, the connection between Rottnest Island and the tropical richness of its fauna is primarily the offshore location of the island and the influence of the Leeuwin Current. This current flows well offshore from mainland reefs and other inshore islands, but bathes the western portion of Rottnest Island. Not only does the current carry the propagulcs of tropical reef fishes to the island, but it also maintains relatively high water temperatures during the winter. In addition, Rottnest Island provides numerous protected habitats which are conducive to settlement by tropical larvae. The corals in these areas obviously provide the shelter and food preferred by them. Some tropical species probably reach mainland reefs in the wind-blown upper surface layers of the water column, but only the very hardiest of species survive the less favourable inshore conditions. East Australian Current The presence of significant numbers of tropical reef fishes in south-eastern Australia during summer has long been an attraction for recreational divers and aquariasts (Lawler 1984, Fallu 1985). Although a relatively rich tropical fauna occurs all year round on the reefs of northern New South Wales (approx. 30°S), this fauna gradually diminishes as the latitude increases, becoming more of a seasonal phenomenon in central New South Wales (approx. 33°S), and eventually disappearing in the region of the Victorian border (latitude 37°30'S). Nevertheless some localities near the southern extremity of this range possess prominent numbers of tropical fishes during the summer months. An unpublished report (Kuiter 1981) on the fishes of Montague Island (36°15’S) listed 75 species of tropical reef fishes, and a survey by the present author in 1981 at the nearby mainland locality of Merimbula (37°S) produced 50 species. The latter investigation involved a study of the recruitment of tropical reef fishes near the entrance to Merimbula Lake. Here, the recruitment of tropical fishes commences in late November or early December, and continues until about April. These times coincide with the southwards movement of the East Australian Current, which brings warmer tropical waters down the eastern Australian coast to Bass Strait, mainly in summer (Rochford 1984). In autumn the current moves away from the coast and tropical recruitment ceases. With the onset of unfavourable environmental conditions in winter, most individuals of these tropical species disappear during May and June. At Montague Island, however, a few of the hardier species {eg the scorpaenid Pterois volitans) may persist throughout the winter, especially when water temperatures remain higher than normal (Kuiter 1981). Comparing the dispersal of tropical reef fishes between south-eastern and south-western Australia, temporal and spatial differences are obvious. In south¬ eastern Australia, recruitment takes place mostly during the summer months, with significant numbers of recruits being found as far south as latitude 37°S, whereas in south-western Australia, recruitment is more of a late autumn to winter phenomenon and virtually ceases at latitude 32°S. This can be best explained by the seasonal differences in flow of the respective currents. The summer-flowing East Australian Current transports its larvae southwards during a period of rising water temperatures, so settlement occurs when conditions are more favourable. In contrast, the Leeuwin Current transports larvae during times when water temperatures are falling, and therefore, only the hardiest are likely to survive. In addition, more propagules of tropical fishes - which generally breed during the warmer months - would be available to a summer- flowing current than a winter-flowing one. Southern Africa The Agulhas Current transports propagules of tropical reef fishes to temperate areas of South Africa in summer (Penrith 1976, Beckley 1985). This dispersal occurs regularly to Algoa Bay (34°S) but occasionally such transients are found in the vicinity of the Cape of Good Hope (Beckley et al. 1987). Generally they do not survive to adulthood in these higher latitudes, gradually disappearing with the onset of winter (Beckley 1985). Many of these tropical species are the same species also being dispersed to temperate waters of south¬ eastern and south-western Australia (Beckley et al. 1987). Tongues of the Agulhas Current can also be found off the lower west coast as isolated eddies entrapped in the cold north-flowing Benguela Current (Penrith 1976, Shannon & Agenbag 1987). However, any larvae originating from the tropical east coast carried by these eddies would be unlikely to settle given the unsuitable environment for reef fishes along the exposed sandy coasts of south-western Africa (Penrith 1976). Furthermore the cold conditions resulting from a combination of the Benguela Current and seasonal cold water upwclling would prevent survival. Species of reef fishes originating from tropical west Africa are rare in south-west Africa, given the absence of a south-flowing current and the presence of large expanses of sandy coast between the Orange River (29°S) and southern Angola (17°S) (Penrith 1976). 83 Journal of the Royal Society of Western Australia, 74,1991 South America Little has been published on the dispersal of tropical reef fishes in temperate areas of South America. However, on the west coast, the combination of upwelling and the cold north-flowing Humboldt Current would obviously restrict the southward movement of tropical forms. Briggs (1974) indicated that the influence of these conditions has pushed the southern limit of the tropical fauna almost to the equator. On the eastern side, the warm Brazil Current flows almost as far south as the waters off the Rio de la Plata (35°S) where it meets the cold north-flowing Falkland Current. Briggs (1974) stated that the Brazil Current maintains a tropical fauna to at least Rio de Janeiro (23°S). Therefore, it could be expected that the dispersal of tropical reef fishes by this current should continue well into the warm temperate regions of Brazil. Acknowledgements: I thank R H Kuiter for making available his unpublished report on the fishes of Montague Island. K Posselt and G Richards contributed much valuable data on the tropical fish fauna of Merimbula Lake, the former also providing logistical support for my work in the area. A F Pearce helped by providing details of the Leeuwin Current, whereas G R Allen kindly supplied information on the breeding seasons of tropical reef fishes in Western Australia. Data on the dispersal of tropical fishes in South African waters was generously made available by L E Beckley. Finally, I am particularly grateful to D A Pollard and DI Walker for many helpful comments on the manuscript. References Beckley L E 1985 The fish community of East Cape tidal pools and an assessment of the nursery function of this habitat. S Afr J Zool 20(l):21-27. Beckley L E, Hutchins J B & Pearce A F 1987 Range extensions of tropical fish by southward flowing boundary currents in the Indian Ocean. 6th National Oceanographic Symposium, University of Stellenbosch, South Africa (abstract only). Briggs J C 1974 Marine Zoogeography. McGraw-Hill, New York. Cresswell G R & Golding T J 1980 Observations of a south - flowing current in the southeastern Indian Ocean. Deep-Sea Res 27A:449-466. Fallu R 1985 Collecting tropical marines down south. Aquarium Life Australia 1(2):8-13. Hodgkin E P & Marsh L M 1957 On the significance of a tropical element in the littoral fauna of south-western Australia. C R 3e Congress de la P I O S A, Tananarive, section B (abstract only). Hodgkin E P & Phillips B F 1969 Sea temperatures on the coast of south Western Australia. J R Soc W Aust 52:59-62. Hutchins J B 1977 The fish fauna of Rottnest Island in relation to those of other offshore island groups in Western Australia. Unpublished B Sc Honours thesis, Murdoch University, Perth. Hutchins J B 1979 A guide to the marine fishes of Rottnest Island. Creative Research, Perth. Kailola P J &c Jones G K 1981 First record of Promicrops lanceolatus (Bloch) (Pisces: Serranidae) in South Australian waters. Trans R Soc S Aust 105:211-212. Kuiter R FI 1981 Rare visitors. Sportfishing 1981 (June/July):66-67. Lawler C 1984 Shiprock: a naturalist’s view. Skindiving in Australia and the South Pacific 14(5):23-29. Maxwell J G H & Cresswell G R 1981 Dispersal of tropical marine fauna to the Great Australian Bight by the Leeuwin Current. Aust J Mar Freshw Res 32:493-500. Michaelsen W 1908 First report upon the publications of the ’’Hamburger sudwest-Australische Forschungreise” 1905. J W Aust nat Hist Soc 5:6-25. Pearce A F 1989 The Leeuwin Current and Rottnest Island. Rottnest Island Authority, Perth. Pearce A F 1991 Eastern boundary currents of the southern hemisphere In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:35-45. Penrith M J 1976 Distribution of shallow water marine fishes around southern Africa. Cimbcbasia (A) 4:137-154. Randall J E 1968 Caribbean reef fishes. T F H Publications, Jersey City. Rochford D J 1984 Effect of the Leeuwin Current upon sea surface temperatures off south-western Australia. Aust J Mar Fresh w Res 35:487-489. Russell B C, Anderson G R V & Talbot F H 1977 Seasonality and recruitment of coral reef fishes. Aust J Mar Freshw Res 28:521-528. Saville-Kent W 1897 The naturalist in Australia. Chapman & Hall, London. Shannon L V & Agenbag J J 1987 Notes on the recent warming in the southeast Atlantic, and possible implications for the fisheries of the region. Colin scient Pap int Commn SE All Fish 1987:243-248. Smith J L B 1949 The sea fishes of southern Africa. Central News Agency, Cape Town. Thresher R E 1985 Australia’s other fishes. Sea Frontiers 31 (l):23-27. Thresher R E, Colin P L & Bell L ] 1989 Planktonic duration, distribution and population structure of Western and Central Pacific damselfishes (Pomaccntridae). Copeia 1989(2):420-434. Victor B C 1986 Duration nf the planktonic larval stage of one hundred species of Pacific and Atlantic wrasses (family Labridae). Marine Biology 90:317-326. Wilson B R & Allen G R 1987 Major components and distribution of marine fauna. In: Fauna of Australia (eds G R Dyne & D W Walton) vol 1A, Australian Government Publishing Service, Canberra, 43-68. 84 Journal of the Royal Society of Western Australia / 74,1991, 85-91 Mass spawning of corals on Western Australian reefs and comparisons with the Great Barrier Reef C J Simpson Environmental Protection Authority, 1 Mount St, Perth, WA 6000, Australia Abstract Multispecific, synchronous spawning or 'mass spawning' of scleractinian corals on Western Australian reefs was observed fortuitously in the Dampier Archipelago in March 1984. Subsequent studies during 1985-1988 have documented this phenomenon along the Western Australian coastline on tropical and temperate coral reefs and spawning has been observed to occur simultaneously on reefs separated by over 12 degrees of latitude. Mass spawning occurred mainly around the third quarter of the moon (ie 7-9 nights after the full moon) on neap, nocturnal, ebb tides. Although most of the corals studied spawned after the full moon in March each year, in some years, some spawned on the same nights after the full moon in April. Studies since 1988 have shown that a small percentage of corals spawn after the full moons in February and April each year. As a result of these studies, 102 species of scleractinian corals from Western Australian reefs are now known to spawn during the austral autumn. A further 44 species were found to contain ripe gonads during the same period and are presumed to participate in the annual coral mass spawning on Western Australian reefs. These records represent 88% of the coral species studied so far or about 46% of the coral species currently described from Western Australia. In Western Australia, coral mass spawning coincides approximately with the annual intensification of the Leeuwin Current, a warm poleward current of tropical origin that flows unidirectionally along the coastline of Western Australia during the austral autumn and winter. This current provides a mechanism for the southward dispersal of planulae and raises the possibility of a unidirectional gene flow between regionally separate coral reefs in Western Australia. Comparisons with the spring coral mass spawnings on the Great Barrier Reef indicate that, apart from the seasonal difference in the timing of spawning, many similarities exist suggesting that the same phenomenon is occurring on both sides of Australia. Comparisons of annual sea temperature patterns however, both within Western Australia and between the east and west coasts of tropical Australia, suggest that sea temperature is not a proximate or ultimate factor in determining the breeding season of scleractinian corals. It is postulated that the different seasonal timing of coral mass spawning on the cast and west coasts of Australia is the result of an endogenous rhythm reflecting the breeding patterns of ancestral corals as a consequence of selective dispersal of larvae from equatorial regions. The simultaneous spawning of many other coral reef invertebrates during the coral mass spawnings suggests that the findings presented here in relation to scleractinian corals have important implications regarding the reproductive patterns of many other coral reef animals. Introduction The recent discovery of multispccific, synchronous spawning or 'mass spawning' of tropical reef corals on the Great Barrier Reef and on Western Australian reefs has greatly increased our understanding of the reproductive patterns of many scleractinian corals (Harrison el al. 1983, Harrison el al. 1984, Simpson 1985, 1987, Babcock el al. 1986). In contrast to the brief, predictable spawning periods of corals on the Great Barrier Reef and on Western Australian reefs, corals in the Red Sea exhibit temporal reproductive isolation (Shlesingcr & Loya 1985) indicating that the reproductive traits of reef corals are highly variable. The proximate (environmental) and ultimate (ecological) factors responsible for these reproductive patterns are poorly understood. As pointed out by Willis el al. (1985), comparisons of the reproductive patterns of corals between geographic regions may 85 Journal of the Royal Society of Western Australia, 74,1991 identify common factors that will possibly pinpoint underlying causes o*f the mass spawning phenomenon. Following the fortuitous discovery of coral mass spawning in the Dampier Archipelago (Fig. 1) during March 1984, studies at the same location in 1985 confirmed that coral mass spawning occurred during a brief predictable period after the full moon in March. The coral species involved were documented and aspects of the physical environment during the spawning period were characterised (Simpson 1985). Similar studies at the Dampier Archipelago and the Ningaloo Reef in 1986 confirmed that coral spawning at these two locations occurred synchronously, and studies at the Abrolhos Islands in March 1987 determined that coral mass spawning on these temperate reefs occurred at approximately the same time as on tropical reefs in Western Australia (Simpson & Masini 1986, Simpson 1987). In addition, volunteer observers were stationed along the coastline of Western Australia in March 1987 and 1988 to determine the extent of coral spawning synchrony within and between regionally separate coral reefs in Western Australia (Simpson 1988). In this paper I summarise the results of coral mass spawning studies on Western Australian reefs between 1985 and 1988 and draw comparisons with the spring mass spawnings on the Great Barrier Reef. An hypothesis explaining the difference in the seasonal timing of the breeding seasons of scleractinian corals on the east and west coasts of Australia is also outlined. Methods Polyp reproductive status was determined by the presence of pigmented eggs either by examination of freshly broken pieces of live coral in the field or under a dissecting microscope in the laboratory. In addition to ripe eggs, the presence of testes and motile sperm with condensed heads were used as criteria for reproductive maturity (Harrison et al 1984). Spawning of corals in the field was determined directly by observation of gamete release in situ , the appearance of eggs on the sea surface, or inferred from the disappearance of mature gametes in sequential samples from tagged colonies. In most years 200+ coral colonies were tagged in the two weeks before the predicted time of spawning and their reproductive status was determined by the methods outlined above. The reproductive status of a representative subset of these colonies, usually numbering about 30-50, was monitored daily. Random collections were also made during the week before and after spawning. Untagged species observed to be spawning during night dives were also recorded or collected. In general, night observations were carried out between sunset and about 2230 h on about 6 nights around the time of predicted spawning, usually 6-11 nights after the full moon. In earlier years searches for coral eggs floating on the sea were made on several nights either side of these dates, usually between sunset and 2100 h. These combined observations, together with the daily monitoring of tagged colonies, ensured that spawning dates were determined accurately. Figure 1 Seasonal variation in the drift of surface waters on the east and west coasts of Australia. The change from a variable or northerly drift to a net southerly drift occurs around March on the west coast and around November on the east coast and coincides approximately with the timing of coral mass spawning at both locations. Details of the species and number of colonies that were sampled in the Dampier Archipelago in November 1984 and March 1985 can be found in Simpson (1985). In 1986, 512 colonies from at least 74 species were sampled at the Dampier Archipelago and the Ningaloo Reef (Simpson, unpublished data). In 1987, 422 colonics from 107 species were sampled from the Abrolhos Islands (Babcock et al. unpublished data). In 1988, 387 colonies from at least 51 species were sampled from Ningaloo Reef. As a result of these surveys, most of the coral species that were studied were sampled repeatedly at two or all of these locations. 86 Journal of the Royal Society of Western Australia, 74, 1991 Table 1 Summary of coral mass spawning observations on Western Australian Reefs during March 1987. Sea temperatures are monthly means from Pearce (1986); tidal data from Easton (1970). (s, semi-diurnal tides; d, mainly diurnal tides; * estimated tide and temperature data). Location Latitude (°S) Tide range (m) Temperature range (°Q Dates of main spawning Nights after full moon Koolan Is. 16 >10/s 31-24 23,24 8,9 Dampier Arch. 20 ~5/s 30-22 23,24 8,9 Lowendal Is.* 20 ~ 4/s 30-23 23,24 8,9 Ningaloo Reef 23 <2/s 27-23 23,24 8,9 Abrolhos Is. 28-29 ~l/d 24-20 25,26 10,11 Results Preliminary studies of coral gametogenesis, the results of tagging studies and random surveys in the weeks before and after spawning, as well as direct observations, indicate that most of the corals studied spawned after the full moon in March each year. During some years, however, a small percentage of corals with ripe gonads did not spawn after the March full moon (Simpson 1985). Observations of a small number (<10) of colonies spawning and a small spawn slick at Coral Bay (M Forde, pers. comm) after the full moon in April in 1986 suggest that these corals spawn a month later. Further observations since 1988 indicate that a small percentage of corals (usually less than 10%) spawn after the full moons in February and April each year. Spawning occurred mainly around 7-9 nights after the full moon during a period of maximum sea temperatures, within 3-4 hours after dark and during neap, ebb tides (Simpson 1985, 1987, 1988). An exception occurred in 1987 at the Abrolhos Islands where spawning occurred mainly on the 10 th and 11 th nights after the full moon. Direct observations and results from daily monitoring of tagged corals indicated that, in most years, a small percentage of the coral population spawned on one or two nights either side of the main spawning nights (Simpson, unpublished data). Simultaneous observations along the coastline of Western Australia in March 1987 indicated that a high degree of spawning synchrony exists within and between regionally separate coral reefs in Western Australia, and that spawning on these reefs occurred under markedly different environmental conditions (Table 1). Similar observations in 1988 support these conclusions (Simpson 1987). A total of at least 165 species of scleractinian corals (41 genera, 13 families) were examined in the periods following the full moon in March between 1985-1988 (Table 2). These studies indicate that at least 102 species of corals that occur on Western Australian reefs shed gametes during a brief, predictable spawning period each year. A further 44 species contained ripe gonads, indicating that spawning was imminent and suggesting that these species also participate in the mass spawning. The total of definite and probable spawning species represents approximately 46% of the 318 species of scleractinian corals, from 70 genera, currently described from Western Australia (Veron & Marsh 1988). A high percentage (88 %) of the coral species that were examined were either definite or probable spawners suggesting that many of the other unexamined coral species that occur on Western Australian reefs are likely to be involved in the annual mass spawning event. None of the coral species that were examined contained or released planulac during this period and mature gonads were not observed in any corals collected during November 1984 or May to October 1985. In addition, preliminary data on the gametogenic cycle of species of Acroporidae, Faviidae, Mussidae and Oculinidae at Ningaloo Reef between 3 October 1985 and 26 March 1986 suggest that the gametogenic cycles of these species are similar to the same or like species on the Great Barrier Reef (Marshall & Stephenson 1933, Kojis & Quinn 1981, Harriot 1983, Babcock 1984, Wallace 1985) although it is offset by 4-5 months. 87 Journal of the Royal Society of Western Australia, 74, 1991 Table 2. Comparison of the number of coral species found to be definite or probable autumn spawners on Western Australian reefs in relation to the total number of species currently known from Western Australia (* from Veron & Marsh 1988) and the number of species examined. FAMILY Number of species Number of species Number of species Number of specie: in Western Australia* examined that spawned with ripe gonads ASTROCOENIIDAE 2 1 0 0 POCILLO PORI DAE 9 1 0 0 ACROPORIDAE 98 60 50 7 PORITIDAE 30 16 6 7 SIDERASTREIDAE 12 5 0 2 AGARICIIDAE 20 7 1 3 FUNG1IDAE 28 4 1 3 OCULINIDAE 3 2 1 1 PECTINIDAE 11 6 4 2 MUSSIDAE 17 8 4 3 MERULINIDAE 7 6 2 3 FAVIIDAE 59 42 30 9 TRACHYPHYLLIIDAE 2 0 - - CARYOPHYLLIIDAE 9 1 1 - DENDROPHYLLIIDAE 11 6 2 4 TOTAL 318 165 102 44 Reproductive swarmings of polychaete worms, predominantly rag-worms (Polychaeta: Nereidae) but including many other species such as Eunice cf australis, occurred simultaneously with the coral mass spawnings in all instances. The epitokous (reproductive) stage of these worms emerged following the onset of coral spawning. Other taxa observed spawning during this period include Alcyonarians and species of Mollusca and Echinodermata (Marsh 1988). Discussion There are many similarities between the coral mass spawning events observed on the Great Barrier Reef and on Western Australian reefs. Most species are simultaneous hermaphrodites and spawn after a full moon, over 2 or 3 consecutive nights, during a period of neap tides and within 3 to 4 hours after sunset. Additionally, mass spawning appears to be a predictable, annual event in both locations and approximately 60% of the species observed to spawn or contain ripe gonads in Western Australia mass spawn on the Great Barrier Reef (Harrison et al. 1984, Babcock el al. 1986, Shlesinger & Loya 1985, Simpson 1985, 1986). The colours, general buoyancy (Babcock et al. 1986) and size range of mature eggs (Marshall & Stephenson 1933, Kojis & Quinn 1981, Harriot 1983, Babcock 1984, Wallace 1985) and the most common spawning behaviour are also similar (Babcock et al. 1986). The synchronous spawning of corals on regionally separate reefs (Table 1, Babcock et al. 1986) and the time between consecutive spring spawnings on the Great Barrier Reef and autumn spawnings on Western Australian reefs (Willis et al. 1985, Babcock et al. 1986, Simpson 1988) are further similarities. The remarkable likeness of coral spawning on the east and west coasts of Australia suggests that the same phenomenon is being observed. Different seasonal timing of mass spawning, and as a consequence, the different environmental conditions that exist during the periods of gametogenesis, spawning, larval development and settlement are the most significant differences between the two locations. For example, mass spawning on the Great Barrier Reef occurs after a period of rapidly rising sea temperatures, although these temperatures are still well below the maxima for these locations (Fig. 2a). In contrast, spawning on Western Australian reefs coincides with the period of maximum seawater temperatures (Fig. 2a, b). There are further differences. Corals on offshore reefs in tropical Western Australia ( eg Lowendal Is) spawn at the same time as corals on inshore reefs (eg Dampier Archipelago, Table 1), in contrast to the one month difference that occurs between Magnetic Island and the offshore reefs on the Great Barrier Reef (Babcock et al. 1986). Corals on the Great Barrier Reef also spawn around neap tides but, because of the aphasic tides in tropical east Australia, spawning occurs mainly three to six nights after the full moon. As a result, spawning on the Great Barrier Reef occurs before, during and after moonrise (Babcock et al. 1986) whereas in Western Australia, corals spawn well before moonrise (Simpson 1985). A further difference is that species of Turbinaria participate in the mass spawning 88 Journal of the Royal Society of Western Australia, 74, 1991 mainly three to six nights after the full moon. As a result, spawning on the Great Barrier Reef occurs before, during and after moonrise (Babcock et al. 1986) whereas in Western Australia, corals spawn well before moonrise (Simpson 1985). A further difference is that species of Turbinaria participate in the mass spawning event in Western Australia but not on the Great Barrier Reef where spawning of this genus occurs in autumn (Harrison et al. 1984). These differences possibly reflect adaptation to local conditions. Coral mass spawning on Western Australian reefs coincides approximately with the annual intensification of the Leeuwin Current, a warm poleward current of tropical origin that flows unidirectionally along the Western Australian coastline in autumn and winter (Cresswell & Golding 1980, Godfrey & Ridgway 1985). This current flows predominantly, but not exclusively, during the austral autumn and winter and provides a mechanism for the southward dispersal of coral planulae which, in turn, raises the possibility of a unidirectional gene flow between regionally separate coral reefs in Western Australia. The reproductive cycles of marine invertebrates are considered to be influenced predominantly by temperature (Orton 1920, Giese & Pearse 1974) and this factor has been suggested as having an important influence in timing the gametogenic cycle of scleractinian corals (Harrison et al. 1984, Babcock et al 1986, Oliver et al. 1988, Richmond & Hunter 1990). Furthermore, differences in temperature patterns have been suggested as possible explanations for observed differences in the timing of coral spawning on the northern (Harriot 1983) and southern (Kojis & Quinn 1981) Great Barrier Reef and between the onshore and offshore reefs on the central Great Barrier Reef (Babcock et al. 1986). Seawater temperatures at the Dampier Archipelago show a pronounced seasonal pattern and are similar to sea temperatures at Magnetic Island, yet the breeding seasons of corals at these two locations are about 5 months apart (Fig. 2a). In contrast, sea temperature patterns along the Western Australian coastline are quantitatively and qualitatively different due to the varying influence of the Leeuwin Current, yet mass spawning occurs synchronously between regionally separate reefs (Fig. 2b, Table 1) and between onshore and offshore reefs (Table 1). These data suggest that sea temperature is not a universal proximate cue or ultimate factor in determining the timing of the breeding season of scleractinian corals. An alternative hypothesis is that the breeding season of corals in Australia is not controlled by environmental cues but is the result of an endogenous rhythm which reflects historical breeding patterns of ancestral corals (Simpson 1987, 1988) and has been termed the Genetic Legacy Hypothesis by Oliver et al. (1988). Ocean circulation patterns at the time of coral spawning on the east and west coasts of Australia and preliminary data on the breeding season of corals in equatorial regions support this hypothesis. Mass spawning of corals occurs in late spring/early summer on the Great Barrier Reef and during autumn on Western Australian reefs. In both locations spawning coincides approximately with the annual change from a northerly or variable net drift to a pronounced net southerly drift of surface waters (Fig. 1) and with periods of calms associated with the seasonal changes in monsoonal wind patterns (Pickard et al , 1977, Williams et al. 1984, Holloway & Nye 1985). The southerly drift of tropical water along the Great Barrier Reef originates from the South Equatorial Current and occurs during late spring and summer whereas during autumn and winter surface waters of the Great Barrier Reef drift northward under the influence of the south-east trade winds. Off Western Australia the opposite situation exists. The Leeuwin Current, originating as inflow from the Western Pacific to the Indian Ocean through the Indonesian Archipelago (Godfrey & Ridgway 1985), flows strongly southward along the coastline during autumn and winter. Weaker flows and periodic reversals occur during late spring and summer when south-westerly winds predominate. Thus, in both areas, albeit in different seasons, a mechanism exists for the southward transport of coral larvae from equatorial regions. 32 # /\MI o 30 \ o 9, 28 V / DA ' - 0 ALK ^ s —- ~ ws CM i_i_i_i ~ i 0 4 8 12 16 20 24 0 4 8 12 16 20 24 Lag (months) Figure 6 Cross correlations of pueruli settlement of Panulirus cygnus with Fremantle sealevel (FMSL) and water temperature (SST). The cross correlation functions of pueruli for 1984- 1989 with sealevel and water temperature are shown in Fig. 6. There was a broad trend of smaller lags (1-3 97 Journal of the Royal Society of Western Australia, 74, 1991 Table 2 The maximum cross correlations of pueruli with the environmental series (seasonal effects removed) showing maximum lags. SOI Sealevel SST Salinity East wind North wind Site max cross lag (mths) max cross lag (mths) max cross lag (mths) max cross lag (mths) max cross lag (mths) max cross lag (mths) corr corr corr corr corr corr SP .23 3 .32 10 .33 9 -.33 15 -.26 0 .24 0 Hor .24 8 .32 8 .38 4 -.37 1 -.29 6 -.23 8 Rat .35 0 .30 8 .43 9 -.24 5 -.21 1 -.27 8 SMB .27 3 .29 10 .53 4 -.37 6 -.36 3 -.29 8 Jur .51 3 .51 3 .51 1 -.57 0 .17 6 -.28 4 Cer .49 3 .50 3 .62 1 -.48 0 .27 11 -.42 1 Aik .30 1 .50 1 .50 4 -.38 0 .25 5 -.35 3 WS .43 1 .47 1 .45 1 or 5 -.31 0 .21 5 -.37 3 CM .50 3 .56 0 .46 3 -.52 0 .19 4 -.34 4 months) for the southern sites (Cape Mentelle, Warnbro Sound, Alkimos, Cervantes, Jurien Bay), with sealevel and water temperature to larger lags (4-10 months) for the northern sites (Seven Mile Beach, Rat Island, Horrocks Beach, Shark Bay). This trend is also evident with salinity. A similar cross correlation analysis of the longer time settlement data from Seven Mile Beach and Jurien Bay against sealevel showed that the maximum cross correlation (0.27) occurred when settlement at Jurien Bay lagged Fremantle sealevel by four months and when Seven Mile Beach lagged Fremantle sealevel by four months (0.26). Satellite imagery We also attempted to relate mesoscale (order 100 km) features of the Leeuwin Current as revealed by satellite images to fluctuations in puerulus settlement along the coast. Pearce & Phillips (1988) suggested, for example, that a strong meander off Geraldton in October 1984 may have contributed to the record settlement at Seven Mile Beach in that month. While there is little doubt that meanders offshore of the Leeuwin Current can rapidly bring phyllosoma larvae towards the shelf (Pearce & Phillips 1992), current measurements along the shelf break have shown that the alongshore flow is very strong (Cresswell et al. 1989). It now seems more likely, therefore, that the pueruli will be carried both south and north by the reversing current system on the shelf and so distributed over a larger area of coastline. Along the inshore boundary of the Current, settlement may be aided or hindered by variations in the currents or water temperature near the coast. For example, zero settlement at Cape Mentelle in 1986 and 1987 (both ENSO years when low coastal sealevels indicated a weak Leeuwin Current) was possibly associated with the Leeuwin Current being along the outer shelf and a cool inshore counter-current in that area (Fig. 7a). In 1988, on the other hand, there was appreciable settlement at Cape Mentelle, and satellite imagery for November 1988 showed warm Leeuwin Current water flooding onto the shelf (Fig. 7b). Discussion Spatial and temporal settlement of puerulus It is clear that settlement along the coast of Western Australia follows a similar seasonal pattern at all sites and that in years when settlement occurs there is a close correlation between adjacent sites. However, the 98 Journal of the Royal Society of Western Australia, 74, 1991 Figure 7 Satellite images of South-Western Australia from the NOAA Advanced Very High Resolution Radiometer (AVHRR): (upper) orbit N9/15219 on 26 November 1987, and (lower) orbit N11/646 on 9 November 1988. The Leeuwin Current is evident as the warm (pale) water along the shelf. The area shown is from Fremantle (in the north) to south of Cape Leeuwin; Cape Mentelle is midway between the two prominent capes (Cape Naturaliste and Cape Leeuwin). Images courtesy of the Western Australian Satellite Technology and Applications Consortium (WASTAC). 99 Journal of the Royal Society of Western Australia, 74, 1991 southern sites receive lower levels of settlement than in the north, and in some years have zero settlement. Both 1986 and 1987, the two years in which no settlement was recorded at the southern sites, were ENSO years when the Leeuwin Current was weaker. Settlement to the southern area may only occur (or be higher) when the Leeuwin Current is strong and running close inshore. Catches in the southern area of the fishery fluctuate much more than in the central and northern areas of the fishery, supporting this hypothesis. Correlation between environmental events and levels of pueruli settlement Only preliminary conclusions can be drawn from the results in view of the short time span of the data. Phillips (1986) recorded fluctuations in the annual index of puerulus settlement at Seven Mile Beach ranging from 15.9 to 182.8 over a 14-year period. Between 1984 and 1989 the values ranged from 60.3 to 128.4, hence it is difficult to assess the exact reliability of the cross correlations recorded. However, the strong cross correlations of SOI, sealevel, sea surface temperature and salinity with settlement at all sites indicate that similar forces are affecting settlement along the whole coast. This is in line with the conclusions of Pearce & Phillips (1988) who suggested that the oceanic processes influencing settlement along the Western Australian coastline operate at length scales of the order of hundreds of kilometres and time scales of many months to a year. Of the environmental factors examined in this study, sealevel, sea surface temperature and salinity may all reflect interrelated expressions of monthly to annual variations in the strength or temperature of the Leeuwin Current. Some of the correlations are curious. It is, for example, difficult to imagine the basis for the between- season time lags of puerulus settlement where the northern sites, such as Seven Mile Beach, follow settlement at the southern sites nine to ten months later. However, the long term (20-year) data from Seven Mile Beach and Jurien Bay show the same picture. This may point to previously unrecognised linkages in oceanic events. Although Pearce & Phillips (1988) demonstrated that there is a clear link between the Leeuwin Current and larval recruitment, the mechanisms which act on the larval or puerulus stage to bring this about arc as yet unknown. This is partly because of a lack of information about the site of transformation to, and thu behaviour of, the puerulus stage, as well as variability of the oceanic environment. Further studies of both thes^ aspects are necessary. Acknowledgements Tidal data for Fremantle were supplied by the Tid^l Laboratory, Flinders University of South Australia, Copyright reserve^ (courtesy Mrs Jean Bye). Dr Paul Stewart of the Bureau of Meteorology provided the wind data. The technical staff of both the CSIRO Marino Laboratories and the Western Australian Marine Research Laboratories bore the burden of the routine puerulus collections. We are grateful to Dr Nick Caputi for helpful comments. References Boland F M, Church J A, Forbes A M G, Godfrey J S, Huyer A, Smith R L & White N J 1988 Current-meter data from the Leeuwin Current Interdisciplinary Experiment CSIRO Mar Lab Rep 19831pp. Cress well G R, Boland, F M, Peterson J L & Wells G S 1989 Continental shelf currents near the Abrolhos Islands, Western Australia. Aust J Mar Freshw Res 40: 113-128. Diggle P 1990 Time Series A Biostatistical Introduction. Oxford Science Publications. Morgan G R, Phillips B F & Joll L M 1982 Stock and recruitment relationships in Panulirus cygnus, the commercial rock (spiny) lobster of Western Australia. Fish Bull 80: 475- 486. Pearce A F & Phillips B F 1988 ENSO events, the Leeuwin Current, and larval recruitment of the western rock lobster. J Cons int Explor Mer 45:13-21. Pearce A F & Phillips B F 1992 Oceanic processes, puerulus settlement and recruitment of the western rock lobster Panulirus cygnus. Proc Boden Conf, Thrcdbo NSW, Feb 5-7,1990 (in press). Phillips B F 1972 A semi-quantitative collector of the puerulus larvae of the western rock lobster Panulirus longipes cygnus George (Decapoda, Palinuridae). Crustaceana 22:147-154. Phillips B F 1981 The circulation of the southeastern Indian Ocean and the planktonic life of the western rock lobster. Oceanogr & Mar Biol Ann Rev 19:11-39. Phillips B F 1986 Prediction of commercial catches of the Western Rock Lobster Panulirus cygnus. Canadian J Fish & Aquatic Sci 43: 2126-2130. Phillips B F, Brown P A, Rimmer D W and Reid D 1979 Distribution and dispersal of the phyllosoma larvae of the western rock lobster, Panulirus cygnus , in the South¬ eastern Indian Ocean. Aust J Mar Freshw Res 30: 773- 783. Phillips B F & Hall N G 1978 Catches of puerulus larvae on collectors as a measure of natural settlement of the western rock lobster Panulirus cygnus George. CSIRO Aust Div Fish Oceanogr Rep 89:1-18. Prata A J, Pearce A F, Wells J B & Carrier J M 1986 Satellite sea surface temperature measurements of the Leeuwin Current. Proc 1st Aust AVHRR Conf Perth October 1986: 237-247. Troup A J 1965 The southern oscillation. Quart J Roy Meteor Soc 91:490-506. 100 Journal of the Royal Society of Western Australia / 74, 1991, 101-114 The influence of the Leeuwin Current on coastal fisheries of Western Australia R C Lenanton 1 , L Jolt 1 , J Penn 1 and K Jones 2 1 Western Australian Marine Research Laboratories, PO Box 20, North Beach, WA 6020, Australia 2 South Australian Department of Fisheries, GPO Box 1025, Adelaide, SA 5001, Australia Abstract The fisheries of the three major southern hemisphere eastern boundary currents are briefly reviewed. In all three systems, physical environmental variables influence fish catches in a major way. Both the Benguela and Humboldt Currents create highly productive and dynamic upwelling ecosystems that are characterised by a succession of dominant finfish species, which can individually support substantial commercial fisheries. By contrast, the Leeuwin Current waters off the west coast of Australia are characterised by low biological productivity. Although a group of finfish species, almost identical to those of the Benguela and Humboldt Current ecosystems is represented in the Leeuwin Current ecosystem, the collective Western Australian catches of these species arc insignificant by world standards. Indeed the major commercial species of this region are demersal invertebrates, some of which (eg rock lobster) support fisheries of international significance. Thus the Leeuwin Current does exert a major influence on the overall ecology of this unique region, and affects the production of both economically important finfish and invertebrate species. Introduction Historically four major eastern boundary current systems were recognised in the world oceans (Wooster & Reid 1963). They comprise very large spatial systems which exhibit unique bathymetry, circulation, biological productivity and trophodynamic relationships of populations. Two of these, the California and Canary Currents, are located in the northern hemisphere, while the other two, the Humboldt and Benguela, are located in the southern hemisphere (Parrish el al. 1983, Weaver 1990). Both the Humboldt and Benguela eastern boundary currents are part of oceanic-scale wind-driven anticyclonic gyres in the southern hemisphere. Because the equatorward flow of water is in the same direction as the prevailing wind, upwelling is associated with these coastal currents (Cushing 1971, Shannon 1985, Bohlc-Carbonell 1989). In recent years, however, a fifth and uniquely different eastern boundary current system has been recognised (Golding & Symonds 1978, Cresswell & Golding 1980, Pearce & Phillips 1988, Pearce & Prata 1989). In contrast to the other four eastern boundary current systems, the Leeuwin Current is driven poleward by a deep alongshore density gradient, whose existence is partly dependent on the flow of warm western equatorial Pacific water through the Indonesian Archipelago (Weaver 1990). The lack of upwelling associated with this current is because of the eastward flow of Indian Ocean water despite the prevailing southerly winds (Godfrey & Ridgway 1985). The life history characteristics, such as spawning, migration, recruitment and feeding patterns, and ultimately the overall production of many ecologically important finfish species have evolved under the influence of such current systems. For example, the upwelling of cool nutrient-rich water, which is a most important characteristic of the Humboldt/Benguela systems leads to high rates of primary production it high biomass of phytoplankton and zooplankton (Cushing 1971, Armstrong et al. 1987, Chavez et al. 1989). This accounts for the substantial populations of pelagic planktivorous fishes found in both of these upwelling systems (Crawford et al. 2983, Parrish et al. 1983, Crawford et al. 1987, Crawford 1987). Indeed the commercial catches of pelagic species from these regions are very significant in the context of world fish production (FAO 1988). By contrast, the Leeuwin Current consists of warm low nutrient water flowing into continental shelf waters. 101 Journal of the Royal Society of Western Australia, 74,1991 Table 1 The prominent families of finfish which comprised the commercial catch from southern hemisphere eastern boundary currents systems during the 1980's. Southern Hemisphere Eastern Boundary Current Systems Finfish Families Humboldt Benguela Leeuwin Clupeidae (true herrings) Sardine Pilchard Round herring Sardinella Pilchard Round herring (maray) Sardinella Engraulididae (anchovy) Anchoveta Anchovy Anchovy Carangidae (trevally) Horse mackerel Horse mackerel = Jack mackerel (or scad) Scombridae (mackerel) Mackerel Bonito Mackerel Bonito Mackerel Bonito Merluccidae (hake) Hake Hake ? (offshore) Gemplylidae (snoek) Snoek Barracouta (or snoek) OVERALL ANNUAL <1-13* CATCH (Million tonnes) <1-4* <0.001+ Source * FAO, 1988, Crawford et al , 1987 + Anon. 1990 which, by Humboldt/Benguela standards, are already low in nutrients (Rochford 1980,1988, Pearce et al. 1985). Although similar pelagic planktivorous fish species are represented in the Leeuwin Current system, the commercial catches of these species are far smaller than those of similar species taken from the Benguela and Humboldt upwelling regions (Table 1). Indeed, demersal species (particularly rock lobster and prawns), that are dependent on benthic production dominate commercial catches taken from the Leeuwin current (Anon. 1990). Effect of the Humboldt and Benguela Currents on Associated Fisheries Environmental change, rather than factors such as recruitment overfishing, predation or pollution, has been identified as the major variable controlling large scale changes in fish abundance in all eastern boundary currents (Sherman 1987). Understanding the complex processes through which, in an overriding sense, climate (Cushing 1982, Sharp 1987) and more specifically the physical environmental properties of the current systems can ultimately affect commercial fish catches is of major importance to the managers of such fisheries. Principal abiotic properties of the current environment that can lead directly to changes in fish abundance (and therefore catches) include: current strength and direction, current motion (or turbulence), water temperature, water salinity, and dissolved oxygen. For example, temperature, turbulence and transport patterns influence the location of spawning grounds and the breeding period of anchovies and pilchards in those eastern boundary currents characterised by upwelling (Parrish et al. 1983). 102 Journal of the Royal Society of Western Australia, 74, 1991 Certainly, instances of strong recruitment for neritic stocks in the southern Benguela system have been linked to environmental anomalies (Crawford et al. 1983). Specifically, sea surface temperature and anchovy recruitment have been positively correlated (Boyd 1979). Sea temperature has also been shown to influence directly the survival of pilchard and anchovy eggs (King 1977). Furthermore, a shift of the predominant species from anchoveta to sardine between 1970 and 1983 resulted in a dramatic increase in the yield of sardine off Chile and Peru (Cushing 1982). This has been partially attributed to El Nifto producing higher sea surface temperatures which in turn reduced the anchoveta habitat size (Muck 1989a), and thereby made them more vulnerable to fishing pressure (Cushing 1982, Csirke 1989). The above abiotic factors can also indirectly affect abundance of important commercial fish species by influencing their food supply, competitors and predators (Wooster & Bailey 1989). For example, high sea surface temperatures associated with El Nino events can indirectly reduce anchoveta abundance by increasing the density- dependent mortality on eggs and larvae, increasing metabolic cost and reducing food availability (Muck 1989a). Moreover, the intrusion of warm oxygen-rich waters from the north into the Humboldt upwelling system during El Nifio events, led to hake extending further south, and thus invading the main anchoveta area. This allowed increased anchoveta predation of hake eggs (Muck 1989b). Effect of the Leeuwin Current on Fisheries Productivity The continental shelf waters off Western Australia are relatively low in nutrients (Pearce et al. 1985) and relatively clear. As a result of the Leeuwin Current, the overall temperature range in the region of its influence is also relatively small. Temperatures are therefore appreciably warmer than at comparable latitudes in other eastern boundary current regions (Pearce 1991). Because of the shape of the Western Australian coastline, and in particular variation in the width of the continental shelf, the impact of the Leeuwin Current appears to be greater on some sections of the shelf than others. Satellite imagery has suggested that the current approaches the coastline between North-West Cape and Shark Bay, in the Geographe Bay/Cape Naturaliste/Cape Leeuwin region, and along the south coast from Pt. D’Entrecasteaux to about Albany (Pearce 1985). In addition, islands near the edge of the shelf, such as the Abrolhos Islands and Rottnest Island, are particularly affected by the warm current waters (Hatcher 1991, Hutchins 1991). Because upwelling is not a feature of the current system, nutrient levels in the coastal waters are largely dependent on terrestrial inputs. Run-off from the largely arid hinterland is particularly low, with the limited river outflow mostly from the south-western region of the State being confined almost entirely to winter/spring (Lcnanton & Hodgkin 1985). The relatively clear coastal waters landward of the Leeuwin Current, which include the large marine embayments of Shark Bay and Geographe Bay, are typified by extensive seagrass meadows and macroalgae dominated coastal reef systems (Kirkman 1985, Walker 1991). Coastal finfish resources of the state are generally confined within these water masses landward of the main current and are largely dependent on the relatively productive estuarine and protected coastal marine ecosystems (Lenanton 1982, Robertson & Lenanton 1984, Lenanton & Potter 1987). Figure 1 The limits of distribution of the Western Australian salmon. Hatching indicates the region of most intensive spawning. In addition to these major fisheries, there are also tropical species appearing in the commercial and recreational catch off the west coast which are dependent on the seasonal flow of the Leeuwin Current (Maxwell & Cresswell 1981, Hutchins 1991). The major invertebrate resources of Western Australia, the western rock lobster Panulirus cygnus (Phillips & Brown 1989), and penaeid prawns Penaeus esculentus and Penaeus latisulcatus (Penn 1981), are similarly dependent on the extensive inshore and relatively productive macrophyte zones. Fisheries Affected by the Leeuwin Current Almost all of the major economically important fish stocks in waters off the western and southern coasts of Western Australia are influenced to some extent by the Leeuwin Current. As will be shown below, fisheries which are specifically affected by the current are (i) the Western Australian salmon and herring fisheries off the south and lower west coasts; (ii) the pilchard purse seine fishery off the south coast; (iii) the western rock lobster fishery off the west coast; (iv) the saucer scallop fisheries in Shark Bay and other areas off the west and 103 Annual Commercial Catch ( Tonnes Live Wt. x 10 Journal of the Royal Society of Western Australia / 74,1991 1940 I \ f kA Q c/ \ I k \ \ I 1,4 o \/\ 0 \ '. JD O / I i ef 1950 r — I — '— T — T — 1 — T — T — T — T — 1 I ' 1960 1970 YEAR 1980 19 ? Figure 2 The annual Western Australian commercial catch of Western Australian salmon between 1944 and 1989. south coasts; and (v) the penaeid prawn fishery in Shark Bay. They include scaly mackerel ( Sardinella lemura), dusky (bronze) whaler shark ( Carcharhinus obscurus), Spanish mackerel ( Scomberomorus spp.) and baldchin groper ( Choerodon rubescens) in marine waters; and bar-tailed flathead ( Platycephalus endrachlensis) and giant herring ( Flops machnata) in the Swan river system. A large component of the finfish catch from the Abrolhos Islands is also made up of tropical species such as narrow-barred Spanish mackerel (Scomberomorus commerson), cod ( Epinephelus spp.), coral trout ( Plectropomus spp.) and baldchin groper. Western Australian salmon fishery Western Australian salmon ( Arripis truttaceus ) is a large pelagic inshore schooling species (Malcolm 1960). It is distributed from Kalbarri on the mid-west coast of Western Australia to about Victoria and western Tasmania on the south coast of Australia (Fig. 1) (Stanley 1980a, Hutchins & Swainston 1986) where it supports substantial commercial net and recreational angling fisheries (Stanley 1980a, Walker 1982, Cappo 1987). The major Western Australian fishery for salmon is located off the beaches between Geographe Bay and just east of Bremer Bay (Stanley 1980a, Walker 1982). All fish located east of the western Great Australian Bight are immature (Malcolm 1960, Stanley 1980a, 1980b), while all the mature fish are located in Western Australian waters. Spawning commences during March, and reaches a peak during early April (Malcolm 1960, Nicholls 1973). It is postulated that large numbers of fertilised eggs and larvae are transported east by the Leeuwin Current to inshore protected nursery grounds located between the western Bight and Victoria (Cresswell 1986, Malcolm 1960, Robertson 1982). Juveniles first appear in the Western Australian nursery grounds in April (Lenanton 1977, 1982), and in 104 Journal of the Royal Society of Western Australia, 74, 1991 South Australia, nursery grounds in June (Jones et al. unpublished). Although juveniles are first fished commercially as 1 year old fish in South Australia (Stanley 1979, Cappo 1987), there is very little commercial exploitation of juveniles in Western Australia (Walker 1982). Fish tend to mature according to size rather than age, and grow much faster in Western Australia than in eastern Australia (Nicholls 1973, Stanley 1979,1980b). At about the end of January/early February, mature and maturing fish migrate from waters adjacent to their nursery grounds west to spawn off the lower west and south coasts of Western Australia (Malcolm 1960, Stanley 1980a). These fish form the basis of the Western Australian commercial and recreational fishery (Walker 1982, Cappo 1987). At the beginning of each season, the Western Australian catch comprises mainly larger resident fish (Malcolm 1960, Stanley 1980a). By about mid March, smaller new recruits dominate the catch (Stanley 1980b, Cappo 1987). Preliminary modelling of the Western Australian salmon fisheries has revealed that stock abundance appears to be dependent mostly on the magnitude of annual recruitment. Major peaks in annual Western Australian catch, in particular those in the late 1960's and early 1980's (Fig.2) are thought to be related to periods of strong recruitment from Western Australian nursery areas (C. Walters, R.C.J. Lenanton and M. Cappo, unpublished). Furthermore environmental change influenced by the Leeuwin Current, rather than fishing, appears likely to be one of the main factors affecting recruitment. Figure 3 Central and western coastline of South Australia showing sites where sealevel is routinely monitored, together with sites where juvenile salmon abundance is measured. Preliminary evidence from South Australia suggests that the Leeuwin Current assists the recruitment of salmon to South Australian nursery areas. First a significant positive correlation has been demonstrated between the Southern Oscillation Index (SOI) and mean annual sealevel at Thevenard (r 2 = 0.42, 0.025

0.05), Port Lincoln (r 2 =0.35, 0.0025

0.005), Port Adelaide (Outer Harbour) (^=0.46, 0.005

0.025) and Victor Harbor (^=0.50, P=0.01) (Fig. 3). In years of low SOI (and weak Leeuwin Current) (eg 1982), relatively low sealevels occurred during June to September w r hen 0+ year old salmon are being distributed across the Great Australian Bight and entering the nursery areas of the west coast bays. South Australian Gulfs and the Coorong waters (Fig.4). In years of high SOI (and strong Leeuwin Current) (eg 1981), the sealevel during these months was relatively high- An annual recruitment index (natural log numbers (In n)) of 0+ year old salmon is available between 1980 and 1990 from the waters of Barker Inlet, adjacent to the Outer Harbour in South Australia (Fig.3) (G.K. Jones, G. Wright & K. Edyvane, unpublished). Further analysis revealed a significant positive correlation between sealevel in August (the usual month when salmon enter the South Australian nursery areas ) at Outer Harbour, and the annual recruitment Index of salmon in Barker Inlet (Fig.5). The commercial catch of salmon in waters of the Coorong comprises mainly 1+ year old fish (Cappo 1987). There is also a significant positive correlation between the commercial catches of 1+ year old salmon in the Coorong waters and the adjacent Victor Harbor August sealevel measured one year earlier (Fig.6). Thus, there is strong circumstantial evidence for a direct, relatively short-term (up to 6 months) process of transportation of larvae from Western Australian spawning areas to South Australian nursery areas. Furthermore, there are indications that both current strength and timing of peak flow are likely to have an important influence on the strength of recruitment to South Australian nursery areas. Attempts to relate mean sealevel at Albany at the time of spawning to the subsequent recruitment of maturing salmon into the Western Australian commercial fishery (ie between 4 to 6 years later) have to date been unsuccessful because of: 1) the complicated size-dependent recruitment process 2) variable annual rates of fishing and natural mortality during the relatively long period of four or more years leading up to recruitment into the fishery. However, there are indications that commercial catches may be adversely affected in years of strong Leeuwin Current flow along the south coast of Western Australia. Preliminary analyses have revealed a negative correlation between the mean monthly sealevel at Albany over the period of the fishery (February - April) and the mean annual south coast log book catch per hour of beach observation in that same year (Fig.7). These data were treated as two separate 105 Journal of the Royal Society of Western Australia / 74,1991 Figure 4 The average monthly sealevel at Thevenard, South Australia during years of weak (1982) and strong (1981) SOI. groups: those from years of low catches (=abundance) and those from years of high catches. The suggestion, supported by observations from commercial fishers, is that in years of strong Leeuwin Current flow, local storm events modify the pattern of the current flow, resulting in "patches” of warmer Leeuwin Current water adjacent to the shoreline. Thus, migrating salmon are forced offshore and deeper, in order to avoid these cells of warmer water, and at such times are not available for capture on beaches located along the affected shoreline. Thus during years of strong Leeuwin current flow the catchability of the stock is reduced. Indeed, log book records kept by two demersal shark gill net fishers (Table 2) show clearly that salmon occurred within 3 m (the approximate depth of the nets) of the bottom in water depths of up to 57 m. Further, a by-catch of salmon has consistently been recorded in the monthly catch and effort returns of a number of demersal gill net fishers who operate in a variety of areas between Geographe Bay and Cheynes Beach (Fig.l). Although the above data suggest that the Leeuwin Current is the major factor influencing larval transport and distribution, the precise details of the process are not known. Observations by fishers suggests that salmon spawn close to the coast. However the Leeuwin Current usually flows along the shelf break. Thus if the current is in fact the medium of larval transport, then either the fish must spawn in locations where the current is close to the coast (eg in the Cape Naturaliste region), or other factors such as local weather conditions must contribute to the transport of fertilised eggs offshore into Leeuwin current waters. If the timing of peak current flow coincided with peak spawning, maximum numbers of larvae could be expected to be transported via the current. Then there is the question of what factors influence the relative size of recruitment to the different shoreline nursery areas located off the southern Australian coast. Do local weather conditions play an important role? Are the larvae transported in a frontal system associated with the current? How are potential competitors/predators affected by the current? Clearly the processes involved are only just beginning to be understood. 106 Journal of the Royal Society of Western Australia, 74,1991 I M.0 r =0.64 0.01 > p > 0.05*’ 84 / / / / / / / 85 / / / / / 83 / / / ' I ' 15 j 0 '"I ' 155 17.0 18.0 Outer Harbour August Sea Level Height (m) Figure 5 Relationship between the annual recruitment index (In n) of salmon in Barker Inlet, and the sealevel in August at Outer Harbour over the period 1979-89 (83 = year of data; • ENSO years O Non-ENSO years). 1 - 1 16 + o 1 c 5 « r *=0.54 0.05>p> 0.025* 82 78 o 79 o 77 75 O 540 550 560 590 600 610 630 610 650 670 Victor Harbour August Sea Level Height (mm) Figure 6 Relationship between the annual commercial catch of 0+ year old salmon caught in the Coorong between 1976/77 and 1984/85 and the Victor Harbor sealevel in August one year earlier (83 = year of data; • ENSO years O Non-ENSO years). 107 Journal of the Royal Society of Western Australia, 74, 1991 Table 2 Western Australian salmon by-catch records extracted from log books kept by two demersal longline and demersal gill net limited entry fishers. Date Fishing area (see Fig. 1) Location Depth (m) Surface water temp (°C) WA salmon catch 1990 290390 Augusta 34°43'S 50 20.1 500 kg 115°18.4'E 300390 Augusta 34°38.5'S 47 21.2 + 115°14.6'E 070490 Augusta 34°43’S 50 21.2 + 115°15.3‘E 080490 Augusta 34°41.6’S 47 21.6 + 115°16.5'E 130490 Augusta 34°27'S 38 20.3 + 115°27.2'E 140490 Augusta 34°26.7’S 38 205 + 115° 27.TE 1991 210291 Cape Pasley/ 33°49'S 57 20.8 7 fish Pt. Malcolm 124°0'E 050391 Cape Pasley/ 34°02'S 55 21.1 5 fish Pt. Malcolm 123°40'E + WA Salmon were caught but precise quantities were not recorded Australian herring fishery The stock of Australian herring (Arripis georgianus ) occupies an almost identical range to salmon, extending from Kalbarri to South Australia and into Victoria (Fig. 1). Like salmon, herring spawn predominantly on the lower west and western south coasts of Western Australia (Lenanton 1978), while the juveniles extend through the Great Australian Bight into South Australia and Victoria. A pre-spawning migration to Western Australia occurs for the first time during the second year of life. The source of recruitment ranges from South Australia and the G.A.Bight region to local marine embayments. particularly Geographe Bay, where juveniles occur abundantly, associated with seagrass and drift macrophytes in the waters inshore of the Leeuwin Current (Lenanton 1982). Ongoing research (G.K. Jones, unpublished) is providing preliminary evidence for a direct link between the abundance of juveniles in South Australia and the size of the spawning stock in Western Australia. Thus it is highly likely that the strong Leeuwin Current flow at the time of spawning in autumn is a critical factor in the transport of larvae across the G.A.Bight to South Australian and Victorian nursery areas. 108 Journal of the Royal Society of Western Australia, 74,1991 250 200 o x £ 9 1501 o S Vi 100 - 50 High Catch Levels r= (-0.44) Not Significant LEGEND ° High Catch 0 Low Catch Low Catch Levels r=(-0.54) 0.01

0.05** 60 70 80 ~~r 90 Mean Sea Level Height at Albany Feb.-Apr. (cm) Figure 7 Relationship between the mean monthly Albany sealevel between February and April, and the mean annual south coast log book catch of salmon per hour of beach observation during years of low and high annual catch. Pilchard purse seine fishery The stock of pilchards (Sardinops neopilchardus ) on the south coast, centred on King George Sound at Albany, provides Western Australia's largest single catch of finfish (Fletcher 1990). The fishery, which has developed over the past decade, reached a peak in catch of 8 300 tonnes in 1988, but has subsequently declined to less than 6 500 t annually (W.J. Fletcher, unpubl). This species is either closely related to, or the same species as, the dominant Sardinops species in the Humboldt and Benguela Currents (Table 1, Parrish et al. 1989). However, catches in Western Australia to date indicate that maximum sustainable production is likely to be around 10 000 t which is orders of magnitude lower than catches of the same or similar planktivorous species in the other current systems (FAO 1988). Survey data (W.J. Fletcher & R.J. Tregonning, unpublished) has shown that the species in Western Australia is found predominantly close inshore and therefore is taken only in bays, usually within the 50 m depth contour. Thus the Lceuwin Current significantly reduces the production of pilchards in this region compared to other boundary current systems. Furthermore, analyses of commercial catch rates, together with computer modeling (W.J. Fletcher, unpubl), suggest that yearly fluctuations in current strength are probably involved in the larger variations in the observed catch. Western rock lobster fishery The western rock lobster (Panulirus cygnus ) stock supporting Australia’s most valuable single species fishery is directly influenced by the Leeuwin Current and other environmental factors. Phillips et al. (1991) and Pearce & Phillips (1992) deal in detail with the impact of the current on the recruitment of the puerulus stage of the life history. These studies have shown that the levels of puerulus settlement in the nursery grounds on the coast are highly correlated with sealevel changes which provide an index of the Leeuwin Current strength (Pearce & Phillips 1988), and with westerly storm conditions during the settlement period (Caputi & Brown 1989). Once settled, the juvenile lobsters remain for approximately 4 to 5 years on the coastal limestone reefs while feeding on the fauna and flora associated with seagrass beds (Joll & Phillips 1984). Thus the Leeuwin Current, which not only regulates recruitment to the stock but maintains the clear water environment essential to the development and survival of the extensive seagrass beds, is closely linked to the overall production of the fishery. 109 Journal of the Royal Society of Western Australia, 74, 1991 A second and possibly more critical influence of the Leeuwin Current on this important fishery is through its impact on the biology of the lobster at the Abrolhos Islands. In this location, the lobster stock matures at a smaller size than on coastal reefs, and spawns before reaching the minimum legal size for capture (C.F. Chubb, unpublished). The lobsters at the Abrolhos Islands account for about half the annual egg production (C.F. Chubb, unpublished) from the total stock, and are critical for the ongoing productivity of the fishery. While the specific effect of the Leeuwin Current on the spawning stock has yet to be precisely determined, evidence from aquarium experiments has shown that elevated water temperatures, such as those caused by the current at the Abrolhos Islands, increase reproductive activity (Chittleborough 1976). A further important aspect of the fishery involving the Leeuwin Current is the effect on catchability of the lobsters through the influence on temperature and salinity. Furthermore, Morgan (1974) has shown that both temperature and salinity variations at the Abrolhos Islands, (again related to the current), have significant effects on the catchability of the rock lobster. Thus, the Leeuwin Current has a major influence on most stages of the life history of the lobster and the catch ultimately achieved by the fishery (Phillips 1986). Saucer scallop fishery The distribution of the saucer scallop (Amusium balloti) extends considerably further south on the western coast of Australia than on the eastern coast. On the western coast it extends as far south as 35°S (off Albany) and east along the southern coast to Esperance (122°E) (Gwyther et al. 1991), whereas on the eastern coast it extends only as far as 27°S (Moreton Bay) (Dredge 1985). This extension of the range on the western side of the continent almost certainly results from the warming influence of the Leeuwin Current. Scallop populations throughout the world are acknowledged as having highly variable recruitment as a result of the influence of environmental factors (Caddy 1989). In the Shark Bay scallop fishery catches of A. balloti have shown a greater than five-fold variation over the period 1983 - 1990, primarily as a result of inter-annual variations in recruitment (Joll & Caputi 1991). Examination of satellite imagery of Shark Bay suggested that the Leeuwin Current may be the environmental factor responsible for this recruitment variation. The imagery showed tongues of warmer water, derived from the south-flowing Leeuwin Current, entering the bay during the spawning season (April to December (Joll 1987)) and possibly affecting recruitment. Populations at locations further south (eg the Abrolhos Islands) spawn at different times of the year (Joll 1989) and are probably less vulnerable to any environmental influences of the Leeuwin Current. Surveys to measure recruitment in Shark Bay have been conducted in November each year since 1983. Growth data from tagged scallops (Joll 1987) showed that scallops from size classes as small as 30-39 mm in November reach a size of around 90 mm by March of the following year, at which size they are acceptable for commercial harvest. Trawling surveys in November, therefore, are capable of estimating the abundance of recruits from the current spawning season which will reach sizes appropriate to enter the fishery in the following year. Data on landings of scallops by vessels operating in the scallop fishery are provided voluntarily by fishermen in their research logbooks and these are checked against wholesale buyers’ receival records. In all years except 1983 the fishery has ceased before the legal closing date when catches have fallen to levels which are not commercially viable. With the exception of 1984, therefore, the catch in each year has been dominated by the new recruits from the previous year's spawning. The strength of the Leeuwin Current is reflected in the coastal sealevel (Reid & Mantyla 1976, Pearce & Phillips 1988), so that data from the Fremantle tide gauge are a useful index of the flow of the Leeuwin Current. The sealevel at Carnarvon may have more accurately reflected the influence of the Leeuwin Current in the Shark Bay area, but these data were not available for the whole of the period of this study. Figure 8 Relationship between the recruitment index for saucer scallops (Amusium balloti ) in Shark Bay over the period 1983 - 1990 and the mean sealevel at Fremantle over the period May to August of the same year. (83 = year of data; • ENSO years; O Non-ENSO years) (spnm = scallops per nautical mile). Spawning activity, which results in recruits detectable in the November survey and which subsequently contribute to the recruitment to the fishery in the following year, occurs mainly in the period, April to July (Joll & Caputi 1991). Therefore, in considering the effects of the Leeuwin Current on ( scallop recruitment, it is the strength of the current in 110 , Journal of the Royal Society of Western Australia, 74, 1991 these months which is likely to be of greatest importance. As the peak in the sealevel at Fremantle due to the current occurs about a month after the peak at Carnarvon, the sealevel over the months of May to August at Fremantle was used as the environmental variable to examine the influence of the Leeuwin Current on spawning / recruitment success in Shark Bay over the period April to July. Over the period 1983-1990, average Fremantle sealevel for the months May to August was negatively correlated with the abundance of recruits measured in the November survey of that year (Fig.8). Similarly, there was a negative correlation between the Fremantle sealevel over the months May to August and the catch of the fishery in the following year (Fig. 9). The catch for 1991 indicated in Fig. 9 is a conservatively estimated figure based on the very high recruitment index recorded in November 1990. By the end of June 1990 the catch of the fishery was over 1 000 tonnes. 1000 V 90/91 (ANTICIPATED) 800 X o £ o 600 400 200 87/88 82/83 # 86/87 O 89/90 O 83/84 O 85/86 O 84/85 O 88/89 0 L -*- 1 -‘- 1 70 80 90 SEA LEVEL (May — Aug) cm Figure 9 Relationship between the annual catch of saucer scallops (Amusium balloti) from Shark Bay over the period 1983 to 1991 and the mean sealevel at Fremantle over the period May to August of the previous year. (83/84 = year of sealevel data / year of catch data; • ENSO years; O Non-ENSO years). (1991 catch data are anticipated). The mechanism by which the Leeuwin Current influences recruitment success in Amusium balloti in Shark Bay has not been determined. However, the data suggest very strongly that in years of a weak Leeuwin Current, both recruitment success and the catch in the following year will be high. The Leeuwin Current is known to be weakest when El Niflo/Southern Oscillation (ENSO) events occur (Pearce & Phillips 1988), so that good recruitment could also be expected to be associated with these events. Massive increases in the abundance of the Chilean scallop (Argopecten purpuratus) were noted to be associated with the ENSO event of 1982/83 by Arntz (1984) and Wolff (1987). While the mechanism of action of the environment on recruitment success of Amusium balloti has not been positively identified, hydrological flushing in years of strong Leeuwin Currents seems a strong possibility. Both Dickie (1955) and Caddy (1979,1989) noted the importance of hydrological flushing in recruitment success of the Atlantic sea scallop ( Placopecten magellanicus). Strong Leeuwin Currents also bring warm, low-nutrient waters into Shark Bay. Thus, other possibilities for the mechanism of action may be negative effects of increased temperatures on spawning or the success of fertilization or a reduction in primary production leading to an inadequate food supply for the larvae. Whichever mechanism or combination of mechanisms is responsible for the observed influence of the Leeuwin Current, it is clear that the effect of the current is to depress fisheries production in an embayment which is otherwise capable of high productivity. Penaeid prawn fishery of Shark Bay The largest Western Australian fishery for western king (Penaeus latisulcatus ) and brown tiger ( Penaeus esculentus) prawns is located in Shark Bay (Penn et al. 1989), a sector of the coast frequently influenced by the Leeuwin Current. The current has two major effects on the prawn fishery particularly the major western king prawn stock. It firstly radically changes the annual temperature cycle on the trawl grounds (Penn 1988) from that found in the more usual annual cycle in inshore waters which are unaffected by the current. The winter Leeuwin Current flow results in the temperature of the shelf water peaking later, usually in May; and a period of lower temperatures extending through spring, when the warm current declines and cooler local inshore waters dominate. This unusual temperature regime alters the burrowing behaviour of western king prawns (Penn 1984) and thereby influences the catches of prawns. The resulting changes in catchability produce high catches and maximum exploitation rates from the start of each season in March through to May/June of each year, followed by significantly lower catches and exploitation rates for the remainder of the year. Alterations in the annual temperature cycle, particularly the timing of the temperature decline in May/June that is almost certainly related to the timing of the current peak, have been simulated in a computer model (N.G. Hall & J.S. Andrews, pers. comm.) which predicts catch variations in the order of 20% with a one month alteration in the time of the temperature decline. Secondly, within each season, there is also a significant correlation (r=0.6) between recruitment catches each year, and the strength of the Leeuwin Current expressed as the mean monthly Fremantle sealevel over the period April to August of that same year. Since the spawning season for king prawns recruited in a particular year is during winter/spring of the proceeding year, the above correlation suggests that the current is having an effect on the survival, or Journal of the Royal Society of Western Australia, 74,1991 growth, of recruits, once the year class has migrated out into the main trawl ground in the northern region of Shark Bay (Penn et al. 1989). As a result of the generally positive relationship between the current strength and prawn catches, and the negative effect of the current on the scallop recruitment to the same trawl fishery, the cycles in prawn and scallop catches are often out of phase. Furthermore, the relatively infrequent occurrence of weak Leeuwin Current years results in consistent, relatively large king prawn catches and on average low scallop catches with occasional very high catches resulting from strong recruitment. Summary In conclusion, under the influence of the Leeuwin Current, a more tropical coastal water environment has evolved off south-western Australia. This situation contrasts markedly with those of other environments characteristic of eastern boundary currents. The oceanic sources of nutrients which support extensive plankton-based food chains on other western continental shorelines where upwelling occurs are not available off Western Australian. Fisheries production in these waters is therefore heavily dependent on benthic-based food webs in near-shore waters, rather than on those associated with oceanic ecosystems. Thus inshore demersal invertebrate fisheries such as rock lobster, rather than pelagic finfish resources, dominate fisheries production in Western Australia. Both the strength and timing of the peak current flow also appear to influence significantly the annual catches of most of the economically important finfish and invertebrate resources of the west and south coasts of Western Australia. Depending on the species being considered, strong current flows can either adversely or favourably affect catches. The precise mechanisms however, are in many instances still poorly understood, although larval dispersal and catchability variations are thought to be the most likely factors. Long-term studies into the important interaction between the Leeuwin Current and Western Australia's major fisheries are ongoing with a view to increasing the level of understanding of the mechanisms underlying the effects of the current. Acknowledgements We thank Chris Dibden and Mark Cliff for much of the data extraction and analysis, and for preparation of most of the figures and text for publication. We also thank Nic Caputi for his assistance with statistical analyses and discussion of environmental relationships. Our colleagues from the WA Marine Research Laboratories and Murdoch University constructively criticized the manuscript. Commercial fishers Bob Bubb and Peter Osborne kindly provided information from their personal logbooks. Tidal data are supplied by the National Tidal Facility, The Flinders University of South Australia, copyright reserved. References Anonymous 1990 State of the fisheries. Government Printer Perth Western Australia, 21pp. Armstrong D A, Mitchcll-Innes B A, Verheye-dua F, Waldrop H & Hutchings L 1987 Physical and biological features across an upwelling front in the southern Benguela. In: The Benguela and Comparably Ecosystems (eds Payne A I L, Gulland J A and Brink K H). SAfrJmarSd 5:171-190. Arntz W E 1984 El Nino & Peru: Positive aspects. Oceanu^ 27:36-39. Bohle-Carbon ell M 1989 On the variability of the Peruvian upwelling system. In: The Peruvian Upwelling Ecosystem: Dynamics & Interactions Proceedings of th^ Workshop on Models for Yield Prediction in thy Peruvian Ecosystem 24-28 August 1987 Callao, Peru (ed^ Pauly P, Muck P, Mendo J & Tsukayama I) ICLARM Conference Proceedings 18:14-32. Boyd A J 1979 A relationship between sea surface temperature variability and anchovy Engraulis capensi$ recruitment off South West Africa 1970-1978. Fish Bull S Afr 12:80-84. Caddy J F 1979 Long-term trends and evidence for production cycles in the Bay of Fundy scallop fishery. Rapp P-v Reun Cons Perm Int Explor Mer 175:97-108. Caddy J F 1989 Recent developments in research and management for wild stocks of bivalves and gastropoda. In: Marine Invertebrate Fisheries: Their Assessment and Management, (ed J F Caddy) John Wiley & Sons NY, 665 - 700. Cappo M C 1987 The fate and fisheries biology of sub-adult Australian salmon in South Australian waters. Fishing Industry Research Trust Account Report 84/75 South Australian Department of Fisheries Unpublished report July, 1987:162 pp. Caputi N & Brown R 1989 The effect of environmental factor^ and spawning stock on the puerulus settlement of thy western rock lobster. In: Workshop on rock lobster ecology and management (ed B F Phillips) CSIRO Marine Laboratories Rep 207: 29 pp. Chavez F P, Barker R T & Sanderson M P 1989 The potential primary production of the Peruvian upwelling i ecosystem, 1953-1984. In: The Peruvian Upwelling Ecosystem: Dynamics and Interactions. Proceedings of the Workshop on Models for Yield Prediction in thy Peruvian Ecosystem 24-28 August 1987, Callao, Peru (eds Pauly P, Muck P, Mendo J & Tsukayama I) 1CLARM Conference Proceedings 18:50-63. Chittleborough R G 1976 Breeding of Panulirus longipes cygnus (George) under natural and controlled conditions. Aust J Mar Freshw Res 27:499-516. Crawford R J M 1987 Food and population variability in five regions supporting large stocks of anchovy, sardine and horse-mackerel. In: The Benguela and Comparable Ecosystems (eds Payne A I L, Gulland J A & Brink K H) S Afr J mar Sci 5:735-757. Crawford R J M, Shelton PA & Hutchings L 1983 Aspects of variability of some ncritic stocks in the southern , Benguela system In: Proceedings of the Expert Consultation to Examine Changes in Abundance and species Composition of Ncritic Fish Resources San Jose, Costa Rica, April 1983, (eds Sharp G D and Csirke J) FAO Fish Rep 291(2):407-448. Crawford R J M, Shannon L V and Pollock D E 1987 The Benguela ecosystem Part IV. The major fish and invertebrate resources. Oceanogr Mar Biol Ann Rev 25:353-505. Cresswell G R 1986 The role of the Leeuwin Current in the life cycle of several marine creatures. UNESCO Technical Papers in Marine Science 49:60-64. Journal of the Royal Society of Western Australia / 74, 1991 Cresswell G R and Golding T J 1980 Observations of a south flowing current in the south-eastern Indian Ocean. Deep-Sea Res 27A: 449-466. CsirkeJ 1989 Changes in the catchability coefficient in the Peruvian Anchoveta ( Engraulis ringens ) fishery. In: The Peruvian Upwelling Ecosystem: Dynamics and Interactions. Proceedings of the Workshop on Models for Yield Prediction in the Peruvian Ecosystem , 24-28 August 1987, Callao, Peru (eds Pauly P, Muck P, Mendo J & Tsukayama I) ICLARM Conference Proceedings 18:207-219. Cushing D H 1971 Upwelling and the production of fish. Adv Mar Biol 9:255-334. Cushing D H 1982 Climate and fisheries. Academic Press, London, 373 pp. Dickie L M 1955 Fluctuations in abundance of the giant scallop, Placopecten magellanicus (Gmclin), in the Digby area of the Bay of Fundy. J Fish Res Bd Can 12:797-857. Dredge M C L 1985 Estimates of natural mortality and yield- per-recruit for Amusium japonicum balloti Bcrnardi (Pectinidae) based on tag recoveries. J Shellfish Res 5:103-109. FAO 1988 Yearbook of fishery statistics: Catches and landings, 1986. FAO Fisheries Series 30 62: Table B- 35,119. Fletcher W J 1990 A synopsis of the biology and the exploitation of the Australasian Pilchard, Sard inops neopilchardus (Steindachner) Part 1: Biology. Fish Res Rep Fish Dept West Aust 88:45 pp. Godfrey J S and Ridgway K R 1985 The large-scale environment of the poleward flowing Leeuwin Current, Western Australia: longshore steric height gradients, wind stresses and geostrophic flow. J Phys Oceanogr 15:481 - 495. Golding T J & Symonds G 1978 Some surface circulation features off Western Australia during 1973-76. Aust J Mar Fresh w Res 29:187-191. Gwyther D , Cropp D A , Joll L M & Dredge, M C L 1991 Fisheries and Aquaculture: Australia. In Scallops: Biology, Ecology and Aquaculture. Developments in Aquaculture and Fisheries Science 21 (ed S Shumway) Elsevier (Amsterdam) 835-850. Hatcher 1991 Coral reefs in the Leeuwin Current - an ecological perspective. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:115-127. Hutchins 1991 Dispersal of tropical fishes to temperate seas in the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:79-84. Hutchins B and Swainston R 1986 Sea fishes of southern Australia. Swainston Publishing, Perth, Western Australia, 180 pp. Joll L M 1987 The Shark Bay scallop fishery. Fisheries Management Paper No 11, Fish Dept West Aust 123pp. Joll L M 1989 Recruitment variation in stocks of the saucer scallop Amusium balloti in the Abrolhos Island area. In: Proceedings of the Australasian Scallop Workshop (eds Dredge M L C, Zacharin W F & Joll L M) Tasmanian Government Printer. 61-67. Joll L M & Caputi N 1991 Environmental influences on recruitment in the Shark Bay saucer scallop ( Amusium balloti ) fishery. Proceedings of "Shellfish life histories and shell fishery models. ICES Symposium, Monctou, Canada, 1990. Rapp P -v Reun Cons Int Explor Mer Joll L M & Phillips B F 1984 Natural diet and growth of juvenile western rock lobsters Panulirus cygnus George. J Exp Mar Biol Ecol 75:145-169. King D P F 1977 Influence of temperature, dissolved oxygen and salinity on incubation and early larval development of the South West African pilchard ( Sardinops ocellata). Invest Rep Sea Fish Brch S Afr 114:1-35. Kirkman H 1985 Community structure in seagrass in southern Western Australia. Aquat Bot 21:363-375. Lenanton R C J 1977 Aspects of the ecology of fish and commercial crustaceans of the Blackwood River estuary. Western Australia. Fish Res Bull West Aust 19:1-72. Lenanton R C J 1978 Age, spawning time and fecundity of Australian herring ( Arripis georgianus Cuvier and Valenciennes) (Pisces: Arripidae) from the waters around Rottnest Island, Western Australia. Aust J Mar Freshw Res 29:599-612. Lenanton R C J 1982 Alternative non-estuarine nursery habitats for some commercially and recreationally important fish species of south-western Australia. Aust J Mar Freshw Res 33: 881-900. Lenanton R C J & Hodgkin E P 1985 Life history strategies of fish in some temperate Australian estuaries. In: Fish community ecology in estuaries and coastal lagoons: Towards an ecosystem integration (ed AYanez- Arandbia) UNAM Press, Mexico, Chapt 13:267-284. Lenanton R C J & Potter I C 1987 Contribution of estuaries to commerdal fisheries in temperate Western Australia and the concept of estuarine dependence. Estuaries 10:28-35. Malcolm W B 1960 Area of distribution, and movement of the western sub-species of the Australian salmon, Arripis trulla esper Whitley. Aust J Mar Freshw Res 11:282- 325. Maxwell J G II & Cresswell G R 1981 Dispersal of tropical marine fauna to the Great Australian Bight by the Leeuwin Current. Aust J Mar Freshw Res 32:493-500. Morgan G R 1974 Aspects of the population dynamics of the western rock lobster, Panulirus cygnus George. II. Seasonal changes in catchability. Aust J Mar Freshw Res 25:248-259. Muck P 1989a Major trends in the pelagic ecosystem off Peru, and their implications for management. In: The Peruvian Upwelling Ecosystem: Dynamics and Interactions. Proceedings of the Workshop on Models for Yield Prediction in the Peruvian Ecosystem , 24-28 August 1987, Callao, Peru (eds Pauly P, Muck P, Mendo J & Tsukayama I) ICLARM Conference Proceedings 18:386-403. Muck P 1989b Anchoveta consumption of Peruvian hake: A distribution and feeding model. In: The Peruvian Upwelling Ecosystem: Dynamics and Interactions Proceedings of the Workshop on Models for Yield Prediction in the Peruvian Ecosystem , 24-28 August 1987, Callao, Peru (eds Pauly P, Muck P, Mendo J & Tsukayama I) ICLARM Conference Proceedings 18:306- 320. Nicholls A G 1973 Growth in the Australian "salmon" Arripis trutta (Bloch and Schneider). Aust J Mar Freshw Res 24:159-176. 113 Journal of the Royal Society of Western Australia, 74, 1991 Parrish R H, Bakun A, Husby D M & Nelson C S 1983 Comparative climatology of selected environmental processes in relation to eastern boundary current pelagic fish reproduction. In: Proceedings of the Expert Consultation to Examine Changes in Abundance and Species Composition of Neritic Fish Resources, San Jose, Costa Rica, April 1983 (eds Sharp G D & Csirke J) FAO Fish Rep 291 (3) :731-777. Parrish R H, Serra R & Grant W S 1989 The monotypic sardines, Sardina and Sardinops : their taxonomy, distribution, stock structure and zoogeography. Can J Fish Aquat Sd 46:2019-2036. Pearce A F 1985 The Leeuwin Current, as viewed from space. HNS 18:3-5. Pearce A F 1991 Eastern boundary currents of the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:35-45. Pearce A F & Phillips B F 1988 ENSO events, the Leeuwin Current, and larval recruitment of the western rock lobster. J Cons int Explor Mer 45:13-21. Pearce A F & Phillips B F 1992 Oceanic processes, puerulus settlement and recruitment of the Western rock lobster Panulirus cygnus. Boden Research Conference, Thredbo, NSW, February 1990. Pearce AF & Prata A J 1989 Mysteries of the Leeuwin Current. Western Angler October-November 198935-36. Pearce A F, Johannes R E, Manning C R, Rimmer D W & Smith D F 1985 Hydrology and nutrient data off Marmion, Perth, 1979-1982. CS1RO Marine Laboratory Report 167.45 pp. Penn J W 1981 A review of mark recapture and recruitment studies on Australian penaeid shrimp. Kuwait Bull Mar Sd 2:227-248. Penn J W 1984 The behaviour and catchability of some commercially exploited penaeids and their relationship to stock and recruitment. In: Penaeid shrimps, their biology and management (eds Gulland J A & Rothschild B J) Fishing News Books, Surrey, England. 173-186 Penn J W 1988 Spawning stock-recruitment relationships and management of the penaeid prawn fishery in Shark Bay, Western Australia Ph D Thesis, Murdoch University. 239pp Penn J W, Hall N G & Caputi N 1989 Resource assessment and management perspectives of the Penaeid prawn fisheries of Western Australia. In: The scientific basis of Shellfish Management (ed J Caddy) John Wiley & Sons, New York. 115-140. Phillips B F 1986 Prediction of commercial catches of the Western rock lobster Panulirus cygnus. Can J Fish Aquat Sd 43:2126-2130. Phillips B F & Brown R S 1989 The West Australian rock lobster fishery: Research for management. In: The Sdcntific basis of shellfish management (ed J Caddy) John Wiley & Sons, New York. 159-181. Phillips B F, Pearce A F & Litchfield 1991 The Leeuwin Current and larval recruitment to the rock (spiny) lobster fishery off Western Australia. In: The Leeuwin Current: an influence on the coastal climate and marino life of Western Australia, (eds. AF Pearce and Dl Walker). J Roy Soc WA 74:93-101. Reid J L & Mantyla A W 1976 The effect of the geostrophi c flow upon coastal sea elevations in the northern North Padfic Ocean. J Geophys Res 81:3100-3110. Robertson A I 1982 Population dynamics and feeding ecology of juvenile Australian salmon (Ar ripis trutta) in Western Port, Victoria. Aust J mar Freshwat Res 33:369-375. Robertson A I & Lenanton R C J 1984 Fish community structure and food chain dynamics in the surf-zone of sandy beaches: The role of detatched macrophyte detritus, j Exp Mar Biol Ecol 84:265-283. Rochford D J 1980 Nutrient status of the oceans around Australia. CSIRO Fish and Ocean Ann Report 1977- 79:9-12. Rochford D J 1988 Seasonal influx of nitrates to the slope and shelf waters off north-west Australia. CSIRO Mar Lab Report 191.23pp. Shannon L V 1985 The Benguela ecosystem. Part 1. Evolution of the Benguela physical features and processes. Oceanogr Mar Biol Ann Rev 23:105-182. Sharp G D 1987 Climate and fisheries: cause and effect or managing the long and the short of it all. In: The Benguela and Comparable Ecosystems (eds Payne A I L, Gulland J A & Brink K H) S Afr J mar Sd 5:811-838. Sherman K 1987 Large marine ecosystems as global units for recruitment experiments. ICES Council meeting, 1987 Copenhagen, Denmark 17 pp. Stanley C A 1979 Decline in Western Australian salmon catch Aust Fish 38(3):14-17. Stanley C A 1980a Australian salmon. CSIRO Div Fish Oceanogr Fish Sit Rep 5 llpp Stanley C A 1980b Lee’s phenomenon in the western sub- spedes of the Australian salmon, Arripis trutta esper. Aust J mar Fresh wat Res 31:13-19. Walker D I 1991 The effect of sea temperature on seagrasses and algae on the Western Australian Coastline. J In: The Leeuwin Current: an influence on the coastal dimate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:71-77. Walker M H 1982 The present state of the Western Australian fishery for Australian salmon. W A Dept of Fish and Wildlife Report 5232 pp. Weaver A J 1990 Ocean currents and climate. Nature 347(4):432. Wolff M 1987 Population dynamics of the Peruvian scallop Argopecten purpuratus during the El Nino phenomenon of 1983. Can J Fish Aquat Sci 44:1684-1691. Wooster W S & Bailey K M 1989 Recruitment of marine fish revisited. In: Effects of ocean variability on recruitment and an evaluation of parameters used in stock assessment models (eds Beamish R J & McFarlene G A) Can Spec Publ Fish, Aquat Sd 108:153-159. Wooster W S & Reid J L 1963 Eastern boundary currents. In: The sea Vol 2 (ed N M Hill) Intersdence, New York 253-276. 114 Journal of the Royal Society of Western Australia / 74, 1991, 115-127 Coral reefs in the Leeuwin Current - an ecological perspective B G Hatcher Oceanography Department, Dalhousie University, Halifax, Nova Scotia,B3H 4J1, Canada. Abstract The Leeuwin Current differs markedly from other eastern boundary currents in that it transports tropical water towards a polar ocean, and inhibits upwclling. As a result, the offshore marine environment of Western Australia is characterised by elevated sea temperatures and reduced dissolved nutrient and particulate concentrations relative to adjacent coastal and south-west Indian Ocean water masses. As in many western boundary currents, the distribution of coral reefs in and near the Leeuwin Current is extensive, and appears to mirror the influence of the current, with the limit of active reef accretion offshore extending well outside the tropics. Causality in the relationship between the regional oceanography and coral reef development has not been established, but the assumed dominance of sea temperature as a factor controlling reef growth is questioned on the basis of the evidence available. The Leeuwin Current's role in maintaining apparently low rates of nutrient delivery to the benthos, in combination with its elevation of sea temperature and advection of planktonic spores and larvae, serves to inhibit the development of marine macrophyte communities, which compete effectively with coral reef-building communities. At reefs where these two benthic community types overlap (ie the Houtman Abrolhos), periodic delivery of high nutrient water is inferred by the extent of non reef-building macrophyte communities. 1 suggest that the primary influence of the Leeuwin Current on coral reef development is to modulate the competition between coral and macrophyte communities. More oceanographic measurements, geological analyses and ecological experiments are required to test the hypothesis. Ridgway (1984), and by Pearce (1991) in this volume. It is Introduction a narrow (<200 km), shallow (<200 m) stream of The role of the Leeuwin Current in conti oiling the development and distribution of coral reefs along the Western Australian coast has been a subject of interest at least since Saville-Kent's (1897) observations of the Abrolhos reefs. There has also been a great deal of speculation on the topic, and precious little hard data collected. In this brief review I will make limited comparisons between the Leeuwin and other current systems, and discuss the most obvious mechanisms by which ocean currents may control reef growth. Finally, I will focus on the Abrolhos as the epitome of Leeuwin Current reefs. relatively warm, low salinity, low nutrient oceanic water of tropical origin which flows southwards at relatively high velocity (0.1 - 0.4 ms-i) along the western continental slope of Australia. The Current is driven by a latitudinal gradient of steric scalevel, is seasonal in its volume flux and sheds large and mesoscale eddies into the Indian Ocean and onto the shelf. The stream is coherent at least from North West Cape to Cape Leeuwin, with clear influence on flows on the Northwest Shelf and in the Great Australian Bight. I refer to this extent of continental shelf and slope as the "Leeuwin Province". Clear definition of terms is required. I have followed Fagerstrom (1987) in defining coral reefs as: "carbonate structures with a framework dominated by the skeletal remains of zooxanthellate, scleractinian corals, supporting a living veneer of those and other calcifying organisms". This definition encompasses the majority of photic zone, Holocene reefs, but specifically excludes aphotic zone reefs dominated by hermatypic corals, non-coral dominated reefs (eg algal-millepora reefs), and communities of corals living on non-coral reef structures such as limestone or sandstone. The Leeuwin Current is well defined by Crcsswell & Golding (1980), Thompson & Veronis (1982), Godfrey & Coral Reef Initiation and Accretion Extant coral reefs in the Leeuwin Province began growth within the past ten thousand years as coral communities recruiting to newly available marine substrata during the Holocene transgression. The basement could be either a former coral reef, killed during a glaciation perhaps, or any other hard surface in shallow water. Under suitable conditions reefs grew upwards at rates which approximated the rate of scalevel rise. Those which reached the sea surface then expanded horizontally. 115 Journal of the Royal Society of Western Australia / 74, 1991 Figure 1 Chart of the earth between the 40° parallels, showing the major boundary currents, the 18°C winter minimum monthly mean isotherm, and the distribution of coral reefs (modified from Neumann & Pierson 1966, Veron 1986). Boundary Current codes: A.C.= Agulhas Current, B.C.= Brazil Current (off S.America) & Benguela Current (off Africa), C.C.= California Current (off N.America) & Canary Current (off Africa), E.A.C.= East Australia Current, G.C.= Guiana Current (off S.America) & Guinea Current (off Africa), G.S.= Gulf Stream, K.S.C.= Kuro Shio Current, L.C.= Leeuwin Current, M.C.= Mozambique Current, P.C.= Peru Current (Humbolt Current). Latitudinal limits of coral reefs in the open ocean are indicated by doubled-ended brackets, those near coasts by single-ended brackets. The essential dimension of reef growth is time. With upward accretion rates ranging from 1 to 10 mm per year, even modest reef structures represent temporal integration of environmental controls over thousands of years. Perhaps the single strongest message from oceanographic research during the last decade has been the variability of even the largest ocean structures. Rapid, glacially-forced change in sealevel is the penultimate control on patterns and rates of coral reef development (Davies & Montaggioni 1985). The ultimate control is the foundation upon which photic zone reefs must rest. Coral reefs cannot develop even under ideal atmospheric and seawater conditions if they have no structure within the upper 100 m of the ocean upon which to initiate growth. The lack of shallow foundation is not a constraint on the shallower portions of the continental shelves off Western Australia, but is absolute on the shelf slope where the Leeuwin Current flows. The dearth of tectonism and related seamounts and guyots means that there are few potential reef sites actually in the Leeuwin Current: most are on the adjacent shelf or coast. Even when the foundation of a Leeuwin reef is of coral construction, historical interruptions in vertical accretion of sufficient duration may have precluded subsequent reef development. The world's oceans are littered with the ghosts of drowned reefs, overlain by a photic zone ideal for coral growth (Menard 1986). Who knows what ghosts lie beneath the Leeuwin Current? The process of reef accretion is the net result of a coral community’s initiation and subsequent development in a particular environment. It requires: 1) A consistent supply of the larvae of reef-building, binding and infilling plants and invertebrates, most of which have a planktonic life stage. The supply can be from both local populations (ie self-seeding) and from distant reefs. Such immigration is a requisite for initial reef colonization. 2) Suitable substrata for larval settlement when they are delivered to a region: there must be available space on the seabed within the photic zone. 3) Suitable environmental conditions for the survival and growth of newly settled recruits through to the adult life stages, including: a) Non-lethal sea temperatures and salinities. b) Adequate irradiance for positive net daily photosynthesis by symbiotic and free-living algae. c) Adequate supplies of inorganic nutrients to meet the demands of autotrophic organisms for growth and calcification. d) Low concentrations of inorganic and organic particulates in the water column which otherwise inhibit coral growth by shading, smothering, and promoting bacterial infection. Clearly many of the requirements for coral reef accretion such as temperature, light and nutrient concentration are the very factors which are strongly influenced by regional oceanography such as boundary currents. Poleward-flowing boundary currents, including the Leeuwin Current, have the potential to influence all of these factors, either directly or indirectly. The transport of tropical reef larvae and moderation of winter sea surface temperatures have received most attention. What can we learn about the Leeuwin’s influence on coral reef development by looking at the global distribution of boundary currents and reefs? 116 Journal of the Royal Society of Western Australia, 74,1991 Boundary Currents and Reef Systems The global distribution of coral reefs as defined here is quite predictable (Fig.l). Reefs occur exclusively within the photic zone, and usually within 18°C mean monthly minimum isotherm. In the open oceans the latitudinal limits of reef development fall close to the tropics, generally within the 25th parallels (eg Midway in the north Pacific). Within semi-enclosed seas (eg Red Sea, Persian Gulf) and ocean margins however, the isotherms are distorted by continental influences and deflection of currents which close the great ocean circulations. Western boundary currents are the strongest, carrying warm water poleward in relatively narrow and deep streams such as the Mozambique - Agulhas Current system of eastern Africa. This current extends the poleward limit of reef development to well south of Durban (Fig.l). On the eastern margins of the ocean basins the westerly flows of the major gyres are deflected towards the equator in slower boundary currents which are usually broad and shallow (Fig.l). Because of Coriolis' force and the trade winds, eastern boundary currents are generally characterized by upwelling at their coastal edges. Thus, the Benguela and Peru Currents are the analogs of the West Australian Current, and should serve as comparative models for the distribution of coral reefs in eastern boundary currents. Such a comparison will perforce be brief. Only three reef systems (including the Galapagos) fall within the precinct of the Peru Current, the southernmost of which is at only 7°S near the coast at Chiclayo, Peru (Fig.l). All of them, including those at the Galapagos Islands over 1000 km from the South American coast, are marginal reefs in terms of their extent, diversity and vertical accretion rates. Coral reefs in the Gulf of Guinea at the northern extreme of the Benguela Current extend only 2°S of the equator. It is apparent that the cold, nutrient-rich waters of eastern boundary currents somehow inhibit the development of coral reefs. Presumably, that is one reason why there are no open ocean atolls on the few seamounts in the West Australian Current. In contrast, the Elizabeth and Middleton Reefs occur at 29 to 30°S near Lord Howe in the East Australian Current. The Leeuwin Current is clearly not a typical eastern boundary current. Its high poleward velocity and narrowness liken it to a western boundary current. But it is differently forced and shallower, so it moves a far smaller volume of water. In some ways it resembles the narrow currents which interpose between the major boundary currents and the coasts off North Africa and the America's, such as the Guinea Current and the Labrador Current extension. Both of these are equatorward currents, however, and are not strictly comparable with the Leeuwin. As an oceanographic feature, the Leeuwin Current appears to be in a class by itself. The distribution of reefs along the Current, however, is repeated in poleward flowing boundary currents throughout the world. In terms of the factors controlling the development of coral reefs, more appropriate comparisons might be made with western boundary currents such as the Agulhas Current, the East Australian Current, the Kuro Shio Current and the Antilles/Florida/Gulf Stream current system. Like the Leeuwin Current, all of these sustain coral reefs in decreasing density and diversity along gradients through the tropics to well beyond them at Durban, Lord Howe Island, Kyushu Island and Bermuda respectively. They originate in regions rich in coral reefs and thus can transport reproductive propagules in a conducive biophysical regime. Hydrologic Factors and Reef Growth Space does not permit extending the comparisons in further detail here, but there is something to learn from examining parameters of reef structure and function along latitudinal gradients, and among boundary current systems. The major conclusion I draw is that while temperature alone is a reasonably good correlate of reef distribution, there are enough exceptions to the 18° 'law' (Wells 1957) at high latitude coral reefs to reject temperature as the sole factor limiting reef development. The apparent correlation of coral reef distribution with minimum ocean temperatures (Fig.l) has led to the assumption of causality (Dana 1843, Rosen 1971, Burns 1985). The physiological data for individual coral species is far from conclusive. Corals and algae exhibit a high degree of plasticity in their responses to temperature extremes (reviewed in Brown & Howard 1985, Hatcher et al. 1989), and extrapolations from species-specific responses of local populations to reef building communities is problematical. Certainly instances of extensive natural mortality resulting from thermal stress have been documented (eg Shinn 1976, Roberts et al. 1982, Burns 1985), but active reef growth has been documented in environments where minimum mean temperatures extend well below 18°C (eg Downing 1985, Tribble & Randall 1986, Coles & Fadlallah 1991), and some species of hcrmatypic corals survive in temperatures as low as 11°C (eg MacIntyre & Pilkey 1969, Veron & Marsh 1988). Identification of causality in controls on reef growth is complicated by the correlation of temperature, salinity, irradiance, nutrient and particulate gradients along geographic axes such as latitude and proximity to shore. Gradients of increasingly unsuitable environmental conditions for reef growth are characterized by a decrease in coral abundance and diversity, and an increase in the abundance of macroalgae (particularly Phaeophytes) at higher latitudes and close to continental land masses 117 Journal of the Royal Society of Western Australia, 74,1991 (Johannes et al. 1983a, Birkeland 1988, Coles 1988, Sheppard 1988). Studies of reefs in the middle of this transition (ie supporting both coral and macrophyte- dominated communities) have led to the conclusion that competition for space, light and nutrients between these two groups of benthic organisms is an important factor controlling the development and distribution of coral reefs near the poleward and landward ends of gradients (Johannes et al. 1983a, Hatcher 1985). Complex (often non-linear) interactions amongst water quality parameters (eg nutrient concentration, turbidity, temperature) render the comparative approach of observing reef structure and performance along gradients inconclusive. For example, neither Liddell & Ohlhorst (1988) nor Sheppard (1988) were able to unequivocally identify the determinants of declining reef development along geographical axes in the Western Atlantic or Arabian Gulf regions from synoptic comparisons of coral abundance and diversity. Clearly, no single factor controls the transition from coral reef to kelp bed. Rather, the total environment must be considered in terms of its effects on the competitive abilities of the pool of available benthic organisms. Oceanographic conditions exert major, but not exclusive controls, on this environment. Coastal runoff and its accompanying sedimentation have great potential to modify the influence of boundary currents on reef distribution and development, because of their strongly negative effects on coral growth and survival (reviewed in Birkeland 1988, Hatcher et al 1989). The arid conditions of Western Australia modify the influence of the Leeuwin Current on reef development along the coast to a lesser extent than, for example, the wet coast of east Africa modifies the influence of the Agulhas Current. Examination of the distribution and structure of coral reefs influenced by the Leeuwin Current reinforces the conclusions drawn above. Coral Reefs of the Leeuwin Province Four major and two minor coral reef systems fall under the influence of the Leeuwin Current (Fig.2, Table 1). From north to south they are: 1) The Rowley Shoals, 2) The Dampier Archipelago & adjacent reefs of the Pilbara coast, 3) The Ningaloo reef tract and adjacent reefs, 4) The western islands of Shark Bay, 5) The Houtman Abrolhos reefs and adjacent banks, and 6) The Pocillopora reef at Rottnest Island. With the probable exception of the Rowley Shoals, all of the Leeuwin reefs were high and dry at the start of the Holocene transgression 10,000 years ago. Their colonization and growth is a geologically recent event. It is important to be precise about the definition of a coral reef in this context. Isolated coral colonies, or spatially restricted groups of corals, and their associated epi- and infauna extend at least as far south as Cape Naturaliste. They do not form the vertically or horizontally accreting structures of coral framework, infilled and cemented by detritus and algae, which are here defined as coral reefs. It is quite possible that there are other small reef structures fringing the coast between Northwest Cape and Perth (eg Point Quobba, just north of Shark Bay, Veron & Marsh 1988) which would meet these criteria, but they have not been described adequately. Knowledge of such reefs would be useful in separating coastal from Leeuwin Current influences on coral reef development. The four reef systems differ markedly in their physiography, geology and degree of interaction with the Leeuwin Current. Figure 2 Chart of the coral reefs of Western Australia which are potentially influenced by the Leeuwin Current, showing the 200 m depth contour, the main stream of the Current, and the northern limit of kelp (Ecklonia ) occurrence (K). Minimum mean monthly sea surface temperatures are shown in °C for reef and coastal locations, with absolute minimum values recorded in brackets (from Veron & Marsh 1988). 118 Journal of the Royal Society of Western Australia, 74, 1991 Table 1. A comparison of relevant attributes of coral reefs of Western Australia which potentially fall under the influence of the Leeuwin Current. Diversity = # genera/# species, N/D = No Data, Extreme lower temperatures in brackets. Reef system Reef Type Lat. °S Dist. from Coast km Dist. between Reefs km Surround Depth m Dist. from Slope km Min. mean Temp °C Water Quality Best coral Growth Abundnt Macro- Algae? Coral Diver sity Rowley Shoals Atoll? 17 250-300 300 200-400 0 245 Oceanic, Tidal Slopes & E lagoons No 52/180 Dampier Archr. Fringe 20 5-25 120 10-20 75-100 202 (18.0) Coastal, Tidal NW. slopes Yes 57/216 Monte- bellos Fringe 20.3 80-120 10 30 25-40 22.7 Shelf, Tidal W. slopes No 30/66 Barrow Island Fringe 205 60 120 25 60 N/D Shelf & Coastal W. shore Yes 15/25 Ningaloo Tract Fringe/ Barrier 21- 23 0.2-7 100 15 8-50 22.1 (20.0) Oceanic & Shelf Flat & Lagoons No 54/203 Shark Bay Fringe 24- 26 0 250 25 60-80 N/D Shelf & Coastal W. shores N/D 28/82 Houtman Abrolhos Plat¬ form 28- 29 45-70 330 30-40 10-20 19/8 (17.6) Oceanic & Shelf E. slopes & Lagoons Yes 42/184 Rottnest Island Fringe 32 20 15-30 20-30 185 Shelf & Coastal S. shore Yes 16/25 The Rowley Shoals The Rowley shoals are shelf-slope platform reefs of classic structure and development (Fairbridge 1971). As such, they probably represent a long history of reef growth, with a relatively thick, coral-rich Holocene stratum on the Pleistocene reef basement (Berry 1982). Their position, 700 km northwest of what is generally depicted as the northern limit of the Current at Exmouth (eg Pearce & Cress well 1985), begs the question of their classification as Leeuwin reefs. Currents on the Northwest Shelf, however, exhibit a distinct SW component which is coherent with the Leeuwin Current flow further south (Holloway & Nye 1985), suggesting that this area forms part of the headwaters of the Current. Eastward flowing water from the northern Indian Ocean (and perhaps a southwestward flow from the Timor Sea) undoubtedly bathes the Rowley Shoals as it assembles for its rush south along the shelf slope in the Leeuwin Current. The Dampier Archipelago & adjacent reefs of the Pilbara coast A similar generalization applies at least to the outer portions of the Dampier Archipelago and adjacent Lowendal, Monte Bello and Barrow Islands, which occupy the inner, southeast portion of the Northwest Shelf. The coral reefs comprise an extensive array of small, barrier and fringing reefs growing in shallow water adjacent to hundreds of small continental islands (Simpson 1988). They are primarily Holocene structures, resting on non-reefal basements of continental rock. A large range of reef habitats, from high energy, coral covered outer reef slopes to stagnant, depauperate lagoons is the result of this geographic diversity (UNEP/IUCN 1988). From about February to June the currents on the shelf adjacent to these reefs are predominantly southwest (Mills et al. 1986). The source of the water in these currents is unclear, but its quality is strongly influenced by benthic resuspension and runoff from 119 Journal of the Royal Society of Western Australia, 74,1991 the adjacent coast (Simpson 1988). Whether it is considered part of the Leeuwin Current or not, much of the water flowing through these reef systems during that period soon joins the Current off Exmouth. The Ningaloo reef tract and adjacent reefs The Leeuwin Current is a clearly delineated entity at the Murion Islands off the north end of the Ningaloo reef tract. Australia's largest fringing reef system is separated from the arid coast by only 0.2 to 7 km of shallow lagoon. The best coral reef development occurs in reef passes and within this lagoon. Although the shelf is narrow here, the outer reef slopes are not characterized by rich coral communities extending to great depth, as found, for example at the Rowley Shoals. It appears that the underlying structure of the reef below about 10 m depth is not a Pleistocene reef matrix, but rather a mixture of aeoleanites, calcarenites and tertiary limestones supporting very few corals (May et al. 1983, Veron & Marsh 1988, UNEP/1UCN 1988). The Ningaloo Reef tract is likely to be a thin layer of coral matrix built on old coastal features during the Holocene transgression, rather than an ancient coral reef. Perhaps reduced precipitation during the Holocene allowed the reef to grow closer to shore than had been possible during previous interglacial periods. Having made statements about the structure of Leeuwin reefs, I emphasize that it is impossible to be certain about their underlying composition because no coring has been done. The holes currently being sunk at the Abrolhos (see Collins et al. 1991) should fill a major gap in our knowledge. The pattern of circulation within the Ningaloo reef lagoon ensures rapid exchange with the waters immediately adjacent to the outer reef edge (Hearn et al. 1986). There are few oceanographic data from outside the reef, but the proximity of the shelf break (Fig.2) strongly suggests that the Leeuwin Current is the major reservoir for lagoonal exchange. Shark Bay Narrow fringing reefs partially line the seaward edges of the islands which form the western boundary of Shark Bay. The islands separate the high salinity water of this large inverse estuary from the shelf waters, and the corals are best developed on their seaward margins (Veron & Marsh 1988). Because there has been little documentation of the biota or physical environment of these reefs, it is difficult to assess the influence of the Leeuwin Current on them. The continental shelf is over 100 km wide at this latitude, however, so the potential influence of the Leeuwin Current is likely to be attenuated. The Houtman Abrolhos reefs and adjacent banks On the edge of the shelf between 28 and 29°S latitude lie a series of submerged and emergent reefs which often receive the full force of the Leeuwin Current (Fig.3). While undeniably the products of coral accretion over geological time periods, the Houtman Abrolhos Reefs are unlike any other Australian reef system in morphology and community structure (Saville-Kent 1897, Veron 1986). A Holocene coral veneer is virtually non-existent on the western, that is, seaward portions of the reef platforms. Instead, there grows an unusual mixture of macroalgae of temperate and tropical affinities (Hatcher 1985, Hatcher et al. 1987). Yet the relicts of classic spur and groove topography, symptomatic of healthy reef accretion in high energy environments, are evident. In the eastern portions of the platforms extensive and diverse accumulations of Holocene coral matrix dominate what is clearly a severely eroded Pleistocene reef topography. In places like Turtle Bay in the Wallabi Group, the Pleistocene reef structures arc exposed to reveal rich coral sequences (Teichert 1947, Collins et al. 1991). Several shoals and banks are located on the shelf and slope both to the north and south of the Abrolhos (Figs.2 & 3). To what extent these represent extant or extinct coral reef structures is unknowm. The Leeuwin Current rides the slope 10 to 20 km west of the Abrolhos Reefs, and there appears to be a persistent, large scale (100 + km) cyclonic eddy in the current at this latitude (Fig.2). Sporadically, the warm waters of the Current flood the shelf edge around the Abrolhos, while at other times it forms a narrow jet well to the west (Pearce & Griffiths 1991, Pearce et al. 1991). Rottnest Island The inclusion of Rottnest Island in a classification of coral reefs is debatable. Most corals on this continental island 20 km off the coast near Perth exist as solitary colonies resting on the submerged limestone platforms which surround the Island (Hodgkin et al. 1959). In this respect it represents coral communities found on the coast from Jurien Bay, 150 km north of Rottnest Island, to as far south as Cape Naturaliste and Espcrance. Pocillopora Reef, on the south coast of Rottnest Island, is a coral reef structure of about 3 m maximum relief which is dominated by the one species, but which supports many other tropical invertebrates and fish. As such, it meets the minimum criteria of the reef definition used here. It is unlikely that the underlying structure is a Pleistocene reef matrix. A relict coral reef of late Pleistocene age outcrops on the south coast of the Island nearby at Fairbridge Bluff (Playford 1988). It appears to represent a thin but extensive sequence of coral matrix now obscured by non-reefal, Holocene deposits, and indicates that conditions suitable for coral reef development here are not restricted to the recent past. Rottnest Island is not as proximal to the course of the Leeuwin Current as are the Abrolhos Reefs, but the presence of the Current is evident in sea surface 120 Journal of the Royal Society of Western Australia, 74,1991 — 28°30'S — 29°S 113°E SHELF Wallabi Group HOUTMAN ABROLHOS 35m Pelsaert Group 47m Figure 3 Chart of the Houtman Abrolhos, showing the three major reef platforms (island groups), the 200 m depth contour and representative shelf depths, and generalized circulation features (based on analysis of NOAA-AVHRR imagery, Pearce el al. 1991). Areas of subtidal reef dominated by macroalgal communities (horizontal lines) and coral communities (diagonal shading) surround the islands (solid black). Three mechanisms by which Leeuwin Current water (light stipple), and upwcllcd or intruded water (dark stipple) can influence the reefs on the shelf are depicted. temperatures and the local hydrography (Pearce et al. 1989). Inter-reef comparisons Comparison of the salient features of these reef systems (Fig.2, Table 1) reveals several interesting generalizations about the relationship between coral reefs and the Leeuwin Current. With the possible exception of the northern extremity, where temperatures increase shorewards towards Broome, the Leeuwin Current maintains warmer sea surface temperatures at offshore reefs than on the adjacent coast. South of Shark Bay, winter temperatures near the coast often go below 18°C in August or September. The absolute minimum values are generally about 2°C colder (Table 1): for the most part this is a patchy data set, requiring cautious interpretation. I suggest that the offshore and coastal means at the same latitude are significantly different, but that the ranges usually 121 Journal of the Royal Society of Western Australia, 74, 1991 overlap. The lack of coastal reef development south of Shark Bay suggests either that it is only the mean (rather than the extreme) temperature that matters, or that factors other than temperature are at work. The reefs of the Leeuwin Province may be divided into offshore and coastal groups, the criterion being distance from the shelf break, rather than distance from the coast. The critical distance appears to be about 50 km from the 200 m depth contour (Table 1). Coastal reefs only occur north of the Abrolhos, and are characterised by more turbid water and greater macroalgal abundance than those offshore. This distinction is most apparent in a comparison between the Monte Bello Islands, which have low macroalgal cover and twice the coral diversity of the reef on nearby Barrow Island, which has abundant macroalgae (UNEP/IUCN 1988). There is considerable evidence to demonstrate that the presence of macroalgae inhibits coral settlement, growth and survival (Johannes 1975, Crossland 1981, Hatcher 1985). Unlike offshore reefs to the north of Shark Bay, those at the Abrolhos support luxuriant macroalgal communities, including the kelp Ecklonia radiata (Hatcher et al. 1987), which has its northern coastal limit near Kalbarri (Fig.2). Proximity to the coast, and all it implies in terms of water quality and biotic communities, must thus be seen as another gradient of Leeuwin Current influence, in addition to the more obvious latitudinal cline. The Role of the Leeuwin Current Coral reefs are surface phenomena, and as such are particularly influenced by shallow circulations like the Leeuwin Current. Coral reefs are also multidimensional structures which cannot be classified on any single axis. In attempting to interpret the relationship between the Leeuwin Current and its reefs, the multiplicity of factors which interact to control reef structure and development precludes simple cause and effect assumptions in the absence of experimental data. Virtually all of the data available at present are either circumstantial or correlative. In its simplest guise, the role of the Leeuwin Current in maintaining coral reefs on the coast of Western Australia can be viewed as the flux balance between southern Indian Ocean water derived from the West Wind Drift, and tropical water delivered by the Leeuwin Current. Where mixing is weak and Leeuwin flows dominate, conditions for reef development are good. Where mixing is rapid and Leeuwin flows are diluted and dissipated by their interaction with other water masses, conditions are poor. In this context, one expects to see an attenuation of reef development as one moves away from the centre of the Current (ie across the shelf towards the coast or out into the Indian Ocean), and as one moves south and the gradients across which mixing processes occur steepen. It would be nice to have a simple mixing parameter which could be used to quantify the relative dilution of Leeuwin water along these longitudinal and latitudinal axes. Thompson's (1987) scaling model of the Current at shelf scale might be a good place to start. Several hypotheses can be erected concerning the role of the Leeuwin Current in maintaining coral reefs along the coast of Western Australia. I have listed some of the more obvious ones here: The first might be termed the "Recruitment Hypothesis": "Advective delivery of the larvae of reef building organisms in the Leeuwin Current replenishes local populations after local extinctions, and maintains populations of reef organisms at sites where they are not reproductively viable." Given typical drifter transit times in the Current (Cresswell & Golding 1980), and a maximum distance between any two Leeuwin reef systems of just over 300 km (Table 1), inter-reef transport during the larval stages of most reef-building organisms is entirely feasible (Maxwell & Cress well 1981). Two other hypotheses deal with the effect of the Current on water quality in and around offshore and nearshore reef systems. The "Advective Influence Hypothesis": "Direct advection of Leeuwin Current water maintains elevated temperatures and depressed dissolved nutrient and particulate concentrations which favour the growth of coral reef organisms, and inhibits the growth of macroalgal competitors at offshore sites where the seabed intersects the photic zone and the Current." The "Mixing Influence Hypothesis": "Mixing of Leeuwin Current water onto the shelf has a similar effect as advected water masses, but it is attenuated by the quality and quantity of coastal water with which it mixes." The Advection Hypothesis primarily concerns latitudinal variations in reef development, while the Mixing Hypothesis concerns cross-shelf variation A fourth hypothesis again concerns water quality for reef development at offshore locations: the "Upwelling Inhibition Hypothesis" ’The Leeuwin Current inhibits wind-driven upwelling of colder, nutrient enriched water on the shelf slope and outer shelf. The degree of inhibition is proportional to the magnitude of the Current in relation to the opposing wind-driven flow." The degree of upwelling along the coast of Western Australia has been a matter of debate at least since Schott’s paper of 1933, and the jury is still out. Certainly it is not a pervasive phenomenon, but even sporadic, local delivery of cold, nutrient-rich water to shallow 122 Journal of the Royal Society of Western Australia / 74, 1991 reefs can have profound effects on community structure and function. Processes of upwelling and terrestrial runoff which deliver cold and/or nutrient-rich waters to shallow benthos favour the growth of macroalgac, and have been implicated in the inhibition of reef development. The best examples come from temperature and nutrient proxy records in the skeletons of massive coral colonies collected from currently dead or dying reefs in the eastern Pacific. Increased intensity of coastal upwelling resulting from an equatorward shift in the tradewinds during the little ice age (1500-1850 AD) is implicated in the death of a reef tract on the SW coast of Costa Rica some 150 to 300 years B.P. (Glynn et al. 1983). Markedly increased surface concentrations of inorganic nutrients from the Equatorial upwelling occurred during this same period at the Galapagos, suggesting greatly reduced El Nino (Panama Current) flows, and a possible explanation for the scarcity of modern corals at this site (Shen et al. 1987, Lea et al. 1989, Linn et al. 1990). On a more recent time scale, isotopic variations in the thickness and density of growth bands in corals from many locations have been correlated with intra and interannual variations in seawater conditions as influenced by local and global (ENSO) oceanography (eg Knutson et al. 1972, Smith et al. 1979, Boto & Isdale 1985, Barnes & Lough 1989, Hudson et al. 1989). Finally, the dissipation and mixing of Leeuwin Current water as it moves south leads to the "Latitudinal Attenuation Hypothesis": "All of the effects of the Leeuwin Current are attenuated with latitude, such that a gradient of decreasing diversity, abundance, growth rates and interconnectedness of coral reef communities occurs." A corollary exists in the compressed gradient towards the coast as Leeuwin water is altered by mixing and terrestrial influences. The last is not really a testable hypothesis, in that Leeuwin Current effects on reef development are confounded by latitudinal variation which is independent of the Current, such as the latitudinal gradient in solar radiation. It is thus important to identify the mechanisms by which observed gradients in coral reef distribution, structure and function along the Western Australian coast are maintained. Three arc obvious. Mechanisms of Leeuwin Current Influence A gradient of decreasing species diversity of reef building organisms to the south and cast is an obvious feature of Leeuwin reefs (Table 1). It is complemented by an inefease in the diversity of competing organisms such as kelp. These changes in community structure provide the simplest explanation for the observed decline in the quantity and quality of reef structures along the Leeuwin gradients. Reefs won't develop in the absence of corals to build framework and associated flora and fauna to cement and infill that framework. If we assume that all shallow areas under the influence of the Current have an equal potential for the survival of recruits, then two mechanisms exist to explain the progressive decrease in the reef-building species pool. Either the species have not reached many potential reef sites in the Leeuwin province from some Indonesian source of radiation, because there has been insufficient time for the advective delivery of critical numbers of larvae, or the larvae have all died en route. The first case is unlikely because the dynamics of the Leeuwin Current favour rapid and direct larval advection, the close proximity of its reefs along stream favour larval advection and "island hopping", and there have been at least 6000 years of relatively stable conditions for dispersal. If there are mechanisms which prevent entrainment of larvae in the Current (eg spawning at times of reduced or reversed Current flow), they must be very closely tuned to the Current's dynamics. The only data available suggest that biological mechanisms serve to enhance, rather than inhibit, entrainment of coral larvae (Simpson 1991). In the second scenario gradients in the physical and biotic environment of the Current cause increased larval mortality due to dispersion, thermal stress and starvation in the south and east of the Current stream. This is a more likely mechanism for producing the observed diversity gradients among Leeuwin reefs, but one for which there is little direct data. At the extreme southern end of Leeuwin reef development at Rottnest Island, the recruitment of Pocillopora colonies with genotypes differing from adjacent colonies is very patchy in space and time (Stoddart 1984). The recruitment of tropical fish species to Pocillopora Reef is also a rare event compared to tropical reefs (Hutchins 1991), suggesting that depletion of both larval abundance and diversity down the current is an important phenomenon. If, on the other hand, we assume that reef-building organisms have equal potential for recruitment to shallow locations within the Leeuwin province, then the observed pattern of reef distribution and development is due to variations in post-recruitment growth and mortality. In this case differences among reefs are due to differences in their local physical, chemical and biotic environments. Observed variation in potentially controlling factors such as reduced sea temperature and increased nutrients, particulates and coral competitors southwards and shorewards are strongly influenced by the Leeuwin Current. Gradients in the marine environment, largely maintained by the Current's flow southwards, provide the most obvious mechanism for maintaining gradients in coral reef development along the Western Australian coast. A third mechanism is variation in habitat diversity. If we again assume that reef-building organisms have equal potential for delivery to all Leeuwin reef sites 123 Journal of the Royal Society of Western Australia / 74,1991 (admittedly unlikely), then differences among reefs may be due simply to variation in the range of habitats available for colonization and suitable for accretion at any given site. Habitat diversity is undoubtedly a major determinant of coral species diversity at the scales of individual reefs (Rosen 1971, Veron 1986). There is no evidence, however, for the consistent geographic variation in habitat diversity at the scale of the Leeuwin Current, which would produce the observed patterns of species diversity (Table 1). Indeed, some of the most southerly sites, such as Rottnest Island and the limestone reefs north of Perth, offer extreme habitat complexity, yet support species-poor coral and coral- associated communities (Hodgkin et al. 1959, Marsh 1974). The mechanisms proposed here are not mutually exclusive. Undoubtedly all three operate to some extent in producing the pattern of reef development within the Leeuwin Province. On the available data, it appears that the Leeuwin Current exerts its major influence by maintaining gradients of temperature, dissolved macronutrients and particulate organics which determine the growth and survival of reef¬ building organisms in planktonic and benthic stages of their life histories, through both direct physiological effects, and indirect competitive (and predation) effects. Understanding the influence of the Leeuwin Current on the structure and function of its coral reefs is not simply a question of determining effects on the environmental factors which currently control benthic community structure, growth rates and other parameters of existing reefs, although that is a good place to start. Also required is a knowledge of the geological history of their development, both in terms of changes in the Leeuwin Current itself, and changes in global parameters such as sealevel, surface radiation and source pools of colonizer species. Fortunately, while temporally integrating environmental influences on their development, coral reefs also record many of those environmental signals. Our best clues to the past influence of the Leeuwin Current on coral reef development off Western Australia lie within the calcium carbonate matrix of the coral skeletons themselves. This record is only now being exposed (Collins et al. 1991, Pearce et al. 1991). The Houtman Abrolhos - Australia’s Galapagos The Houtman Abrolhos epitomize the anomalous ecological consequences of the Leeuwin Current. Southernmost in the Indian Ocean, the Houtman Abrolhos are coral reef communities at the limits of existence, extending, with the help of the Leeuwin Current, well into a region dominated by macroalgal communities. Northernmost in the South Pacific, the macroalgal communities of the Galapagos are at their limits of existence also, extending well into a region dominated by coral reef communities, with the help of the Peru Current and upwelling of the Equatorial Undercurrent. In both archipelagos, coral and macroalgal communities vie for dominance of the substrata, with the outcome apparently dependent on their local oceanography. One of the most striking and counter-intuitive aspects of the Abrolhos reefs is the lack of vital coral communities on the most characteristic coral-built structures: the windward reef slopes (Fig.3). Clearly, what used to be actively accreting coral structures are now kelp beds (Wilson & Marsh 1979, Hatcher 1985). When did this transition take place? What was the cause of it? Is it an ongoing process? What is its relationship to the structure and dynamics of the Leeuwin Current? Analysis of NOAA AVHRR satellite images demonstrates that both direct advection of Leeuwin Current water onto the platforms (shelf flooding and cross-shelf streams), and cross-shelf mixing between Leeuwin & shelf water can influence the Abrolhos (Pearce et al. 1991, Fig.3). Whatever the mechanism, the portions of the Abrolhos likely to experience the strongest Leeuwin Current influence are the western reef margins and adjacent lagoons: the areas with the poorest reef communities (Fig.3), and no apparent accretion. In contrast, the eastern margins and lagoons experience more contact with cooler, nutrient enriched coastal water masses moving north along the shelf under the influence of the prevailing winds (CresswclJ et al. 1989). Yet these portions of the reefs support thu richest coral communities south of Ningaloo (Table 1, Fig.3), high community production and calcification rates (Smith 1981), and obvious vertical and horizontal accretion (Wilson & Marsh 1979, Hatcher 1985, Collins et al. 1991, Hatcher, unpublished data). It is possible that the present distribution of corals within the Abrolhos is related to human activities there (Hatcher et al. 1990). Perhaps the distribution reflects differences in wave energy. Many of the robust corals which characterise wave swept reef fronts do not occur at the Abrolhos (Veron & Marsh 1988, Veron pers.comm.): have those species simply not been able to survive the trip? There is no evidence to suggest that corals adapted to high energy environments are less capable of dispersion. The alternative explanation is that conditions for the growth of corals on the exposed portions of the Abrolhos reefs are unsuitable for coral growth and survival. Experimental results demonstrate that macroalgae are able to outcompete corals for light and space at the Abrolhos, particularly in the absence of intense herbivory (Crossland 1981, Johannes et al. 1983a, Hatcher & Rimmer 1985, Hatcher, unpublished data). The nutrient concentrations required to maintain high biomass communities dominated by macroalgae are also known to characterize the lagoons and adjacent waters of the Abrolhos reefs (Johannes et al. 1983b, Crossland et al. 1984, Hatcher 1985). Indeed, nutrient 124 Journal of the Royal Society of Western Australia, 74, 1991 concentrations at the Abrolhos are the highest ever recorded in an unpolluted coral reef ecosystem (Crossland 1983), averaging 3 to 12 times the mean values in the centre of the Leeuwin Current stream. The concentration gradient is hypothesized to be largely the result of the decomposition of macroalgae advected into the lagoons from the windward reef slopes (Crossland et al. 1984, Hatcher 1983, 1985, unpublished data). While nutrients are effectively sequestered and recycled by the reef systems of the Abrolhos, the efficiency cannot be perfect. New nutrients necessary to maintain the observed macroalgal growth and concentration gradient must be supplied from outside the reef systems at time scales at least approximating the turnover times of the dominant macroalgae: seasonally to annually. Cross-shelf mixing of coastal waters, enriched by terrestrial inputs (Cresswell et al. 1989) is a possible source of new nutrients to the Abrolhos ecosystems, but the distribution of macroalgal communities on the reefs argues against it as the major source. The satellite images occasionally show tongues of cold water off the western margins of the reefs (Fig.3; Pearce, Hatcher & Wyrwoll, unpublished data). If these represent upwelled water, enriched in nutrients from off reef sources, then they could explain the high biomass of macroalgae (including the kelp Ecklonia) at these sites. A temperature logger at 5 m depth on the north end of the Western Reef in the Easter Croup (Fig.3) records sporadic, several degree drops in sea temperature lasting 2 to 6 days (Pearce & Hatcher, unpublished data). Without simultaneous nutrient data this is not conclusive evidence of upwelling. It is notable that the upper depth limit of E. radiata on the western reef slopes at the Abrolhos occurs at about 5 m, while the most luxuriant growth is from 15 to 45 m. Upwelling (or "uplifting" cf. Rochford 1991) on these reef slopes may rarely reach the sea surface. The case of Ecklonia growing on the fringing coral reefs of the coast in the Gulf of Oman provides a fascinating counterpoint to the Abrolhos situation. There the kelp also does not grow above 5 m depth. It has developed an annual growth strategy which is tuned to the well-documented and highly seasonal monsoon-driven upwelling. The nutrient-needy sporophyte is found only during the upwelling period. For the remainder of the year, the plants survive as microscopic gametophytes, and the benthic community looks like a typical coral reef (Barratt et al. 1984). In the Gulf of Oman, upwelling allows the development of kelp beds, on coral reefs within the tropics. The answers to the related questions of the source of nutrients to support macroalgal growth, and the time course of macroalgal versus coral domination of benthic communities at the Abrolhos lie in an improved understanding of the regional oceanography. and of the geological growth history of these reefs. The necessary research is in progress (Pearce et al. 1991, Collins et al. 1991) and the results will provide the best evidence to date on the role of the Leeuwin Current in the development of coral reefs along the coast of Western Australia. Conclusion It was obvious to naturalists of the last century that the regional oceanography of Western Australia controlled the development of reefs along its coast (Saville-Kent 1897). We are now in a position to state the obvious with some authority, and to frame testable hypotheses about the mechanisms of control. The definitive statements, however, must await the next century. Acknowledgements. I thank W.R. Black for presenting this paper orally, and C.J. Crossland and A.I. Hatcher for providing insightful criticism of the manuscript. References Barratt L, Ormond R F G, Campbell A C, Hiscock S, Hogarth P J & Taylor J D 1984 Ecological study of rocky shores on the south coast of Oman. IUCN/UNEP, Cambridge/Geneva, 104 pp. Barnes D J & Lough J M 1989 The nature of skeletal density banding in scleractinan corals: fine banding and seasonal patterns. J Exp Mar Biol Ecol 126:119-134. Berry P F (ed) 1982 Report on a preliminary survey of the fauna of the Rowley Shoals by the Western Australian Museum in July 1982. West Aust Museum, Perth, 46pp. Birkeland C 1988 Geographic comparisons of coral reef community processes. Proc Sixth Int Coral Reef Symp 1 : 211 - 220 . Boto K G & Isdale P 1985 Fluorescent bands in massive corals result from terrestrial fluvic acid inputs to nearshore zone. Nature 315:247-255. Brown BE & Howard L S 1985 Assessing the effects of "stress" on reef coral. Adv Mar Biol 22:1-63. Burns T P 1985 Hard coral distribution and cold-water disturbances in South Florida: Variation with depth and location. Coral Reefs 4:117-124. Coles S L 1988 Limitations in reef coral development in the Arabian Gulf: Temperature or algal competition? Proc Sixth Int Coral Reef Symp 3:211-216. Coles S L & Fadlallah Y H 1991 Reef coral survival and mortality at low temperatures in the Arabian Gulf: New species-specific lower temperature limits. Coral Reefs 9:231-237. Collins L B, Wyrwoll K-H & France R E 1991 The Abrolhos carbonate platforms: geological evolution and Leeuwin Current activity. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. AF Pearce and DI Walker). J Roy Soc WA 74:47-57. Cresswell G R, Boland F M, Peterson J L & Wells G S 1989 Continental shelf currents near the Abrolhos Islands, Western Australia. Aust J Mar Freshw Res 40:113-128. Cresswell G R & Golding T J 1980 Observations of a south- flowing current in the southeastern Indian Ocean. Deep-Sea Res 27A.449-466. 125 Journal of the Royal Society of Western Australia, 74,1991 Crossland C J 1981 Seasonal growth of Acropora cf. formosa and Pocillopora damicornis on a high latitude reef (Houtman Abrolhos, Western Australia). Proc Fourth Int Coral Reef Symp 1:663-667. Crossland C J 1983 Dissolved nutrients in coral reef waters. In: Perspectives on Coral Reefs (ed DJ Barnes), Aust Inst Mar Sd, Townsville. 56-68. Crossland C J, Hatcher B G, Atkinson M J & Smith S V 1984 Dissolved nutrients of a high latitude coral reef, Houtman Abrolhos Islands, Western Australia. Mar Ecol Prog Ser 14:159-163. Dana J D 1843 On the temperature limiting the distribution of corals. Am J Sd 45:130-131. Davies P J & Montaggioni L 1985 Reef growth and sea-level change: the environmental signature. Proc Fifth Int Coral Reef Symp 3:477-511. Downing N 1985 Coral reef communities in an extreme environment: the northwestern Arabian Gulf. Proc Fifth Int Coral Reef Symp 6343-348. Fagerstrom J A 1987 The Evolution of Reef Communities. John Wiley & Sons, New York, 600pp. Fairbridge R W 1971 Coral Reefs of the Australian Region. In: Landform studies from Australia and New Guinea (eds JN Jennings & JA Mabbutt) ANU Press, Canberra. Godfrey J S & Ridgway K R 1984 The large-scale environment of the poleward-flowing Leeuwin Current, Western Australia: longshore steric height gradients, wind stresses, and geostrophic flow. J Phys Oceanogr 15:481 - 495. Glynn P W, Druffel E R M & Dunbar R B 1983 A dead central American coral reef tract: Possible link with the Little Ice Age. J Mar Res 41:605-637. Hatcher A I, Wright G D & Hatcher B G 1990 Resolving the conflict between conservation values and extractive use of the Abrolhos coral reefs. Proc Ecol Soc Aust 16:55-70. Hatcher B G 1983 The role of detritus in the metabolism and secondary production of coral reef ecosystems. In: Proceedings of the Inaugural Great Barrier Reef Conference (eds JT Baker et al.), James Cook Univ, Townsville. 317- 325. Hatcher B G 1985 Ecological Research at the Houtman Abrolhos: High latitude reefs of Western Australia. Proc Fifth Int Coral Reef Symp 6:291-297. Hatcher B G, Kirkman H & Wood W F 1987 The growth of the kelp Ecklonia radiata near the northern limit of its range in Western Australia. Mar Biol 95:63-73. Hatcher B G, Johannes R E & Robertson A I 1989 Review of research relevant to the conservation of shallow tropical marine ecosystems. Oceanogr Mar Biol Ann Rev 27:337-414. Hatcher B G & Rimmer D W 1985 The role of grazing in controlling community structure on a high latitude coral reef: Measurements of grazing intensity. Proc Fifth Int Coral Reef Symp, Antenne Muscum-EPHE, Moorea, 6:229-236. Hearn C J, Hatcher B G, Masini R J, Simpson C J 1986 Oceanographic processes on the Ningaloo reef. Western Australia. Centre for Environmental Fluid Dynamics, Univ of West Aust, Perth, Report ED-86-171,121pp. Hodgkin E P, Marsh L M & Smith G G 1959 The littoral environment of Rottnest. J R Soc West Aust 42:85-88. Holloway P E & Nye H C 1985 Leeuwin Current and wind distributions on the southern part of the Australian North West Shelf between January and July 1983. Aust J Mar Freshw Res 36:123-137. Hudson J H, Powell G V N, Robblee M B & Smith T J 1989 A 107 year-old coral from Florida Bay: Barometer of natural and man-induced catastrophes? Bull Mar Sci 44:283- 291. Hutchins J B 1991 Dispersal of tropical fishes in temperate seas of the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. AF Pearce and DI Walker). J Roy Soc WA 74:79-84. Johannes R E 1975 Pollution and degradation of coral reef communities. In: Tropical marine pollution (eds EJF Wood & RE Johannes) Elsevier, Amsterdam, 13-51. Johannes R E, Wiebe W J, Crossland C J, Rimmer D W & Smith S V 1983a Latitudinal limits of coral reef growth. Mar Ecol Prog Ser 11:105-111. Johannes R E, Wiebe W J & Crossland C J 1983b Three patterns of nutrient flux in a coral reef community. Mar Ecol Proe Ser 12:131-136. Knutson D W, Buddemier R W & Smith S V 1972 Coral chronometers-seasonal grow'th bands in reef corals. Science 177:270-272. Lea D W, Shen G T & Boyle E A 1989 Coralline barium records temporal variability in equatorial Pacific upwelling Nature 340:373-376. Liddell W D & Ohlhorst S L 1988 Comparison of Western Atlantic coral reef communities. Proc Sixth Inti Coral Reef Symp 3:281-286. Linn L J, Delaney M L &c Druffel E R M 1990 Trace metals in contemporary and seventeenth century Galapagos coral: Records of seasonal and annual variations. Geochim Cosmochim Acta 54:387-394. Macintyre I G & Pilkcy O H 1969 Tropical reef corals: tolerance of low temperatures on the North Carolina continental shelf. Science 166:374-375. Marsh L M 1974 Benthic communities of Limestone reefs off Marmion, Western Australia. M.Sc. Dissertation, The Univ. of Western Australia, 119pp. Maxwell G J H & Cresswell G R 1981 Dispersal of tropical marine fauna to the Great Australian Bight by the Leeuwin Current. Aust J Mar Freshw Res 32:493-500. May R F, Lenanton R C J & Berry P F 1983 Ningaloo Marine Park. Report and Recommendations by the Marine Park Working group. National Parks Authority Rept No.l, Perth, 67pp. Menard H W 1986 Islands. Scientific American Books. New York. 230pp. Mills D A, Pitt D R & Simpson C J 1986 Summary of current meter data collected from the Dampier Archipelago: 1981-1984. W Aust Dept Conserv Environ Note No. 178, Perth, 11pp. Neumann G & Pierson W J 1966 Principles of Physical Oceanography. Prentice-Hall, New Jersey, 545pp. Pearce A F 1991 Eastern boundary currents of the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc W A 74:35-45. Pearce A F & Cresswell G 1985 Ocean circulation off Western Australia and the Leeuwin Current. CSIRO Div Oceanogr Info Sheet No. 16-3, 4pp. Pearce A F & Griffiths R W 1991 The mesoscale structure of the Leeuwin Current. J Geophys Res C96:16739-16757. 126 Journal of the Royal Society of Western Australia, 74, 1991 Pearce A F, Prata A J & Manning C R 1989 Comparison of NOAA/AVHRR-2 sea-surface temperatures with surface measurements in coastal waters. Int J Remote Sensing 10:37- 52. Pearce A F, Hatcher B G & Wyrwoll K-H 1991. Using satellite imagery to interpret the growth history and paleoceanography of the Houtman Abrolhos reefs: patterns of the Leeuwin Current around the Abrolhos. CSIRO/UWA Collaborative Research Grant - Final Report, 20pp. Playford P E 1988 Guidebook to the Geology of Rottnest Island. Geol Soc Aust, Perth, 67pp. Roberts H H, Roberts L J, Walker N D, Hudson J 11 1982 Cold w r ater stress in Florida Bay and Northern Bahamas: A product of winter cold-air outbreaks. J Sediment Petrol 52:145-155. Rochford D J 1991 'Upwelling': does it need a stricter definition? Aust J Mar Freshw Res 42:45-46. Rosen B R 1971 The distribution of reef coral genera in the Indian Ocean. Symp Zool Soc London 28:263-299. Saville-Kent W 1897 The naturalist in Australia. Chapman & Hall, London, 302pp. Schott G 1933 Upwelling off Western Australia? Yes or No? (In German) Ann d Hydr usw 61:225-233. Sheppard C R C 1988 Similar trends, different causes: responses of corals to stressed environments in Arabian seas. In: Proc Sixth Inti Coral Reef Symp (eds Choat el al .) James Cook Univ Townsville, 3:297-302. Shen G T, Boyle E A & Lea D W 1987 Cadmium in corals as a tracer of historical upwelling and industrial fallout. Nature 328:794-796. Shinn E A 1976 Coral reef recovery in Florida and the Persian Gulf. Env Geol 1:241-254. Simpson C J 1988 Ecology of Scleractinian corals in the Dampier Archipelago, Western Australia. W Aust Environ Protect Auth Tech Ser No. 23, Perth, 238pp. Simpson C J 1991 Mass spawning of corals on Western Australian reefs and comparisons with the Great Barrier Reef. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. AF Pearce and DI Walker). J Roy Soc WA 74:85-91. Smith S V 1981 The Houtman Abrolhos Islands: carbon metabolism of coral reefs at high latitude. Limnol Oceanogr 26:612-621. Smith S V, Buddemeier R W, Redalje RC & Houck J E 1979 Strontium-Calcium thermometry in coral skeletons. Science 204:404-407. Stoddart J A 1984 Genetic differentiation amongst populations of the coral Pocillopora damicornis off southwestern Australia. Coral Reefs 3:149-156. Teichert C 1947 Contributions to the geology of the Houtman’s Abrolhos, Western Australia. Proc Linn Soc NSW 71:145-196. Thompson RORY 1987 Continental-shelf-scale model of the Leeuwin Current. J Mar Res 45:813-827. Thompson R O R Y & Veronis G 1982 A poleward boundary current off Western Australia. Aust J Mar Freshw Res 34:173-185. Tribble G W & Randall R H 1986 A description of the high- latitude shallow water coral communities of Miyake- jima, Japan. Coral Reefs 4:151-159. UNEP/IUCN 1988 Australia, Western. In: Coral Reefs of the World. Volume 2: Indian Ocean, Red Sea and Gulf. UNEP Regional Seas Directories & Bibliographies. IUCN, Gland, Switzerland & Cambridge, UK/UNEP, Nairobi, Kenya. 389pp. Veron JEN 1986 Corals of Australia and the Indo-Pacific. Angus & Robertson, Sydney, 644pp. Veron J E N & Marsh L M 1988 Hermatypic corals of Western Australia. Rec W Aust Mus Supplt No.29, Perth, 136pp. Wells J W 1957 Coral Reefs. Geol Soc Am Mem 67:609-631. Wilson B R & Marsh L M 1979 Coral reef communities at the Houtman Abrolhos, Western Australia, in a zone of biogeographic overlap. Proc Int Symp Mar Biogeog South Hemisphere, NZ DSIR Info Ser 137, pp.259-278. Journal of the Royal Society of Western Australia, 74, 1991,129-132 Seabird abundance, distribution and breeding patterns in relation to the Leeuwin Current R D Wooller, J N Dunlop, N I Klomp, C E Meathrel & B C Wienecke Biological Sciences, Murdoch University, WA 6150, Australia Abstract Lack of upwelling, and low marine productivity, results in seabirds being much less abundant off Western Australia than along the coasts of western South America and south-west Africa. The breeding and non-breeding distributions of seabirds appear to be influenced by the presence of the Leeuwin Current, as do the timing and success of their breeding activity. For instance, in a year of strong Leeuwin Current flow. Little Penguins near Perth carried less food, were in poorer condition and laid eggs much later than in a year of weaker flow. Abundance On the western coasts of southern continents, regions of coastal upwelling often support rich, but locally restricted, seabird assemblages. The diversity of species in these communities is often similar but their abundances may differ greatly. Off Peru, about four million Peruvian Boobies Sula variegala breed, together with three million Guanay Cormorants Phalacrocorax bougainvillii and one million Peruvian Pelicans Pelecanus thagus (Duffy et al. 1984). The total numbers of seabirds off Peru are thought to have varied between four million and 20-30 million over the last century (Duffy & Siegfried 1987). Off the south¬ western coast of Africa the most abundant breeding seabirds are the Cape Cormorant Phalacrocorax capensis (280,000), the African Penguin Spheniscus demersus (170,000) and the Cape Gannet Sula capensis (80,000 birds). Off the western coastline of Australia, upwelling is much less pronounced than along the coasts of western South America or south-west Africa (Pearce 1991), and populations of seabirds arc much smaller (Serventy et al. 1971). Indeed, seabird densities off Western Australia were never large enough to accumulate massive guano deposits like those harvested off south¬ western Africa, Chile and Peru. In most years, the Bengucla and Humboldt Current systems off the western coasts of southern Africa and South America, and the associated upwellings which bring nutrients into the euphotic zone, result in large schools of pilchards, sardines, anchovies and sprats, which provide food for seabirds. It has long been known that seabird populations in the Peruvian and Ecuadorian coastal areas periodically experience major natural disasters due to malnutrition. A warm southward countercurrent usually only displaces the cold, north-flowing water between December - January and March - April, but occasionally strengthens and persists for a year or more, thereby reducing fish stocks (Schreiber & Schreiber 1984). It is unclear how far the Leeuwin Current affects the abundances of seabirds off south-western Australia but the current has been linked to the changing distributions of seabird species in this region and to their patterns of breeding (Dunlop & Wooller 1990). Non-breeding distributions Pelagic seabird species from the Pacific Ocean recently recorded from the waters off Western Australia have been attributed to a marine continuity between the Leeuwin Current and the tropical, western Pacific Ocean (Dunlop et al. 1988a). These species include Tahiti Petrels Pterodroma rostrata , Bulwcr's Petrel Bulweria bulwerii , Streaked Shearwaters Calonectris leucomelas, Hutton's Shearwaters Puffinus huttoni, Fluttering Shearwaters P. gavia and Matsudaira's Storm-petrels Oceanodroma matsudairae (Dunlop el al. 1988a, 1988b). Breeding distributions Along the western coast of Australia, some tropical seabirds breed much further southward than in eastern Australia, or even elsewhere in the world (Serventy et al 1971). Common Noddies Anous stolidus, Lesser Noddies A. tenuirostris and Sooty Terns Sterna fuscata all have large breeding populations on the Houtman Abrolhos, well beyond their normal latitudinal limits 129 Journal of the Royal Society of Western Australia, 74,1991 (Harrison 1983). Roseate Terns Sterna dougallii breed in Warnbro Sound and Bridled Terns S. anaethetus off Cape Leeuwin, both reaching the southernmost limits of their worldwide breeding ranges off the Western Australian coast. Some marked extensions of breeding range have also been recorded relatively recently. After colonising Lancelin Island, probably from Abrolhos populations, the Roseate Tern started breeding in the Fremantle area about 1982 (Dunlop & Wooller 1986). Bridled Terns occurred no further south than the Abrolhos in 1839-1843/ reached the Safety Bay islands by 1920, bred on Hamelin Island by 1955 and off Cape Leeuwin by 1957 (Serventy et al 1971). On Penguin Island, in Shoalwater Bay, no Bridled Terns bred in 1940-42 but a substantial, and growing, breeding population had become established by the early 1980s (Dunlop et al. 1988c). The Red-tailed Tropicbird Phaethon rubricauda has bred intermittently on the Houtman Abrolhos and from 1957 to 1959 on Rottnest Island (Storr 1964, Serventy et al. 197D/ although the nearest large, stable breeding population of this species is on Christmas Island, in the eastern Indian Ocean (Harrison 1983). More recently, a small population of Red-tailed Tropicbirds have bred off Cape Naturaliste since 1966 (Serventy el al 1971). The distributions of all these essentially tropical species appear to reflect the influence of the Leeuwin Current. Seabird assemblages off southwestern Australia are often paradoxical, for instance near Fremantle, where tropical Bridled and Roseate Terns breed beside cool- water Little Penguins Eudyptula minor. On the Abrolhos, tropical Sooty Terns, Common Noddies and Lesser Noddies breed on the same islands as cool- water species such as Little Shearwaters Puffinus assimilis, White-faced Storm-petrels Pelagodroma marina and Pacific Gulls Larus pacificus. Off the south¬ west corner of Australia, Bridled Terns and Red-tailed Tropicbirds nest alongside species from cooler southern waters, such as the Fleshy-footed Shearwater Puffinus carneipes . Such paradoxes may be accounted for by the southward extension of tropical seabirds associated with the Leeuwin Current. Breeding seasons In south-eastern and southern Australia, seabirds typically breed in spring/summer, whereas in tropical northwestern and northern Australia most breed between March and June. However, on the mid- western and southwestern coasts of Australia, several seabird species breed both in autumn (March-June) and in spring (August-Novembcr), some breeding continuously from autumn to spring. Double-breeding or protracted breeding are seen in Crested Terns Sterna bergii, Bridled Terns, Roseate Terns, Pied Cormorants Phalacrocorax varius, Silver Gulls Larus novaehollandiae and Little Penguins (Dunlop & Wooller 1986). Roseate Terns breed on islands from the Houtman Abrolhos to the Fremantle area in either autumn or spring, the seasonally distinct breeding groups being interspersed throughout these islands. On the Abrolhos, and several other islands, both autumn¬ breeding and spring-breeding colonies occur on the same island. Of the two recently established colonies. Roseate Terns breed in spring on Lancelin Island but in autumn in Shoalwater Bay (Dunlop & Wooller 1990). Most Crested Terns in southwestern Australia breed from August to December, but autumn-breeding colonies are known from the Houtman Abrolhos, the Fremantle area and east of Hopetoun. On Rottnest Island, autumn breeding did not appear to start until about 1977-1978, presumably as a result of an invasion of autumn-breeders from colonies on the Abrolhos or off the Pilbara coast (Dunlop & Wooller 1990). Detailed observations of individually marked Crested Terns in the Fremantle area have shown that individuals have a broadly circannual reproductive cycle and comprise two groups which are reproductively distinct, although some young born in spring have joined the autumn-breeding group (Dunlop 1985, Dunlop & Wooller 1990). In the Silver Gull and the Little Penguin, breeding is greatly protracted and egg-laying shows two or more peaks. Both species are potentially double-brooded and readily replace lost clutches (Nicholls 1974, Wooller & Dunlop 1979, Dunlop et al. 1988b). Thus, protracted breeding results from sequential, successful, and unsuccessful, breeding attempts by the members of a single population, rather than by separate populations, as seen among terns. Breeding success The timing, strength and characteristics of the Leeuwin Current vary seasonally and from year to year. This variability appears to affect the reproductive performance of some seabirds, such as Little Penguins. Since 1986, most of the five hundred Little Penguins breeding on Penguin Island, near Rockingham, have been individually marked, measured and their reproductive success monitored in 55 nest-boxes and a similar number of natural nest-sites in bushes. The stomach contents of penguins coming ashore at dusk in 1986 and 1989 were also analysed (Klomp & Wooller 1988, Wienecke 1989). In 1989, the mean body weights of male and female Little Penguins were significantly less than during 1986 (Table 1; t = 13.77 for males and t = 7.90 for females, both p < 0.01). These samples did not include moulting penguins but even non-moulting penguins vary seasonally in weight. Therefore, a condition index was calculated (body weight (g) + (head length (mm) + beak depth (mm))) to compare condition independently of size. In 1989, both males and females had significantly lower condition indices than during 1986 (Table 1). 130 Journal of the Royal Society of Western Australia / 74, 1991 Table 1 The weights and condition of Little Penguins on Penguin Island, Western Australia, and their reproductive performance in 1986 and 1989. Sample sizes are shown in parentheses. 1986 1989 Leeuwin Current weaker stronger Sea surface temperature cooler warmer Mean (±S.E.) body mass (g) Males 1570 ±19 (152) 1432 ± 16 (145) Females 1363 ±13 (143) 1181 ±13 (132) Mean (± S.E.) condition index (g) Males 125 ±12 (208) 117 ±12 (200) Females 118 ±10 (182) 106 ±10 (159) Mean clutch size 1.9 1.9 Mean (±S.E.) egg weight (g) 53.8 ±0.7 53.8 ±0.2 Mean (±S.E.) egg length (mm) 56.2 ±0.5 56.6± 0.3 Mean (±S.E.) egg width (mm) 42.6 ±0.3 42.3 ±0.3 Percentage eggs hatched 64% (198) 65% (120) Mean young per pair 0.6 (104) 0.7 (63) The first penguin eggs were laid in April in both years and laying continued until December, typical of other years monitored. However, the main laying period was 1-2 months later in 1989 than 1986 (Figure 1). In addition, a lower proportion of birds bred in 1989 than in 1986, although clutch size, egg dimensions, hatching success and overall reproductive output were almost identical in both years (Table 1). Of the 128 penguins sampled in 1989, 15.6% carried no food, significantly more than the 6.3% of 234 penguins sampled in 1986 (Xj 2 = 72.1, pcO.OOl). The samples from penguins with food in 1989 (12.5 ± 2.1 g) were much smaller than in 1986 (57.2 ± 10.3 g). However, the prey taken, mostly small, schooling fish. Were very similar in both years (Figure 2). Figure 1 The cumulative monthly percentage of eggs laid by Little Penguins monitored on Penguin Island, Western Australia, during 1986 and 1989. 1986 1989 N = 234 N = 128 from Little Penguins ashore on Penguin Island, Western Australia, during 1986 and 1989. Sea surface temperature at the edge of the continental shelf, near Perth, varied seasonally in a similar manner in both 1986 and 1989, but was up to 2°C warmer in 1989 (Figure 3). Temperatures closer inshore, where penguins feed, paralleled those further offshore, although they were more variable, with seasonal ranges of about 8°C inshore and 5°C offshore. The mean annual sea level at Fremantle, an indicator of the presence and strength of the Lceuwin Current (Pearce & Phillips 1988), was high in 1989, following a similarly high level in 1988. In contrast, 1986 and 1987 had low mean annual sea levels at Fremantle, presumably reflecting a much weaker Leeuwin Current, and cooler surface waters, in those years. Food samples were not taken from Little Penguins during 1987 and 1988. However, the condition of the penguins was relatively good during 1987 but poorer in 1988. In 1989, the later laying period, poorer condition and lower proportion of birds breeding may have resulted from a warmer sea surface temperature adversely affecting the Little Penguin, which is essentially a cool-water species. This effect is, presumably, mediated through a lower abundance or availability of the schooling fish which form the diet of these penguins. Commercial catches of baitfish were. 131 Journal of the Royal Society of Western Australia, 74,1991 indeed, lower in 1989 than in 1986 (R. Lenanton, pers. comm.) but, in the absence of detailed information on fish stocks and distribution, it is only possible to infer the links between oceanographic events and seabird reproductive success at present. Figure 3 Mean monthly sea surface temperatures (°C) off Rottnest Island, Western Australia, during 1986 and 1989. Data kindly provided by A.F. Pearce. The invariability in the reproductive effort and success of those penguins which did breed may represent the minimal viable effort in accord with genetic fitness. The non-breeding individuals appear to defer breeding until later in the season, or until a later year, rather than attempt to breed sub-optimally. This strategy may have evolved under the variable oceanographic, and probably trophic, conditions produced by the Leeuwin Current. Interestingly, the Bridled Tern, a tropical species, appears to show a converse effect, arriving and laying earlier in years of stronger current flow and warmer sea surface temperatures (Dunlop & Jenkins, unpubl. obs.). Conclusions Variation in the strength of the Leeuwin Current appears to influence the reproduction and mortality of seabirds off Western Australia less dramatically than variation in the El Nino -Southern Oscillation can affect seabirds in the eastern and central Pacific. However, the presence of more tropical, and fewer cool-water, seabird species breeding off southwestern Australia seems clearly linked to this warm water current. The timing and duration of breeding seasons in the region also appear to reflect the seasonal effects of the Leeuwin Current, although much remains to be learned about the mechanisms underlying these relationships. References Duffy D C & Siegfried W R 1987 Historical variations in food consumption by breeding seabirds of the Humboldt and Benguela upwelling regions. In: Seabirds: feeding ecology and role in marine ecosystems (ed J P Croxall) Cambridge University Press, Cambridge. 327-346. Duffy D C, Hays C & Plenge M A 1984 The conservation status of Peruvian seabirds. In: Status and conservation of the world’s seabirds (eds J P Croxall, P G H Evans & R W Schreiber) ICBP Technical Publication No 2, Cambridge. 245-259. Dunlop J N 1985 Reproductive periodicity in a population of Crested Terns Sterna bergii Lichtenstein in southwestern Australia. Aust Wildl Res 12:95-102. Dunlop ] N & Wooller R D 1986 Range extensions and the breeding seasons of seabirds in southwestern Australia. Rec West Aust Mus 12:389-394. Dunlop J N &c Wooller R D 1990 The breeding seabirds of southwestern Australia: trends in species, populations and colonies. Corolla 14:107-112. Dunlop J N, Cheshire N G & Wooller R D 1988a Observations on the marine distribution of tropiebirds, sooty and bridled terns, and gadfly petrels from the eastern Indian Ocean. Rec West Aust Mus 14:237-247. Dunlop J N, Klomp N I & Wooller R D 1988b Penguin Island,Western Australia. Corolla 12:93-98. Dunlop J N, Wooller R D & Cheshire N G 1988c Distribution and abundance of marine birds in the eastern Indian Ocean. Aust J Mar Freshw Res 39:661-669. Harrison P 1983 Seabirds: An identification guide. Croom- Helm, Beckenham. Klomp NI & Wooller R D 1988 Diet of Little Penguins Eudyptula minor from Penguin Island, Western Australia. Aust J Mar Freshw Res 39:633-639. Nicholls C A 1974 Double-brooding in a Western Australian population of the silver gull Laws novaehollandiae Stephens. Aust J Zool 22:63-70. Pearce A F 1991 Eastern boundary currents of the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. AF Pearce and DI Walker). J Roy Soc WA 74:35-46. Pearce A F & Phillips B F 1988 ENSO events, the Leeuwin Current, and larval recruitment of the western rock lobster. ] Cons Int Explor Mer 45:13-21. Schreiber R W & Schreiber E A 1984 Central Pacific seabirds and the El Nino Southern Oscillation: 1982 to 1983 perspective. Science 225:713-716 Serventy D L, Serventy V & Warham J 1971 The handbook of Australian seabirds. AH &AW Reed, Sydney. Storr G M 1964 The avifauna of Rottnest Island, Western Australia. I. Marine birds. Emu 64:48-60. Wienecke B C 1989 The breeding patterns of Little Penguins on Penguin Island, Western Australia, in relation to dietary and oceanographic factors. Unpubl. B.Sc. Hons. Thesis, Murdoch University. Wooller R D & Dunlop J N 1979 Multiple laying by the Silver Gull Larus novaehollandiae Stephens, on Carnac Island, Western Australia. Aust Wildl Res 6:325-355. 132 Journal of the Royal Society of Western Australia / 74, 1991,133-140 Implications of long-term climate change for the Leeuwin Current C B Pattiaratchi 1 & S J Buchan 2 ^Centre for Water Research, The University of Western Australia, Ned lands , WA 6009, Australia 2 Steedman Science and Engineering, 31 Bishop Street, Jolimont, WA 6014, Australia Abstract The Leeuwin Current is an anomalous, poleward flowing, eastern boundary current, which brings water (and associated marine biota) of warm tropical origin to the temperate south-west and the Great Australian Bight. The Current is driven by an alongshore steric height gradient which is due to the inter-connection between the Indian and Pacific oceans through the Indonesian Archipelago and the density structure of the Indian Ocean. The Current flows all year round but exhibits a strong seasonality with the stronger flows occurring during the winter months (May - July) and is weaker during the summer (December to January). This is reflected in the coastal sealevcls off Western Australia which may be used as an indication of the strength of the current. During ENSO events, the Current is also weaker due to changes in the equatorial Pacific Ocean. Under an 'Enhanced Greenhouse warming scenario, there is potential for the driving force of the Leeuwin Current, and its consequent influence on the biota of coastal waters, to be changed. This paper reviews the driving mechanisms of the Current and its annual and inter-annual variability. Selected scenarios under an enhanced greenhouse warming are examined to determine their impact on the strength and location of the Leeuwin Current. It is shown that, although there is a degree of uncertainty on the likely manifestations of the enhanced greenhouse effect, various scenarios indicate a possible decrease in the alongshore steric height gradient resulting in a weaker Leeuwin Current. An increase in the northward wind stress during the summer months is also predicted which could lead to a weaker Current during the summer months than present and more frequent upwelling along the West Australian coast. Introduction Increased emissions of carbon dioxide, methane, chlorofluorocarbons (CFC’s) and nitrous oxide from human activity have resulted in additional warming of the Earth's surface. This is termed the 'Enhanced Greenhouse Effect’ (IPCC 1990). The global mean air temperature is predicted to rise at a rate of 0.3°C per decade (with an uncertainty range of of 0.2°C to 0.5°C per decade) over the next century. Associated with this warming is a re-distribution of heat resulting in changes to the global climate system. This, in turn, will alter the global precipitation patterns, weather systems, frequency of climate extremes and also produce a rise in the mean sealcvel (IPCC 1990). The ocean circulation is driven mainly by the global heat budget and, hence, it is envisaged that changes to the global ocean circulation may result from the enhanced greenhouse effect. This paper examines the possible effect of climate change on the strength and location of the Leeuwin Current off Western Australia. The Leeuwin Current, a poleward eastern boundary current off the West Australian (WA) coast, is a shallow (< 300 m) narrow band (< 100 km wide) of relatively warm, lower salinity water of tropical origin that flows southward, mainly above the continental slope from Exmouth to Cape Leeuwin (Crcsswell & Golding 1980, Pearce & Cress well 1985, Church el al . 1989, Crcsswell 1991). At Cape Leeuwin it pivots eastward, spreads onto the continental shelf and flows towards the Great Australian Bight. Satellite imagery has shown that the Current is a complex of meanders, jet-like streams and eddies, and the structure and behaviour of the Current vary monthly (Legeckis & Crcsswell 1981, Pearce 1985, Prata & Wells 1990, Pattiaratchi el al. 1990). The Current is an important feature locally as it influences the climate of Western Australia (Gentilli 1991) and the local fishing industry (Pearce & Phillips 1988, Stequert & Marsoc 1989, Lenanton el al. 1991). Similar to the other southern hemisphere ocean basins, the Indian Ocean accommodates a general anti-clockwise gyre which includes the westward flowing 133 Journal of the Royal Society of Western Australia, 74,1991 South Equatorial Current from 5°S to 15°S latitude, the strong southward flowing Agulhas Current off the east coast of Africa and the eastward flowing West Wind Drift south of 40°S latitude. Traditional models of ocean circulation postulate a broad, northward return flow off the WA coast, termed the West Australian Current (see for example Tchernia 1980). However, although a net northward flow of water must exist to maintain continuity of the water circulation, field studies undertaken during the 60s and 70's failed to provide any evidence for this equatorward flow. With the advent of satellite tracked drogues and infra-red satellite imagery, the poleward flowing Leeuwin Current was identified (Cresswcll 1991). It is now known that the West Australian Current is located seaward of the Leeuwin Current and may extend over more than half of the Indian Ocean as a very slow return flow towards the equator (Thompson k Vcronis 1983). A unique feature of the Indian Ocean circulation is the inflow from the Pacific Ocean through the Indonesian Archipelago. With the exception of the common inter-connection between all the oceans with the Southern Ocean, this is the only connection between any two ocean basins and is an important factor in the generation of the Leeuwin Current. This paper investigates the possible effect of long¬ term climate change on the strength and location of the Leeuwin Current by reviewing the proposed generating mechanisms of the Current and the observed variability on seasonal and inter-annual time scales. Various greenhouse scenarios (see for example, IPCC 1990) arc examined to identify predictions which may alter the driving forces of the Current. Assuming that any change to the generating forces, as a result of climate change, will influence the strength and location of the Current, predictions are made on the likely behaviour of the Leeuwin Current under an enhanced greenhouse scenario. Generating mechanisms Mechanisms for the generation of the Leeuwin Current have been studied by several investigators (Church et al. 1989, Godfrey k Ridgway 1985, Pearce k Cresswcll 1985, Thompson 1984, 1987, Weaver k Middleton 1989, Batteen k Rutherford 1990, Godfrey k Weaver 1991). There is general consensus that the driving force of the Leeuwin current is an alongshore 134 Journal of the Royal Society of Western Australia, 74,1991 steric height gradient which overwhelms the opposing equatorward wind stress. The source of the Leeuwin Current water is from the Indian Ocean from the west and a component (which originates from the Pacific Ocean) from the North West continental shelf. The South East Trade Winds, in the Pacific Ocean, drive the South Equatorial Current westwards advecting warm surface waters towards Indonesia. This results in the flow of warm, low-salinity water from the western Pacific Ocean through the Indonesian Archipelago into tropical regions of the Indian Ocean. This, together with geostrophic inflow of water from the Indian Ocean, results in the sealevel in the tropics being some 55 cm higher than that along the southern coast of Australia (Pearce & Cresswell 1985). The formation and location of the Leeuwin Current are illustrated schematically in Fig. 1. The meridional gradient of steric height induces a weak geostrophic eastward flow of central Indian Ocean subtropical water toward the coast, between latitudes of 15°S and 35°S. The easterly flow of subtropical water is deflected southward along the edge of the West Australian continental shelf. In the north, the inflow is augmented by tropical water from the North West Shelf. Further south, the continuous inflow from the west accelerates the flow towards Cape Leeuwin, before it turns eastward into the Great Australian Bight. The relative contributions of North West Shelf and central Indian Ocean water, and the mechanism for sustaining the strong meridional steric height gradient, are still under investigation. Initial investigations (Godfrey & Golding 1981) suggested that the Pacific-Indian Ocean throughflow may be sufficient to sustain the Leeuwin Current. More recent modelling studies (Godfrey & Weaver 1991, Weaver & Middleton 1989) indicate that the Leeuwin Current is largely unaffected by the throughflow magnitude (though the density profile water in the Indonesian region is important, since it controls the longshore pressure gradient along the Leeuwin Current). McCreary et al. (1986) suggest that vertical mixing is necessary to support the steric height gradient. The modelling work of Battcen & Rutherford (1990) confirms that the Leeuwin Current can be maintained by the mean thermal structure of the Indian Ocean, but it may be enhanced by the addition of warmer North West Shelf waters. Godfrey & Weaver (1991), using climatological data from Levitus (1982), argue that the propagation of internal Kelvin waves through the Indonesian Archipelago and subsequent western propagation of internal Rossby waves, allows for approximate equilibrium of the specific volume anomaly (SVA) profiles (in the upper few hundred metres) on the North West Shelf, with those in the western equatorial Pacific. The resulting relatively warm pool of surface water, means that surface temperatures are above global equilibrium temperatures (Haney 1971) for west coast water south of 15°S latitude. Consequently, the water of the Leeuwin Current may be expected to lose heat to the atmosphere, resulting in convective overturn and the formation of deep mixing layers. This is confirmed by the observations of Hamilton (1986). There now seems to be general acceptance of the importance of this surface cooling in maintaining the strength of the meridional steric height gradient. Observed variability Seasonal changes In order to examine the fluctuations in the strength of the Leeuwin Current over seasonal and inter-annual time scales, some measure of the intensity of the Current over the complete geographic area of influence is required. In the absence of continuous field measurements of currents, some other measurement related to the Leeuwin Current must be used. Using sea surface temperature (SST) distributions derived from satellite imagery, Prata et al. (1989) have defined a 'Leeuwin Current Index (LCI)'. However, the availability of satellite derived SST distributions is limited (approx. 10 years of data) and hence cannot be used to examine long-term changes. Many investigators (see for example, Sturges 1974, Reed & Schumacher 1981, Pearce & Phillips 1988) have shown that changes in mean sealevel monitored at tide gauges may be used to derive oceanographic information such as variations in flow and/or changes in thermohaline properties. For the Leeuwin Current, Pearce & Phillips (1988) have assumed that changes in the strength of the Current are reflected in mean sealevel changes which have an annual mean amplitude of 20 cm (Fig. 2). During October to March the Leeuwin Current is weaker as it flows against the maximum southerly winds, whereas between April and August the Current is stronger as the southerly winds are weaker (Godfrey & Ridgway 1985). This is reflected in both the SST distributions derived from satellite imagery (Prata et al. 1989, Pearce & Prata 1990) and the mean sealevel at Fremantle (Fig. 2). Month Figure 2 Monthly mean sealevel at Fremantle between 1959 & 1989 indicating that the sealevel has a seasonal amplitude of 20 cm and the maximum occurs during June (Data courtesy of the Tidal Laboratory, Flinders Institute for Atmospheric and Marine Science). 135 Journal of the Royal Society of Western Australia, 74, 1991 Here, the sealevel is higher between April and August when the Leeuwin Current is stronger (lower wind stress) and lower between October and January when the Current is weaker (high wind stress). Geographical distribution of the seasonal variations in mean sealevel along the west coast of Australia indicates a progressive feature (Fig. 3 and Pariwono et al. 1986). On the North West Shelf, the maximum occurs during March whilst in the South West corner, the maximum occurs in May or June (Fig.s 2 and 3); this seasonal movement of the sealevel maximum reflects the southward passage of the Leeuwin Current pulse (Church et al. 1989). Port Group Figure 3 Geographic distribution of the seasonal variation in sealevel (cm) along the Western Australian coastline. Port Groups: NW - north-west (Darwin and Port Hedland); W - West (Geraldton, Fremantle and Bunbury); SW - south-west (Albany and Espcrance). (after Pariwono et al. 1986). In summary, although the Leeuwin Current flows all year round, it exhibits a strong seasonality with the stronger flows occurring during the winter months (May - July) which is reflected in the coastal mean sealevel (Fig. 2). Godfrey & Ridgway (1985) have also shown that there is a very good correlation between the coastal mean sealevel at Geraldton and the steric sealevel. Hence, the mean sealevel at Fremantle (or at any other south-west coast station) may be used as an indicator of the strength of the Current. Inter-annual Changes El Nino-Southern Oscillation (ENSO) events are the result of complex interactions between the ocean and the atmosphere in the tropical Pacific Ocean and have been associated with climatic and environmental anomalies around the world (Philander 1990). Two or three times each decade anomalously warm water, approximately 2-4°C above normal, appears off the coast of Peru and Ecuador and persists for a number of seasons. Normally the Peruvian coast is a region of strong coastal upwelling (Pearce 1991). During ENSO events, however, warm equatorial water from the western Pacific Ocean is transported eastward and flows southwards along the Peruvian coast to replace the cold, nutrient-enriched waters. It is now known (Philander 1990) that during an ENSO event, there is high surface pressure over the western and low sea surface pressure over the south-eastern tropical Pacific Ocean. This coincides with heavy rainfall, unusually warm surface waters and relaxed Trade Winds in the central and eastern tropical Pacific (Philander 1990). 1969 1972 1976/77 1982/83 1986 Figure 4 Time series of the annual Southern Oscillation Index (the normalised difference in surface atmospheric pressure between Darwin and Tahiti, a measure of the potential ENSO events), west coast sealevel (a measure of the strength of the Leeuwin Current) and the Puerulus Settlement Index (a measure of rock lobster recruitment) along the West Australian coast (from Pearce & Phillips 1988). The arrows indicate ENSO events. 136 Sea le'Jet Journal of the Royal Society of Western Australia, 74,1991 Pearce & Phillips (1988) have demonstrated a strong correlation between the Southern Oscillation Index (SOI, the normalised difference in surface atmospheric pressure between Darwin and Tahiti, a measure of the potential of ENSO events), west coast sealevels (a measure of the strength of the Leeuwin Current, see above) and the Puerulus Settlement Index (a measure of recruitment to the rock lobster fishery). During normal years, the coastal annual mean sealevels are relatively high indicating that the Leeuwin Current is strong and the settlement of pueruli in coastal reefs is relatively high. During ENSO years, coastal sealevels fall and the inferred transport in the Leeuwin Current is weaker (Fig. 4). Extension of this time series to include the annual Fremantle sealevel data for the period 1897 to 1990 indicates that each ENSO event during this period (extracted from Quinn et al. 1987) is associated with a transient decrease in the annual mean sealevel (Fig. 5). This confirms the findings of Pearce & Phillips (1988) and Prata el al. (1989) that the Leeuwin Current is weaker during ENSO years. A weaker Leeuwin Current during an ENSO event may be explained as follows: in a 'normal' situation, the South East Trade Winds in the Pacific Ocean set up high steric heights at the north end of the Australasian continent; the gradient between these high steric heights and the thermally-set low steric height off southwestern Australia drives the Leeuwin Current. During ENSO years, the Trade Winds relax and the steric height at the north end of the Australasian continent is lower. This results in a decreased alongshore pressure gradient along the West Australian coastline resulting in a weaker Leeuwin Current. Figure 5 Annual mean sealevel anomalies at Fremantle between 1887 and 1990. The arrows indicate the occurrence of ENSO events as documented by Quinn et al. 1987. A decrease in the mean sealevel is seen to be associated with ENSO events (Data courtesy of the Tidal Laboratory, Flinders Institute for Atmospheric and Marine Science). Scenarios The dominant issue in the current discussion deals with scenarios for long-term climate change caused by the enhanced greenhouse effect due to increased concentrations of carbon dioxide, methane, CFC's and other trace gases since the industrial revolution. Various scenarios have been proposed based on general circulation model (GCM) simulations (see for example IPCC 1990, Evans 1990) and there is substantial uncertainty on the likely manifestations of the enhanced greenhouse effect. Of these, the following are of relevance in examining the potential changes to the strength and location of the Leeuwin Current. (i) A global mean air temperature increase of approximately 1°C above the present value by 2025 and 3°C by the end of the next century will lead to a coincident increase in the sea surface temperature. This warming will not be uniform throughout the globe. It is expected that the warming of the mid-latitudes will be higher than in the equatorial regions (IPCC 1990). (ii) GCM's have so far had limited success in simulating realistic ENSO events (McCreary & Anderson 1991). Hence, there is no clear indication as to the likely changes in the frequency of ENSO events. (iii) The West Wind Belt (the Roaring 40's) may contract poleward by 5° to 10° latitude, therefore increasing the equator-pole pressure gradients. The Sub-tropical High Pressure Ridge should also move south and may broaden. Weaker Trade Winds are likely (Siegfried et al. 1990). (iv) The mid-troposphere in the tropics will warm to a greater degree than the lower atmosphere, suppressing vertical convection and enhancing wind shear (Evans 1990). (v) The equatorward alongshore wind stress during the summer months may increase (Bakun 1990). The different heating rate between the tropics and mid-latitude waters (see (i) above) may result in a decrease in the alongshore steric height gradient driving the Leeuwin Current contributing to a decrease in the strength of the flow. Weaker Trade Winds in the equatorial Pacific (see (iii) above) may result in a decrease in the strength of the South Equatorial Current (i.e. a decrease in the pooling of warmer water against the Indonesian Archipelago). In terms of the ocean circulation, this effect would be similar to that observed during ENSO events resulting in a weaker Leeuwin Current. The resultant alteration in the specific volume anomaly (SVA) profile in the surface waters would be transferred to the North West Shelf waters, and perhaps lead to a decrease in surface water cooling. This in turn may also reduce the alongshore steric height gradient, weakening the driving mechanism of the Leeuwin Current. Godfrey & Weaver (1991) have shown, from 137 Journal of the Royal Society of Western Australia, 74, 1991 modelling studies, that if the SVA profile in the western Pacific is replaced by that in the eastern Pacific, a ’’Peru Current”, i.e. equatorward flow together with upwelling, would be established along the West Australian coast. However, current GCM’s are unable to predict the likely changes in the SVA profile in the ocean under an enhanced greenhouse scenario and, hence, it is not possible to predict whether a reduced Leeuwin Current or even a "Peru Current” would be present off Western Australia. Although there is no clear indication as to the likely changes in the frequency of the ENSO events (see (ii) above) a long-term change in the mean value of the Southern Oscillation Index (SOI) would be important for the intensity of the Leeuwin Current. If under an enhanced greenhouse scenario the mean value of SOI increases (decreases), then the Leeuwin Current will be stronger (weaker). The strong equatorward alongshore wind stress during the summer months is maintained by a strong atmospheric pressure gradient between a thermal low- pressure cell that develops over the heated land mass and the higher barometric pressure over the cooler ocean (Bakun 1990). It has been shown that the Leeuwin Current is weaker during the summer months as it flows against the maximum southerly wind stress (Godfrey & Ridgway 1985). Under an enhanced greenhouse scenario, the gradient between the two pressure systems over land and the ocean may be enhanced, resulting in an intensification of the equatorward wind stress (see (v) above). This would lead to further weakening of the Current during the summer months with the possibility of more frequent upwelling. Analysis of wind data from the major oceanic upwelling areas (Peru, California, Canary Current systems) have shown (Bakun 1990) that the equatorward alongshore wind stress has increased over the past 40 years leading to an intensification of the coastal upwelling systems. This result may indicate that the equatorward wind stress has already increased due to the enhanced greenhouse effect (Bakun 1990). Implications for Biota The presence of tropical marine organisms off the west coast of Australia and in the Great Australian Bight has been attributed to the Leeuwin Current (Maxwell & Cresswell 1981, Cresswell 1985). The Current also plays an important role in the life cycle of the southern blue fin tuna (Thunnus maccoyii) which has it's spawning area off the North West Shelf (Fig. 6). The larvae and young fish (< 2 years old) are carried southwards by the Leeuwin Current and are found in the Great Australian Bight and off the east coast of Australia. Papers appearing in this issue have also identified the role of the Leeuwin Current in the distribution of seagrass and algae (Walker 1991), coral spawning and distribution (Simpson 1991, Hatcher 1991), western rock lobster (Pearce & Phillips 1988), coastal scallop and fin fish stocks (Lenanton el ai 1991) and the sea bird distribution (Wooller el al. 1991). Hence, it is clear that the Current plays a major role in the biota off the west and south coasts of Australia With regard to possible changes in the Leeuwin Current under an enhanced greenhouse scenario, those marine organisms which are dependent on higher temperatures associated with the Current may not be greatly influenced, as a slackening of the Current may be countered by a global sea surface warming. Biota which are dependent on the advective processes of the Current, such as the western rock lobster and southern blue fin tuna, may be more seriously affected. However, a weaker Leeuwin Current and an increase in the northward wind stress may also give rise to more frequent coastal upwelling (see above). This alternate nutrient enrichment may enhance the productivity associated with the continental shelf waters. Figure 6 Spawning area and principal migration routes of the Southern blue fin tuna (Thunnus maccoyii). The dotted areas show the Australian fleet fishing grounds (from Stequert & Marsac 1989). Conclusions This paper has reviewed the proposed generating mechanisms of the Leeuwin Current and observed changes at annual and inter-annual time scales. Possible future changes to the location and strength of the Current and associated biota as a result of the enhanced greenhouse warming have been discussed. Based on these, the main conclusions are: (a) The Leeuwin Current is driven by an alongshore steric height gradient which is generated due to the inter-connection between the Indian and Pacific Oceans through the Indonesian Archipelago (Godfrey & Ridgway 1985) and the density structure of the Indian Ocean (Batteen & Rutherford 1990). (b) The Current flows all year round but exhibits a strong seasonality with the stronger flows occurring during the winter months as reflected in the mean sealevel at coastal stations. This annual variation in the Current is due mainly to changes to the northwards component of wind stress and also to a 138 Journal of the Royal Society of Western Australia / 74,1991 slight reduction in the steric height gradient. During October to March the Leeuwin Current is weakest as it flows against the strong northwards wind stress, whereas between April and August the Current is strongest as this wind stress is weaker. (c) The mean sealevel at Fremantle (or at any other south-west coast station) may be used as an indicator of the strength of the Current. (d) During ENSO events, the Trade Winds relax and the South Equatorial Current in the equatorial Pacific is weaker with a corresponding decrease in the alongshore pressure gradient resulting in a weaker Leeuwin Current. (e) Although there is great uncertainty on the likely manifestations of the enhanced greenhouse effect, various scenarios relevant to determining the strength and location of the Current indicate a possible decrease in alongshore steric height gradient which may result in a weaker Leeuwin Current. Acknowledgements: The authors would like to thank. Prof. G. Lennon (Tidal Laboratory, Flinders Institute for Atmospheric and Marine Science) for providing the tidal data for Fremantle and Dale Robertson, Stuart Godfrey and Alan Pearce for critically reading the manuscript. This paper is Centre for Water Research reference No. ED 566 CP. References Bakun A 1990 Global climate change and intensification of coastal ocean upwelling. Science 247:198-201. Batteen M L & Rutherford M J 1990 Modelling studies of eddies in the Leeuwin Current: the role of thermal forcing. J Phys Oceanogr 20:1484-1520. Church J A, Cresswell G R & Godfrey J S 1989 The Leeuwin current. In: Poleward flow along eastern ocean boundaries, (eds Neshyba S, Mooers CNK, Smith R L & Barber R T) Springer-Verlag Lecture notes on Coastal and Estuarine studies 34: 230-252. Cresswell G R 1991 The Leeuwin Current - observations and recent models. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:1-14. Cresswell G R 1985 The role of the Leeuwin Current in the life cycles of several marine creatures. Proc Int Conf on Pelagic Biogeogr UNESCO Technical papers in marine science 49:60-64. Cresswell G R & Golding T J 1980 Observations of a south¬ flowing current in the southeastern Indian Ocean. Deep Sea Res 27:449-466. Evans J L 1990 Envisaged impacts of enhanced greenhouse warming on tropical cyclones in the Australian region. CSIRO Div Atmos Res Tech Paper No 20,31pp. Gentilli J 1991 I lomologous peri-oceanic west coast climates in the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:15-33. Godfrey J S & Golding T J 1981 The Sverdrup relation in the Indian Ocean, and the effect of Pacific-Indian throughflow on Indian Ocean circulation and on the East Australian Current. J Phys Oceanogr 11:771-779. Godfrey J S & Ridgway K R 1985 The large-scale environment of the poleward- flowing Leeuwin current. Western Australia: Longshore steric height gradients, wind stresses and geostrophic flow. J Phys Oceanogr 15:481 - 495. Godfrey J S & Weaver A J 1991 Is the Leeuwin Current driven by Pacific heating and winds ? Prog in Oceanogr 27:225-272 Hamilton L J 1986 Statistical features of the oceanographic area off south-western Australia, obtained from bathythermograph data. Aust J Mar Freshw Res 37:421-436. Haney R L 1971 Surface thermal boundary condition for ocean circulation models. J Phys Oceanogr 1:241-248. Hatcher B 1991 Coral reefs in the Leeuwin Current - an ecological perspective. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:115-127. IPCC (International Governmental Panel on Climate Change) 1990 Scientific assessment of climate change. WMO/UNEP 25pp. Legeckis R & Cresswell G R 1981 Satellite observations of sea- surface fronts off the coast of western and southern Australia. Deep Sea Res 28:297-306. Lenanton R C, Joll L, Penn J & Jones K 1991 The influence of the Leeuwin Current on coastal fisheries of Western Australia. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:101-114. Levitus S 1982 Climatological atlas of the world ocean. NOAA Prof Paper 13 US Dept of Comm Nat Ocean and Atmos Admini. McCreary J P, Shetye S R & Kundu P K 1986 Thermohaline forcing of eastern boundary current: with application to the circulation off the west coast of Australia. J Mar Res 44:71-92. McCreary J P & Anderson D L T 1991 An overview of coupled ocean-atmosphere models of El Nino and the southern oscillation. J Geophys Res 96 (Suppl): 2125-3150. Maxwell J G IT & Cresswell G R 1981 Dispersal of tropical marine fauna to the Great Australian Bight by the Leeuwin Current. Aust J Mar Freshw Res 32:493-500. PariwonoJ I, Bye J A T & Lennon G W 1986 Long-period variations of sea-level in Australia. Geophys J R Astronom Soc (Lond) 87:43-54. Pattiaratchi C B, Parslow J, Pearce A F &c Hick P T 1990 Application of archived coastal zone colour scanner (CZCS) data for productivity and circulation studies of the Leeuwin Current, Western Australia. Proc 5th Australasian Remote Sensing Conference Perth Volume 1:252-255. Pearce A F 1985 The Leeuwin Current as viewed from space. FINS 19:3-5. Pearce A F 1991 Eastern boundary currents of the southern hemisphere. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:35-45. Pearce A F & Cresswell G R 1985 Ocean circulation off Western Australia and the Leeuwin Current. CSIRO Inf Sheet, 16-3: 4pp. Pearce AF & Phillips B F 1988 ENSO events, the Leeuwin current and larval recruitment of the western rock lobster. J duCons int Explor Mer, 45:13-21. Pearce A F & Prata A J 1990 Sea surface temperature transects across the Leeuwin Current. Proc 5th Australasian Remote Sensing Conference Perth Volume 1:252-255. Philander S G 1990 El Nino, La Nina and the southern oscillation. Academic Press, 293pp. 139 74,1991 Prata A J, Pearce A F, Wells J B, Hick P T, Carrier J C & Cechet R P 1989 A satellite sea surface temperature climatology of the Leeuwin Current, Western Australia. Unpubl Final Report to the Marine sciences & Technology Council, 65pp. Prata A J & Wells J B 1990 A satellite sensor image of the Leeuwin Current, Western Australia. Int J Rem Sens 11:173-180. Quinn W H, Neal V T & Antunez de Mayolo S E 1987 El Nino occurrences over the past four and a half centuries. J Geophys Res 92:14449-14461. Reed RK & Schumacher J D 1981 Sea level variations in relation to coastal flow around the Gulf of Alaska. J Geophys Res 86: 6543-6546. Siegfried W R, Crawford R J M, Shannon L V, Pollock D E, Payne A I L & Krohn R G 1990 Scenarios for global- warming induced change in the open-ocean environment and selected fisheries of the west coast of Southern Africa. S Afr ] Sd 86:281-285. Simpson C 1991 Mass spawning of corals on Western Australian reefs and comparisons with the Great Barrier Reef. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). ] Roy Soc WA 74:85-91. Stequert B & Marsac F 1989 Tropical tuna - surface fisheries in the Indian Ocean. FAO Fisheries Technical Paper 282:238pp. Sturges W 1974 Sea level along continental boundaries. ] Geophys Res 79:825-830. Tchernia B 1980 Descriptive regional oceanography. Pergamon Press, 253pp. Thompson RORY 1984 Observations of the Leeuwin Current off Western Australia. J Phys Oceanogr 14:624-628. Thompson RORY 1987 Continental-shelf scale model of the Leeuwin Current J Mar Res 45:813-827. Thompson R O R Y & Veronis G 1983 Poleward boundary current off Western Australia. Aust J Mar Freshwat Res 34:173-185. Walker D I 1991 The effect of sea temperature on sea grasses and algae on the Western Australian coastline. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:71-77. Weaver A J & Middleton J H 1989 On the dynamics of the Leeuwin Current. J Phys Oceanogr 19:626-648. Wooller R D, Dunlop ) N, Klomp N I, Meathrel CE & Wienecke B C1991 Seabird abundance, distribution and breeding patterns in relation to the Leeuwin Current. In: The Leeuwin Current: an influence on the coastal climate and marine life of Western Australia, (eds. A F Pearce and D I Walker). J Roy Soc WA 74:129-132. rarpy i miPPiPin nnuommfint Printer Western AU stf 04343/10/91-1250-L/5577 GARRY L DUFFIELD, Government Printer. Western