Scientific Investigation Into Eel Deaths in Western Victoria

SCIENTIFIC REPORT

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

PAUL LEAHY, RENEE PATTEN, ALEX LEONARD

Publication 1173 December 2007

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

EPA Victoria 40 City Road, Southbank Victoria 3006 AUSTRALIA

PAUL LEAHY, RENEE PATTEN, ALEX LEONARD

December 2007 Publication 1173 ISBN 0 7306 7666 8

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

TABLE OF CONTENTS EXECUTIVE SUMMARY...... 4 1. INTRODUCTION...... 6 2. METHODS ...... 9 Site location and description...... 9 Collection of in-situ water parameters...... 9 Chemical analysis ...... 11 Algal analysis...... 11 Eel collection ...... 11 Blood plasma collection and analysis ...... 11 Eel measurements and maturity determination ...... 11 Estimation of eel mortality...... 12 3. RESULTS ...... 12 Climate conditions in 2006—07 ...... 12 Fish deaths in the summer of 2006—07 ...... 12 In-situ water quality ...... 13 Salinity ...... 17 Turbidity...... 17 pH...... 18 Dissolved oxygen ...... 20 Water conditions in Belfast Lough...... 22 Major ion analysis of water...... 22 Eel maturity and sexual development ...... 23 Eel blood plasma ion analysis in western Victoria...... 24 Treatment of data ...... 24 Comparison of plasma ion concentrations between lakes ...... 25 Comparison of eel blood plasma ion values in western Victoria with published literature...... 26 Metals analysis at Lake Bolac...... 26 Algal analysis...... 26 4. DISCUSSION ...... 31 Water quality conditions in the lakes ...... 31 Blood plasma in western Victorian eels...... 31 The role of pH in eel deaths in western Victorian lakes ...... 31 Potential for fish deaths in Lake Colac...... 32 Climate change implications for fish in shallow western Victorian lakes...... 32 5. CONCLUSIONS...... 34 6. ACKNOWLEDGEMENTS...... 34 7. REFERENCES...... 35

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

EXECUTIVE SUMMARY Findings

Prior to the onset of below-average rainfall in 1997, Major outcomes of 2006–07 monitoring more than 10 inland south-west Victorian lakes held • The common link between all eel deaths was a substantial short-finned eel (Anguilla australis) salinity rise of 20—40 per cent in the four weeks populations. As of autumn 2007 lakes Colac and prior to the deaths and a pH of around 9—9.5. Purrumbete are the only ones that have a large eel population. The remaining water bodies have either • EPA sampled the blood plasma of eels to see if dried up or experienced large-scale eel deaths. their salt levels were within normal ranges. This work showed that the blood plasma salt levels of In Lake Modewarre, Lake Tooliorook and Lake Bolac dying eels in Lake Tooliorook and Lake Bolac was sudden mass eel deaths have occurred. In early 2005 much higher than eels from two other western EPA Victoria began a scientific investigation into the Victorian waterways. cause of the deaths. Over the course of the • The blood plasma salt level found in dying eels investigation, a number of theories as to the cause of from Lake Bolac is higher than any previously the deaths were examined. The available evidence did recorded for eels in scientific literature. not support a number of theories that had been proposed, including pollution, disease and • The actual salinity does not appear to be as overpopulation. important as the rate of increase. Other factors also make the salinity rise lethal. EPA developed a further investigation for the summer • Rising salinity is exacerbated by extremely high pH. of 2006—07 with the aid of the Eel Death Investigation When the water pH is over 9 eels are more Reference Group, which consisted of representatives susceptible to rising salinity. This is likely to be due from key stakeholder groups and technical experts. to the combined role of the gills in regulating both The key focus of the investigation was to examine salinity and pH. whether eel physiology or age influences their response to declining water quality conditions in • Other factors like dissolved oxygen and turbidity western lakes, particularly increased salinity. may play a role in the eel deaths but are not as critical as rising salinity and high pH. Background • There is no evidence of toxic algae being the cause of the eel deaths. Between Spring 2006 and Autumn 2007 a monitoring program was designed to assess changes in water • Metal contamination was investigated again at conditions and eels. The four lakes monitored were Lake Bolac and there is no evidence to suggest Lake Modewarre, Lake Colac, Lake Tooliorook and that this was the cause of the eel deaths. Lake Bolac. EPA collected information on the Future conditions in the lakes using water quality loggers, which recorded information at every hour throughout Major recreational fisheries and a large commercial eel the period. Eel blood was analysed for salt content and industry have been affected by the changes to their sexual maturity was determined. Water sampling Victoria’s Western District lakes. Their recovery for toxin-producing algae was also undertaken. depends on a period of sustained higher rainfall. If sufficient rainfall occurs and some of the salt is Two large eel death events occurred between spring flushed from the lakes, then there is no reason the 2006 and autumn 2007. In late November to early lakes cannot recover. December 2006, an estimated 75,000 eels died at Lake Tooliorook, five kilometres south-west of Drying and associated high salt levels are a natural Lismore. In March 2007 the largest documented eel part of many lake systems in Australia historically. The deaths occurred at Lake Bolac, where an estimated return of a wetter climate has seen a rapid return by 95,000 eels died. It is unlikely that eels remained in the plants and animals that inhabit the lakes. either lake at the end of Autumn 2007. Both of these lakes supported significant commercial eel fishing Western District lakes and climate change operations. The greatest threat to the recovery of the Western The common link between deaths in the 2006—07 District lakes in the long term is likely to be climate summer and deaths in previous summers was a high change. These lakes tend to act like large rainfall evaporation rate, leading to rapidly rising salinity in gauges, with lake levels heavily dependent on rainfall. the lakes. Salinity doubled compared with the previous Sustained below-average rainfall and low run-off into summer at Lake Tooliorook. At Lake Bolac the change lakes leads to low water levels and, eventually, in salinity over the summer was remarkable. In complete drying. The investigation into eel deaths has November 2006 the salinity was one-third that of underlined the susceptibility of even hardy fish like seawater, but by March 2007 the salinity had risen to eels to low rainfall conditions. over twice that of seawater.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

The effect of drought on south-west Victoria in the last 10 years has been an insight into the consequences of climate change on Victoria’s Western District lakes.

The potential for fish deaths in Lake Colac Lake Colac is one of the last large inland lakes in south-western Victoria with a substantial fish population. In Autumn 2007 it was estimated that perhaps over 500 tonnes of fish were present in the lake. If the trends of increasing salinity and pH in Lake Colac continue as they have for the last three years, then it is likely that large-scale fish deaths will occur. EPA’s research has shown that carp are most susceptible to the changes in salinity, followed by redfin and then eels. Unless there is substantial summer rainfall then we predict that deaths will occur in the 2007—08 summer. If pH remains low then we expect that carp and redfin deaths will be followed by eel deaths. If pH rises quickly then deaths may happen in quick succession.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

1. INTRODUCTION 2006—07 summer. In consultation with the Eel Death Investigation Reference Group, which consisted of Since December 2004, EPA Victoria has been representatives from key stakeholder groups and investigating large fish and eel deaths in south- technical experts, a monitoring program was designed western Victoria. Previous reports by EPA have to investigate these issues. The program began in highlighted the extreme conditions present in the spring 2006 and involved continuous water quality Western District lakes during eel deaths. Investigations measurement through deployed water quality loggers, over the 2006—07 summer centred on developing a monthly monitoring of the lakes for algae and water better understanding of short-finned eel (Anguilla chemistry and sampling of eels for maturity and blood australis) populations in western lakes and a detailed chemistry. knowledge of their response to water conditions in the In late November to early December 2006 a large lakes. number of eels died in Lake Tooliorook, five kilometres Leahy et al (2007) summarised EPA’s investigation to south-west of Lismore. In March 2007, probably the date and highlighted remaining questions to be largest number of eels so far died at Lake Bolac answered in the 2006—07 summer. The key areas to (Figures 1 and 2). These came after large eel deaths at be addressed were: Lake Modewarre and Lake Bolac in previous summers (Leahy et al 2007). This report provides an overview • the impact of eel maturity on eel mortality of EPA’s findings on the cause of eel deaths. • the nature of water quality changes in the lakes Monitoring in previous summer periods identified a • the physiological response of eels to changing relationship between rapidly rising salinity and the eel water conditions deaths, suggesting the eels were unable to regulate • the potential presence of algal toxins in the lakes. the rising salinity in their blood. Part of the 2006-07 sampling focused on collecting data to determine if A review of the causes of eel deaths in previous this was the case. summers highlighted the likelihood of deaths occurring in Lake Bolac and Lake Tooliorook in the

Figure 1: Dead eels among seagrass mats at Lake Tooliorook, December 2006

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Figure 2: Dead eels on the shore of Lake Bolac, March 2007

Fish maintain salt and water balance internally blood plasma salinity is only around 25 per cent through osmoregulation, which occurs within a range (Jobling 1995). Numerous studies on several eel that permits normal physiological functioning. The species have shown that eels maintain plasma sodium major osmoregulatory organs in fish are the gills and concentrations within a very narrow range, even when kidneys and, to a lesser extent, the skin, urinary comparing eels adapted to freshwater or to seawater system and gut (Jobling 1995). (for example Kirsch 1972a). Although research has In freshwater, fish must avoid the loss of salts by found that freshwater and seawater-adapted eels have reabsorbing them and by the production of large stable blood plasma concentrations, during the period volumes of dilute urine (Martinez et al 2005). when eels are adapting to seawater, they have Conversely, in saline water, fish must remove excess elevated salt concentrations in their blood plasma, a salt and avoid the loss of water from their bodies phenomenon known as ‘hypermineralisation’ or ‘sea- (Shuttleworth and Freeman 1973). shock’. Gills are the point of exchange between fish and the A review of the scientific literature found there have aquatic environment. This includes the regulation of been no studies of the blood plasma ions of the internal concentration of salts, removal of wastes A. australis, nor A. reinhardtii, the other south-eastern and exchange of gases like oxygen and ammonia Australian eel species. Given the rise in salinity and pH (Maetz 1971). observed in western Victorian lakes (Leahy et al 2007), blood plasma may be an important indicator of The plasma component of blood is the body’s major the response of A. australis to changing water transport mechanism for salt. Vertebrates generally conditions. keep the salt concentration in their blood plasma within a relatively narrow range. Tight control of The major ion salts regulated by osmoregulation are sodium (Na+) and chloride (Cl-), with much lower plasma salt levels is necessary for normal + 2+ physiological functioning (Lutz 1972). concentrations of potassium (K ), magnesium (Mg ), 2+ 2- calcium (Ca ) and other ions including sulfate (SO4 ), The salinity difference between freshwater and 2- - carbonate (CO3 ) and bicarbonate (HCO3 ). In saline seawater is a factor of about 200. However, the water, the regulation of the monovalent ions (Na+, Cl- difference between freshwater and seawater fish and K+) largely takes place in the gills; specifically, in

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

the chloride cells (Jobling 1995). The divalent ions but a small number of fish able to tolerate extreme 2+ 2+ 2- (Mg , Ca and SO4 ) are predominantly regulated conditions because of their specialised adaptations. through the kidney in saline waters (Cleveland and Another possible influence on the osmoregulatory Trump 1969). ability of the eels is their stage of sexual maturation. In considering the osmoregulation of fish, pH must Chan et al (1967) found that, in comparison with also be considered. Water pH is an important yellow, immature European eels (A. anguilla), silver, consideration for fish health. At low pH, fish are mature European eels showed a range of changes susceptible to the high bioavailability of some metals upon transfer to seawater, suggesting a better ability like aluminium (Alabaster and Lloyd 1982). At high pH, to cope with higher salinity. like the levels recorded in western Victorian lakes As described in McKinnon (2006), eels of the genus (Leahy et al 2007), gill ammonia excretion can be Anguilla have complex life cycles, with both freshwater blocked, leading to death (Eddy 2005). Ammonium can and marine phases. Eels undertake extensive seaward substitute for sodium and potassium, blocking the migrations to spawn, after spending much of their lives normal osmoregulation processes (McKenzie et al in freshwater. Eels undergo a process of morphological 2003). One consequence of this is that fish are and physiological change, which enables them to susceptible to combined stress upon the gills: for prepare for a long migration and reproduction. example, high salinity and low dissolved oxygen or Characteristic changes with the onset of migration high salinity and high pH. include the initiation of sexual maturity and migratory High salinity (greater than 50 per cent seawater) and behaviour. Such changes include shrinking of the gut, high pH (9 to greater than 10) are common changes in colour, lengthening of the pectoral fins, characteristics of the Western District lakes in this enlargement of the eyes and development of sexual study. These conditions are similar to those that Wilkie organs (Pease 2004, Tesch 2003). This process is and Wood (1996) describe as being inhospitable for all known as ‘silvering’ due to the characteristic silver colour of migrating eels.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

2. METHODS Fairy, in February 2007. Belfast Lough is geographically close to the lakes, has a similar Site location and description catchment geology and is kept open to the sea for boating. The salinity of Belfast Lough should remain Four lakes were monitored during the 2006—07 relatively constant and similar to that of seawater. monitoring period: Modewarre, Colac, Tooliorook (also Therefore the condition of the eels from this location known as Ettrick Lake) and Bolac (Figure 3). is of interest for comparison to the inland lakes that Table 1 shows data collected during the 2006—07 experienced large changes in salinity. monitoring period. The monitoring program is given in All four monitored lakes have a lunette-type more detail in the following sections. formation, with sandy eastern shores and typically These lakes have been the focus of previous reports muddy western shores. The maximum depth of the by Leahy et al (2007), with the exception of Lake lakes is less than four metres while, in recent times, Tooliorook, which is located five kilometres south-west none has been deeper than one metre. of Lismore. Lake Tooliorook was added as a Lakes Modewarre and Tooliorook are characterised by monitoring location because it had similar abundant growth of submerged seagrass (Ruppia sp.) characteristics to Lake Modewarre and was predicted and typically have low turbidity. In contrast, Lakes to be susceptible to eel deaths in the summer of Bolac and Colac are historically turbid and have had 2006—07. high algal concentrations in their water columns All four lakes have historically had large populations of (Leahy et al 2007). short-finned eel (A. australis), supporting a commercial Collection of in-situ water parameters fishery. These lakes have also been important recreational fisheries, with redfin (Perca fluviatilis) and In-situ water quality was monitored by the use of two brown trout (Salmo trutta) the most important calibrated water quality loggers in each lake. Water recreational species. temperature, dissolved oxygen, pH and electrical conductivity were measured and recorded by all For comparison with the inland lakes, water quality, loggers, with turbidity measured and recorded on chemistry and eels were collected from Belfast Lough, selected loggers only. All measurements were the estuarine lake section of the Moyne River, Port

Figure 3: Location of study water bodies in south-western Victoria

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Table 1: Overview of data collected from October 2006 –April 2007

EELS WATER Location Month Eel Morphology Blood Chemistry Otolith Aging Major Ions In-situs Logger Data Lake Bolac October 2006 November 2006 December 2006 January 2007 February 2007 March 2007 April 2007 Lake Colac October 2006 November 2006 December 2006 January 2007 February 2007 March 2007 April 2007 Lake Modewarre October 2006 November 2006 December 2006 January 2007 February 2007 March 2007 April 2007 Lake Tooliorook October 2006 November 2006 December 2006 January 2007 February 2007 March 2007 April 2007 Belfast Lough October 2006 November 2006 December 2006 January 2007 February 2007 March 2007 April 2007

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

recorded hourly. The loggers were cleaned and preservative, to keep the algae alive. Microscope cell recalibrated on average every three weeks, more concentration counts were performed on the Lugol’s frequently where biofouling was an issue. Water preserved sample, while the unpreserved sample was quality was checked independently using calibrated used for confirmation of identifications. Algae were handheld meters. generally identified to the lowest practical level, Two loggers were located in each lake with generally genus. deployment limited to the shores of the lakes due to shallow water. Although a logger was deployed in the Eel collection centre of Lake Bolac, it had to be removed in Fleets of fyke nets were set in all lakes and Belfast November 2006 as boat access was prohibited by the Lough. Nets were set for two days to ‘fish in’, cleared lower water levels. The loggers were deployed by and then left for 24 hours prior to fish collection. wading out to approximately 100 m from shore and Fishermen chose locations most likely to yield fish. positioned at a depth of 5—20 cm in a minimum of Nets were evenly distributed within Lake Tooliorook 30 cm of water. and Belfast Lough. Due to the soft sediments and difficulty in access on In lakes Bolac, Colac and Modewarre, the placement of the western sides of the lakes, loggers were placed in nets was restricted to their eastern side because of the east and south-east sides at lakes Modewarre and constraints in depth and substrate type. Commercial Bolac and in the south-east and north east at fishermen report reduced net mortality on the eastern Tooliorook. At Bolac, the logger situated in the east shores of these lakes. A random subsample of around was relocated to the north-west shore, due to the 30 eels was taken from the fishermen’s nets, although receding waterline. In Lake Colac, loggers were placed the actual number analysed varied depending on on the eastern and western shores to avoid sabotage, fishing success and mortality in transit. as these were the most secure locations in this lake. During the eel deaths at lakes Tooliorook and Bolac Chemical analysis numerous eels were observed moving lethargically through the water. These eels were easily collected by In addition to the continuous logging of water quality, hand nets and by seine netting and were also water samples were collected for major ions and analysed. alkalinity. Samples were analysed by two National Association of Testing Authorities (NATA) accredited Blood plasma collection and analysis laboratories, both following Standard Methods (APHA For blood analysis, 1—3 mL of blood was collected from 2005). Major ions analysed were sodium (Na+), calcium the truncus arteriosus in euthanased eels using an 18- (Ca2+), magnesium (Mg2+), potassium (K+), chloride (Cl-) gauge syringe. Blood was expelled from the syringe and sulfate (SO 2-). 4 with the needle removed, to avoid physical damage to As reported in Leahy et al (2007) metals, particularly the red blood cells. The blood was allowed to clot at aluminium and selenium, were high in water samples room temperature for two hours, before being collected from Lake Bolac in January 2006 during the centrifuged at 6000 RPM until plasma was clear. The eel deaths. This was presumed to be due to the high plasma was removed and stored at –20 °C until levels of suspended sediments in the water column analysis. (that is, clay and silts). The high suspended sediment Ion analysis was performed within a week of sample levels during the January 2006 sampling prevented collection to avoid concentration of the plasma. Some filtration of the water samples to determine if the red tainting of the plasma was observed, and clearly metals were likely to be soluble and, hence, tainted plasma samples were not analysed. Plasma bioavailable. ions were analysed by two NATA-accredited During the Lake Bolac eel deaths in March 2007 laboratories. metals sampling was repeated to determine whether there were high levels of bioavailable aluminium. Eel measurements and maturity determination Lower suspended sediment in March 2007 meant that water could be filtered through a cellulose acetate Each eel was weighed and external measurements 0.45 μm filter. Filtered and unfiltered samples were were recorded for the left pectoral fin, left eye acidified in the field with nitric acid to pH 1—2. Metals diameter (vertical and horizontal) and total body analysis was undertaken by a NATA-accredited length. For each eel sampled at Bolac, Colac, laboratory, following Standard Methods. Tooliorook and Belfast Lough (no eels were caught at Lake Modewarre), the stage of maturity was calculated Algal analysis using the silvering index developed by Van den Thillart et al (2005). The silvering index uses length, weight, During the monthly algae sampling two samples were mean eye diameter and pectoral fin length. collected: one sample was preserved with Lugol’s solution and another was kept at 15 °C with no

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

The Gonadosomatic Index (GSI= [gonad weight/body term average) in Victoria for the 12 months prior to weight] x 100), a commonly used method of calculating April 2007. sexual development in fish, was also calculated for a Rainfall in south-west Victoria was in the lowest 5 to 10 subsample of eels where gonads could be sampled and per cent of records during this 12-month period. The weighed. consequence of this and the low rainfall since 1997 was that many lakes in the region dried up. By Autumn Estimation of eel mortality 2007 both Lake Modewarre and Lake Bolac were close To estimate the number of dead eels, every visible eel to drying up, with the water depth less than 30 cm. was counted within 10 evenly distributed 100 m long lake edge segments. Eels were counted if visible from Fish deaths in the summer of 2006—07 shore within the 100 m segment. The average number Eel deaths occurred in Lake Tooliorook and Lake Bolac of eels counted per 100 m segment was then in the 2006—07 monitoring period. At Lake Tooliorook extrapolated to the whole lake using the length of the eels began dying around 23 November 2006. Sick and perimeter of the lake. This method is likely to dying eels were observed up until around 18 December underestimate the number of dead eels, as some may 2006. The number of dead eels by mid-December was not be visible below the surface of the water or estimated to be 75,000. Healthy galaxiids were towards the centre of the lake. observed during seine netting in late November 2006 3. RESULTS during the eel deaths. No dead galaxiids were observed during the eel deaths, but this may have Climate conditions in 2006—07 been due to their small size. Previous reports by EPA have highlighted the Because of the high eel mortality, the commercial eel relationship between the deaths and the low rainfall in fisherman ceased fishing the lake in mid-December south-western Victoria (Leahy et al 2007). Figure 4 2006. It also became increasingly difficult to set nets shows the rainfall deficiency (difference from long- in the lake due to lack of depth and dense seagrass

Figure 4: Rainfall deficiencies in Victoria, 1 April 2006 — 31 March 2007 (Source: Bureau of Meteorology)

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

beds and so no eel survey netting was undertaken the actual number of dead eels may be higher than after the eel deaths in December 2006. estimated. At Lake Bolac an estimated 50 to 100 golden perch Small numbers (fewer than 50) of European carp died in early February 2007, but this was not reported (Cyprinus carpio) died in Lake Colac in January and to EPA at the time of the deaths. High eel mortality March 2007. Both incidents were linked to run-off rates of over 25 per cent in fyke nets were observed from rainfall events. Carp were observed showing by commercial eelers from around 13 February 2007 spawning behaviour, and dissection of the carp onwards. However, other than dying in the nets, eels showed they were heavily laden with eggs or male were not observed dead in the lake until 11 March gonads. 2007. Eels were then observed dying and dead over a Carp mortality associated with spawning is frequently two-week period. By 29 March 2007, no live eels were reported to EPA, particularly after rainfall events. The sighted and no eels were caught by commercial eelers carp become lethargic and can get stranded at such in their nets. The estimated number of eels dead at times. In these cases, the carp deaths are considered Lake Bolac in March 2007 was 95,000. to be due to a different process than the eel deaths at Commercial eelers estimated that the number of eels Lake Tooliorook and Lake Bolac. in the lake at the start of summer 2006—07 was around 150,000. Given the heavy fishing prior to the In-situ water quality deaths and the likely underestimate of death event The water quality monitoring results for lakes size, it is unlikely that live eels remained in the lake at Modewarre, Bolac and Colac up until the start of the the end of the 2006—07 summer period. 2006—07 summer period have been presented in Eels were observed attempting to burrow into Leahy et al (2007). For comparison, longer-term sediments in Lake Bolac during the deaths. Eels have electrical conductivity sourced from the Victorian been known to burrow into sediments as a survival Water Quality Monitoring Network (VWQMN) and mechanism (Tesch 2003), and this may be one way earlier EPA investigations for Bolac and Tooliorook is that eels could survive in Lake Bolac. Up to 500 presented in Figures 5 and 6. Water quality monitoring pelicans were observed at Lake Tooliorook in early during the 2006—07 summer period for the four lakes December 2006 and many were observed feeding on is presented in Figures 7—18. eels. The consumption of eels is another reason why

100000 Fish death event WQ monitoring 90000 SE logger North logger 80000 Spot measures East logger 70000 Central logger

60000

50000

40000

30000

Conductivity@25C (µS/cm) Conductivity@25C 20000

10000

1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

Figure 5: Long-term electrical conductivity in Lake Bolac, 1977 — April 2007 (Monitoring data from www.vicwaterdata.net)

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

60000

Spot measures 65 mm rain

50000 Centre logger

SE logger

NE logger 40000 Eel death event

30000

20000

Conductivity@25C (µS/cm)

10000

0 4 2 1 3 2 1 3 1 8 2 -D 3 4 -M 2 1 0 9 -J 7 -J -M -J -A -S -N a -F e a a u n e c n a y n u e o - b -0 - r -0 g p v 0 5 0 -0 6 -0 -0 -0 -0 7 -0 6 6 6 6 6 6 7

Figure 6: Electrical conductivity in Lake Tooliorook, December 2005 — April 2007 (Monitoring data from www.vicwaterdata.net)

100000 Fish death event Central logger 90000 SE logger

North logger 80000 Spot measures 70000 East logger 65 mm rain

60000

50000

40000

30000 High eel net Conductivity@25C (µS/cm) Conductivity@25C mortality >100,000 20000 begins eels die <100 golden 10000 perch die

2 9 2 1 8 2 1 9 2 0 -N 9 9 -J 8 7 -M 9 -O -N -D a -J -F -M o n a e a c v o e - n b r ar- t-06 -0 v c 0 - -0 6 -0 -0 7 0 -0 7 0 6 6 7 7 7

Figure 7: Electrical conductivity in Lake Bolac, October 2006 — March 2007

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

100000 SE logger NE logger 90000 Eel death event 80000 Spot measures

70000

60000

50000

40000

30000

Conductivity@25C (µS/cm) 20000 65 mm rain 10000

0 2 9 2 1 8 2 1 9 2 0 -N 9 9 -J 8 7 -M 9 -O -N -D a -J -F -M o n a e a c v o e - n b r a t-06 -0 v c 0 - -0 r 6 -0 -0 7 0 -0 7 -0 6 6 7 7 7

Figure 8: Electrical conductivity in Lake Tooliorook October 2006 — March 2007.

100000 East logger SE logger 90000 Spot measures

80000

70000

60000

50000

40000

30000

(µS/cm) Conductivity@25C 28 mm rain

20000

10000

2 9 2 1 8 2 1 9 2 0 -N 9 9 -J 8 7 -M 9 -O -N -D a -J -F -M o n a e a c v o e - n b r a t- -0 v c 0 - -0 r 0 6 -0 -0 7 0 -0 7 -0 6 6 6 7 7 7

Figure 9: Electrical conductivity in Lake Modewarre, October 2006 — March 2007.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

30000 East logger West Logger 25000 Spot measures

Carp spawning

20000

15000

17 mm rain 10000

Conductivity@25C (µS/cm)

5000

45 mm rain

0 20 9-No 29 19 8-Jan-07 28 17 9-Ma 29 18 -O -Nov -Dec -Jan-07 -Feb -M -Apr-07 c v r a t- -06 -07 r 0 - - - - 6 06 06 07 07

Figure 10: Electrical conductivity in Lake Colac, October 2006 — April 2007 (Different scales used for Lake Colac compared with other lakes)

<100 golden High eel net >100,000 perch die mortality eels die begins 2500 Spot measures Fish death event 2000 Logger daily average

1500

1000 Turbidity (NTU) Turbidity

500

2 9 2 1 8 2 1 9 2 0 -N 9 9 -J 8 7 -M 9 -O -N -D a -J -F -M o n a e a c v o e - n b r a t- -0 v c 0 - -0 r 0 6 -0 -0 7 0 -0 7 -0 6 6 6 7 7 7

Figure 11: Turbidity in Lake Bolac, October 2006 — March 2007

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

3000

Spot measures

2500 Logger daily average

2000

1500

(NTU) Turbidity 1000

500

20-Oct-06 9-No 29-Nov-06 19-De 8-Jan-07 28-J 17-Feb-07 9-Ma 29-Mar-07 18-Ap

v-0 an-07 r c- -07 r-07 6 0 6

Figure 12: Turbidity in Lake Colac, October 2006 — April 2007

Salinity • 20 to 35 per cent at Lake Modewarre in January 2005 and January/February 2006 Electrical conductivity (a measure of salinity) increased at least twofold in all the monitored lakes • 35 per cent at Lake Bolac in January 2006. over the course of the 2006—07 summer monitoring Turbidity period (Figures 7 to 10). Although groundwater and surface run-off may add a small amount of salt to the Turbidity was much higher in Lake Colac and Lake lakes studied, a doubling of the conductivity equates Bolac, which typically greatly exceeded 100 NTU roughly to a halving of the volume of the lakes. (nephelometric turbidity units), compared with Lake Modewarre and Lake Tooliorook, which were normally The greatest increase in electrical conductivity was at less than 20 NTU (Figures 11 and 12). The turbidity of Lake Bolac, which exhibited a fivefold increase lakes Modewarre and Tooliorook are not shown, as between spring 2006 and autumn 2007. The electrical these lakes display similar patterns to lakes Colac and conductivity recorded in the lake was unprecedented Bolac respectively. in 30 years of water quality monitoring (VWQMN 2007) Although turbidity was high at lakes Bolac and Colac, both generally decreased through the monitoring In all the lakes studied, the increase in the summer of period. This was particularly apparent at Lake Bolac, 2006—07 was greater than previously recorded where average turbidity decreased from around 300— summer rises (see Figures 5 and 6, and Leahy et al 500 NTU in November and December 2006 to around 2007). This is likely to relate to the severe rainfall 100 NTU in February 2007. deficiency for much of south-west Victoria (Figure 4) and the increasing rate of evaporation in shrinking The decline in turbidity in Lake Colac was not as great, water bodies, which heat up more readily. however, with the exception of periods of high turbidity in February and March 2007, average Figures 7 and 8 show that the eel deaths at Lake turbidity was clearly lower towards the end of the Tooliorook and Lake Bolac were associated with a monitoring. In both Lake Bolac and Colac, declining period of increasing salinity. In the four weeks turbidity appears inversely related to rising salinity, preceding the eel deaths. the salinity rose as follows: which is commonly seen in saline waters (Gregory • 40 per cent at Lake Bolac in March 2007 2006) (Figures 11 and 12). • 20 per cent at Lake Tooliorook in November 2006

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

pH both Lake Modewarre and Lake Tooliorook. A likely explanation for the high pH in Lake Modewarre and The pH in all four lakes was extremely high, especially Lake Tooliorook is the abundant seagrass, which is at Lake Modewarre and Lake Tooliorook (Figures 13 to likely to increase pH through high productivity. 16). In Lake Modewarre, pH was usually between 9.5 to 10.5, and exceeded 10 for much of the monitoring Although pH was lower in Lakes Bolac and Colac, period. In Lake Tooliorook, logger and spot generally between 8 and 9, pH rose in both lakes by measurements taken around the lake on the same day over 1 pH unit within a two-month period during showed a difference of up to 1 pH unit at different summer. In Lake Bolac, pH peaked at 9.86 in mid- locations in the lake. This suggests a lack of mixing in February 2007, and in Lake Colac pH peaked at 9.6 in the lake, probably because the sea grass, Ruppia sp., February and March 2007. At Lake Bolac, eel mortality forms dense beds, restricting water movement in the in fishing nets rose following this high pH. lake. The pH declined through the monitoring period at

10.5

Fish death event Central logger

SE logger 10 Spot measures North logger East logger 9.5

8.5 >100,000 High eel net eels die mortality begins <100 golden perch die 8 2 9 29-No 19-Dec-06 8-Jan-07 2 1 9 2 0 - 8-Jan-07 7-F -M 9-M -Oc Nov ar e a t- -0 v b-07 -07 r-0 06 -06 6 7

Figure 13: pH in Lake Bolac, October 2006 — March 2007

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

10.5

SE logger

NE logger

Eel death event 10.0 Spot measures

9.5

pH 9.0

8.5

8.0

2 9 2 1 8 2 1 9 2 0 -N 9 9 -J 8 7 -M 9 -O -N -D a -J -F -M o n a e a c v o e -07 n b r a t-0 -06 v c - -07 r -0 -06 07 -07 -07 6 6

Figure 14: pH in Lake Tooliorook, October 2006 — March 2007

10.5

10.0

9.5

9.0

East logger

8.5 SE logger

Spot m easures

8.0 2 9 29- 19-Dec-06 8-Jan-07 2 1 9 29-Mar-07 0- -Nov- 8- 7- -M Oct N Jan-0 F o eb-07 ar- - 0 v- 07 06 6 06 7 Figure 15: pH in Lake Modewarre, October 2006 — March 2007.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

10.5

Carp spawning

East logger 10.0 West logger

Spot m easures

9.5

9.0

8.5 17 mm rain

8.0 45 mm rain

7.5 20- 9-N 29-Nov 1 8-Jan-07 28-Jan-0 17-Feb-07 9 2 18- 9- -Mar-07 9- Oct-0 o D Mar- A v- ec-06 pr-07 06 -06 0 6 7 7

Figure 16: pH in Lake Colac, October 2006 — April 2007

Dissolved oxygen rotting eels and seagrass presumably acted to increase oxygen demand in the lake. Areas of stagnant Figures 17 and 18 show dissolved oxygen data for lakes anoxic water with purple sulphur bacteria were Bolac and Tooliorook. Dissolved oxygen data for Colac observed in patches in the lake. and Modewarre are not presented, as these lakes have dissolved oxygen patterns similar to Bolac and At Lake Bolac dissolved oxygen declined before the eel Tooliorook, respectively. deaths in March 2007. This is most apparent in the daily minimum dissolved oxygen, which declined in Hourly monitoring data has been presented as the February 2007. High mortality was seen in nets at this daily average, maximum and minimum of two loggers time, indicating netted eels were subjected to multiple in each lake. Inspection of the dissolved oxygen stressors. In the week before the widespread deaths, monitoring data showed diurnal patterns in all lakes the daily minimum dissolved oxygen was close to zero. (not shown), with the daily maximum typically Minimum dissolved oxygen recovered to 20 to 50 per corresponding to a mid-afternoon peak and the cent when eels were dying. minimum to an early-morning trough. It is understood that eels can tolerate low and In the first half of the monitoring period, dissolved temporarily anoxic conditions (McKinnon 2006) and so oxygen fluctuation was greatest in the two low dissolved oxygen is not considered to be the sole macrophyte-dominated lakes, Modewarre and cause of the eel deaths. However, the low oxygen prior Tooliorook, ranging between 50 and 250 per cent to the deaths may have exacerbated the effect of saturation (see Figure 18 for Lake Tooliorook results). other stressors in the lake such as high pH and salinity. In Colac and Bolac, dissolved oxygen was much more variable in the second half of the monitoring period, which relates to greater algal activity during this period, ranging between close to zero and 350 per cent saturation (see Figure 17 for Lake Bolac results). Algal concentrations were higher in these two lakes in March and April 2007 (Figure 23). At Lake Tooliorook, minimum (that is, overnight) dissolved oxygen dropped substantially following the eel deaths. The combined effect of a large biomass of

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

High eel net mortality >100,000 <100 golden eels die perch die begins 350 Fish death event Spot measures 300 Logger daily mininum Logger daily average Logger daily maximum 250

200

150

Dissolved oxygen % oxygen Dissolved

100

0 1 3 1 9 2 1 7 2 1 0 0 9 -D 9 8 -F 7 9 -O -O -N -D -J e -F -M e a b e c c o c e n b a t- t- v -0 c - -0 r 0 0 -0 6 -0 0 7 -0 -0 6 6 6 6 7 7 7

Figure 17: Dissolved oxygen (%) in Lake Bolac, October 2006 — March 2007

300

Eel death event Spot 250 measures Logger daily average

200 Logger daily minimum Logger daily maximum 150

Dissolved oxygen(%) 100

2 9 2 1 8 2 1 9 2 0 -N 9 9 -J 8 7 -M 9 -Oc -N -D a -J -F -M o n a e a v o e - n b r- a t- -0 v c 0 -0 0 r- 0 6 -0 -0 7 -0 7 0 6 6 6 7 7 7

Figure 18: Dissolved oxygen (%) in Lake Tooliorook, October 2006 — March 2007

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Table 2: Water quality data coincident with blood plasma sampling

Water body Lake Bolac Lake Tooliorook Lake Colac Belfast Lough Lake Bolac Lake Bolac Date Nov-06 Dec-06 Dec-06 Feb-07 Feb-07 Mar-07 Status Pre-kill Kill No kill No kill Pre-kill Kill Total alkalinity (mg/L) 444 945 507 163 290 407 Sodium (Na+) (mol/L) 0.161 0.220 0.074 0.511 0.387 0.681 Magnesium (Mg2+) (mol/L) 0.045 0.051 0.012 0.062 0.099 0.180 2+ Calcium (Ca ) (mol/L) 0.006 0.001 0.003 0.011 0.008 0.013 Potassium (K+) (mol/L) 0.002 0.002 0.001 0.012 0.003 0.006 Chloride (Cl-) (mol/L) 0.243 0.367 0.086 0.635 0.564 1.034 2- Sulfate (SO4 ) (mol/L) 0.007 0.006 0.001 0.032 0.019 0.028 2- Carbonate (HCO3 ) (mol/L) 0.000 0.006 0.001 0.000 0.001 0.000 - HCO3 (mol/L) 0.009 0.006 0.009 0.003 0.003 0.008 Salinity* (ppt) 14.92 20.33 5.82 39.92 33.76 60.67 Dissolved oxygen % 76.6 83.0 104.6 61.6 90.0 108.5 pH 8.5 9.4 8.7 8.1 9.1 8.3 Temperature (oC) 17.6 19.3 20.1 20.0 21.4 20.7 Turbidity (NTU) 981 1 206 6 104 113 * calculated from the sum of Na, K, Mg, Ca, Cl, SO4 and HCO3

Table 3: Proportion of cations in lakes Bolac, Colac, Modewarre and Tooliorook and Belfast Lough in comparison with seawater

Sodium % Magnesium % Calcium % Potassium % Waterbody n Mean Range CV Mean Range CV Mean Range CV Mean Range CV Seawater 1 86.44 - - 9.78 - - 1.90 - - 1.88 - - Lake Modewarre 12 91.25 89.17-92.61 1.13 8.20 6.86-10.12 11.73 0.17 0.13-0.21 17.95 0.38 0.27-0.50 18.25 Belfast Lough 4 85.73 85.19-86.19 0.58 0.00 - - 1.93 1.85-2.02 3.59 1.98 1.86-2.28 10.06 Lake Colac 12 83.51 75.74-85.43 3.19 12.93 11.39-21.49 21.38 2.25 1.60-3.28 26.41 1.31 1.14-1.54 9.98 Lake Tooliorook 8 80.55 79.71-81.80 0.96 18.57 17.31-19.52 4.10 0.29 0.21-0.42 27.72 0.59 0.49-0.65 10.66 Lake Bolac 15 77.30 75.33-82.03 2.16 20.17 15.30-22.80 8.13 1.87 1.43-2.69 21.14 0.66 0.44-0.73 11.75

CV= Coefficient of Variance

Seawater values derived from Atkinson and Bingman (1998)

Water conditions in Belfast Lough saturation This is not high, but was sufficient for fish survival. In-situ water quality and major ions were assessed at Belfast Lough, Port Fairy in February 2007, during Major ion analysis of water netting for eels. In situ measurements were collected at two locations in the estuary on 14 and 16 February Analysis of major ion proportions at lakes Bolac, Colac, 2007. Conditions were similar at both locations on Modewarre and Tooliorook showed that the ionic both days and average in-situ water quality composition, calculated from molar concentrations, measurements and major ion concentrations are remained relatively stable throughout the 2006—07 presented in Table 2. sampling period (Table 3). Differences in average + The pH and electrical conductivity suggest Belfast composition of the cations sodium (Na ), calcium 2+ 2+ + Lough is similar in composition to seawater. Salinity at (Ca ), magnesium (Mg ) and potassium (K ), in Belfast Lough was around 40 ppt or 58,000 μs/cm, comparison to seawater composition as reported by which is higher than the salinity of Lake Tooliorook at Atkinson and Bingman (1998), are also shown in Table the time of the deaths, but lower than Lake Bolac 3. during deaths there. Samples collected in February 2007 showed that the Electrical conductivity varied by less than 10 per cent ionic composition of Belfast Lough water is in the different sampling events. The pH in Belfast comparable to that of seawater (Table 3). The western Lough was around 8, and generally much lower than lakes studied showed similar proportions of sodium in the values recorded for the four lakes. Dissolved comparison with seawater, with most variation shown oxygen in Belfast Lough was around 60 per cent in proportions of magnesium, potassium and calcium.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Lake Tooliorook and Lake Bolac have significantly Eel maturity and sexual development higher proportions of magnesium, with their recorded On the basis of the silvering index of Van den Thillart ranges clearly higher than that reported for seawater. et al (2005), five stages were identified in the 285 eels This is indicative of the location of these lakes within sampled at Lake Colac, Lake Bolac, Lake Tooliorook volcanic plains. Lake Colac also has a greater and Belfast Lough (Figure 19 and Table 4) The stages proportion of magnesium in comparison to seawater, identified were: undifferentiated stage 1 eels; female but not to the same degree as Bolac and Tooliorook. stage 2, 3 and 4 (FII, FIII and FIV) eels; and male stage The proportion of potassium in the western lakes 2 (MII) eels. studied is much lower than ranges reported for Undifferentiated and FII eels are considered resident seawater or Belfast Lough. eels (immature growth phase), FIII are pre-migrant Significantly lower proportions of calcium in eels and FIV are migrant eels. Only one male was Tooliorook and Modewarre in comparison to the other identified from the silvering index and was not water bodies studied may be indicative of their considered in further analysis. macrophyte dominance. Calcium is a micronutrient Table 4 shows the mean and 95 per cent confidence required only by some algal species, whereas it is a limits for the morphological variables measured for universal requirement of macrophytes and other undifferentiated and female eels, grouped by the higher order plants for normal metabolic processes stages identified in the silvering index. (Wetzel 2001). The silvering index was found to categorise eels into The western lakes studied showed no significant maturity stages with distinctive ranges in length, changes in ionic composition during the 2006—07 weight, eye diameter and pectoral fin length. The summer. This suggests that change in proportions of mean length and weight of migrant eels as determined ions was not a mechanism for the eel deaths. by the silvering index in this study are comparable to values obtained for A. australis classified as migrant by De Silva et al (2002).

100%

90%

80%

70%

60%

50%

40%

Relative abundance 30%

20%

10%

0% Lake Lake Lake Bolac Lake Bolac Lake Bolac Lake Bolac Lake Lake Belfast Colac Oct Colac Dec 1 Nov 30 Nov Feb 2007 Mar 2007 Tooliorook Tooliorook Lough Feb 2006 2006 2006 2006 Nov 2006 Dec 2006 2007 stage I- resident stage II- resident stage III- pre-migrant stage IV- migrant

Figure 19: Relative abundance of silvering stages for eels sampled at lakes Colac, Bolac and Tooliorook and Belfast Lough.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Table 4: Mean and 95% confidence limit (CI) of morphological parameters measured in A. australis for each developmental stage.

Resident Pre-migratory Migratory Stage 1 n=7 Stage FII n=131 (39) Stage FIII n=136 (25) Stage FIV n=11 (7) Stage MII n=1 95% CI Mean 95% CI Mean 95% CI Mean 95% CI Mean value Length (mm) 417.0-469.0 443.0 653.1-679.1 666.1 759.6-777.5 768.5 840.2-896.2 868.2 541.0 Weight (g) 93.9-158.4 126.1 553.9-631.3 592.6 849.5-913.7 881.6 1322.0-1542.6 1432.3 398.0 Eye diameter (mm) 3.0-3.8 3.4 5.04-5.3 5.3 6.9-7.2 7.0 7.90-8.70 8.3 7.6 Pectoral fin length (mm) 13.9-16.7 15.3 28.1-29.8 28.9 38.9-40.6 39.8 43.1-49.9 46.5 34.9 Gonad weight (gm) - - 2.9-5.0 4.0 7.4-12.9 10.1 20.8-27.7 24.2 0.6 GSI% - - 0.5-0.8 0.7 0.8-1.3 1.0 1.4-1.8 1.6 -

Note: ( ) indicates smaller sample size for gonad weights

Table 5: Output of regression of blood sodium concentration and mean eye diameter of eels sampled at Lake Bolac, Lake Colac, Lake Tooliorook and Belfast Lough, 2006—07.

Sampling event n R2 value P value

Lake Bolac November 2006 21 0.14 0.09 Lake Tooliorook December 2006 18 0.01 0.71 Lake Colac December 2006 21 0.16 0.07 Belfast Lough February 2007 15 0.27 0.05 Lake Bolac February 2007 14 0.00 0.88 Lake Bolac March 2007 21 0.28 0.01 Lake Bolac March 2007- Female Stage 3 Eels 15 0.07 0.36

Note: Statistically significant values (P< 0.05) are shaded

Further support for the applicability of the silvering chloride cell identified as the mechanism for this index to A. australis is seen in the gonad development change. (Sasai 1998). of the different stages. Gonad development At each sampling event for Lake Bolac, Lake Colac, (calculated from the gonadosomatic index, GSI = Lake Tooliorook and Belfast Lough, eels sampled [gonad weight/bodyweight] x 100), increased with showed variability in their maturity as determined by each class identified by the silvering index (Table 4). Van den Thillart’s (2005) silvering index (Figure 19). Mean GSI of females increased from 0.67 per cent Van den Thillart (2005) found that, from the four ±0.16 (±95% CI) in resident stage II eels, to 1.03 per variables used in this model, mean eye diameter was cent ±0.24 in pre-migrant stage III eels, and to 1.62 per found to be the most important explanatory variable, cent ±0.19 in the migrant stage IV eels. These values followed by fin length and weight. Body length was the are similar to those obtained by Van den Thillart et al least significant variable. Therefore mean eye (2005), who recorded a mean GSI of 0.54 per cent for diameter was used as a representative measure of stage II females, 0.8 per cent for stage III females and maturity in a regression of blood sodium 1.5 per cent for stage IV females. concentration with eye diameter, to explore the These results suggest that the use of the silvering possible relationship of maturity on the blood plasma index (Van den Thillart et al 2005), although based and osmoregulatory ability of the eels sampled at each upon another species of eel, was appropriate to sampling event. determine the maturity of A. australis collected during The regression results (R2) are shown in Table 5. this study. Mean eye diameter was found to have no significant Eel blood plasma ion analysis in western Victoria effect (p >0.05) on blood sodium concentration of eels in all but those eels sampled during the eel deaths in Treatment of data Lake Bolac in March 2007 (p=0.01, n =21). It has been suggested that the osmoregulatory ability Three maturity classes were identified in the sample at of eels increases with morphological change and Lake Bolac in March 2007 (stage II females n= 5, stage sexual development, with changes in the type of III females n= 15, stage IV females n= 1), but the sample sizes for all but the stage III females were small.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Regression of mean eye diameter and blood sodium (Figure 20). Axis 1 represents a blood ion concentration of these stage III female eels showed no concentration gradient, with the position of the significant relationship (p=0.36). clusters along the axis due primarily to the Regression suggests that maturity has no effect on concentration of blood sodium and chloride, which the concentration of blood sodium of eels sampled increase in later sampling events at Lake Bolac and during 2006—07. Although a statistically significant Lake Tooliorook. relationship was apparent in eels sampled at Lake Eels sampled during the death event at Lake Bolac in Bolac in March 2007, the sample sizes of the smallest March 2007 form a cluster clearly separated from and largest eels were too small to make further other samples, signifying the extreme concentrations conclusions. Blood ion concentrations were therefore of ions in the blood of these eels. examined as averages by sampling event for further The sodium/potassium ratio is positively correlated analyses. No distinction between maturity stages was with axis 2, while potassium is negatively correlated to made. this axis. Samples collected at Lake Tooliorook during the death event in November/December 2006 are Comparison of plasma ion concentrations between lakes positively associated with axis 2, reflecting a higher Principle component analysis was performed on the sodium/potassium ratio (generally) in comparison with blood concentrations of sodium, chloride, potassium, other samples and have a greater dispersion across magnesium, calcium and sodium/potassium ratio of 110 axis 2. This plot shows that samples for each sampling eels from four waterways (Figure 20). event display clear groupings on the ordination plot, With the exception of the sodium/potassium ratio and indicating the blood ion concentrations of eels for potassium concentration, all blood ion variables are each sampling event are relatively similar. positively correlated with axis 1 of the ordination plot

Figure 20: Ordination plot of blood ion concentrations of 110 eels sampled at Lake Bolac, Lake Colac, Lake Tooliorook and Belfast Lough 2006—07

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Comparison of eel blood plasma ion values in western Metals analysis at Lake Bolac Victoria with published literature Sampling in January 2006 for metals returned high Plasma values from studies of eels in freshwater, results, particularly for aluminium and selenium. High those adapting to seawater (in ‘sea-shock’) and those levels of suspended sediment during sampling in adapted to seawater are compared with the results of January 2006 prevented filtration of the samples. this study in Figures 21 and 22. Filtered samples are usually used to determine if the They show that mean sodium values for all western metal is bioavailable and hence a risk to organisms. In Victorian locations in this study fall outside the range other lakes, filtered samples were below guidelines, that would be expected by linear interpolation even if unfiltered (total) results were over guidelines between upper literature values for freshwater and (Leahy et al 2007). seawater-adapted eels (Figure 21). This is even Metals sampling was repeated at Lake Bolac in March apparent in the mean sodium plasma concentration for 2007 during eel deaths, on unfiltered and filtered the lowest water salinity sample from Lake Colac. samples. The results of this analysis and guidelines This suggests that either A. australis has relatively from ANZECC–ARMCANZ (2000) are presented in higher plasma sodium concentrations or that there are Table 6. unmeasured differences between samples collected in The hardness of water (calcium and magnesium) the field and laboratory conditions. considerably reduces the toxicity of many metals, and The fact that all non-death event samples fall on the the guidelines are elevated for higher hardness, line of best fit, as determined for eels sampled in the following ANZECC-ARMCANZ (2000). A hardness of western lakes during this study, is suggestive of a 840 mg/L was used, as this is the minimum hardness systematic difference between A. australis and other recorded in Lake Bolac since 2006. The hardness eels. Analysis under laboratory conditions would be during the eel deaths was substantially higher (over required to further elucidate this. 2000 mg/L) and so the use of the guidelines is conservative. Comparing the average plasma sodium in the two eel death sampling events, only Lake Bolac samples With the exception of aluminium and zinc, all metals during the March 2007 death event are elevated in were below guideline levels during the Lake Bolac eel comparison to the straight-line increase for deaths (Table 6). Aluminium results were all below A. australis. The average plasma sodium for the Lake guideline values for filtered samples. This shows that Tooliorook deaths falls on the line of best fit for the aluminium is not likely to be bioavailable. The high A. australis non-death event samples. total aluminium is likely to be due to the clay minerals present in the water column, of which aluminium Mean chloride plasma values for the lower salinity forms a major component. sampling events (Lake Colac, December 2006 and Lake Bolac, November 2006) fall within the range of Total zinc concentrations were below guideline levels, linear increase between upper literature values (Figure but filtered zinc concentrations exceeded guidelines. 22). The most likely explanation for this is contamination during the filtration procedure, as filtered samples Analysis of inter-laboratory comparisons showed there should be lower than or similar to unfiltered values. In is some uncertainty in the absolute chloride values, any case, the filtered concentrations recorded are not but the relative difference between samples in this likely to cause acute fish deaths, particularly given study should be accurate. Based on the water salinity that the water hardness during the deaths (over during the deaths at Lake Bolac and Lake Tooliorook, 2000 mg/L) was much higher than the value used for average plasma chloride values clearly exceed what hardness correction (840 mg/L), which would further would be expected based on the linear increase for A. reduce the bioavailability of zinc (ANZECC-ARMCANZ australis non-death event values. 2000). This suggests that eels during the death events were The results of the metals testing at Lake Bolac suggest failing to regulate chloride (and sodium at Lake Bolac). that metals were not a factor in the eel deaths at Lake The chloride values recorded for A. australis during Bolac. This is consistent with the results of metals the death events show a similar degree of rise to the testing in the other lakes (Leahy et al 2007). sea-shock values recorded for A. anguilla (Kirsch 1972a) and A. dieffenbachii (Shuttleworth and Freeman Algal analysis 1973) (Figure 22). Literature sea-shock values refer to animals that have been subjected to osmotic shock The concentrations of algae recorded during the from being transferred directly from freshwater to monthly algal monitoring are presented in Figure 23. seawater. The results of algal analysis during deaths at Lake Tooliorook and Lake Bolac are presented in Table 7 and Table 8.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Figure 21: Mean sodium concentration in blood plasma plotted against salinity in this study and in published literature. (Error bars are ± standard error of measure — SEM)

Figure 22: Mean chloride concentration in blood plasma plotted against salinity in this study and in published literature. (Error bars are ± standard error of measure — SEM)

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Table 6: Metals testing in Lake Bolac during eel deaths, March 2007

Sample date 13/03/2007 13/03/2007 13/03/2007 13/03/2007 14/03/2007 14/03/2007 14/03/2007 ANZECC Guideline Location South South North South East South East East East *Adjusted to minimum Total/filtered Total Filtered Total Total Filtered Total Filtered hardness of 840 mg/L Limit of reporting Mercury 0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0.06 Aluminium 10 6400 20 2810 6160 10 1170 40 55 Selenium2 4344343 5 Arsenic 0.5 36.1 33.9 35.2 33.4 33 35.8 31.5 221 Cadmium 0.2 <0.2 0.3 <0.2 <0.2 <0.2 <0.2 <0.2 4* Chromium 0.5 13.1 <0.5 5.8 13 <0.5 2.2 <0.5 51* Copper 1 4 2 2 4 2 <1 2 248* Lead 0.2 4.5 0.4 2.3 4.6 0.3 0.8 0.4 234* Nickel 0.5 20.1 14.2 17.8 18.8 13.2 16.6 12.9 187* Zinc 5 9 62 8 12 44 6 41 41*

Note: Values exceeding guidelines are shaded. Filtered samples were passed through a 0.45 μm cellulose acetate filter to estimate metal bioavailability

1000000

100000

10000

1000

100

Bolac 10 Colac

(cells/mL) concentration phytoplankton Total Modewarre Tooliorook

7 7 -06 -06 -07 b-07 r-0 -0 ov ec e ar N -D -Jan -Ma -M 04- 24-Nov-06 14 03 23-Jan-07 12-F 04 24 13-Apr-07

Figure 23: Total phytoplankton concentration in four western lakes, November 2006 — April 2007. (The y-axis is a logarithmic scale)

Many of the species recorded in the four lakes are The monitoring of algae during the summer of 2006- characteristic of marine waters. During the Lake Bolac 07 did not detect a toxic algal species responsible for eel deaths, algae were abundant but had low diversity, the eel deaths. There is no evidence from EPA’s and were dominated by marine diatoms, particularly monitoring since 2004 that algal toxins were Cylindrotheca closterium (Table 7). In contrast, Lake responsible for the fish deaths in south-western Tooliorook samples had a diverse range of algae at low Victoria. abundance (Table 8).

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Table 7: Algae results for Lake Tooliorook collected during eel deaths

Lake Tooliorook Lake Tooliorook Lake Tooliorook South east West North

30-Nov-06 30-Nov-06 30-Nov-06

Estimated Estimated Estimated Cells/mL Cells/mL Cells/mL

Total Phytoplankton 98 279 23,375 Diatoms 8 14 832 Dinoflagellates 21562 Other flagellates 88 250 22,482

Diatoms Amphora sp. 11 Aulacoseira sp. 13 Chaetoceros spp. 1 Cocconeis spp. 595 Cymbella sp. 15 Cylindrotheca closterium 1 2 770 Diploneis sp. 5 Encyonema sp. 120 Gomphonema sp. 1 Melosira sp. 2 Naviculoid spp. 18 Nitzschia spp. 1118 Thalassiosira sp. 1

Dinoflagellates Gymnodinioid spp. 1 10 44 Gyrodinium spp. 143 Heterocapsa rotundata 18 Peridinium sp. 113 Prorocentrum sp. 3 Prorocentrum cordatum 113

Chrysophytes Calycomonas sp. 115 Ochromonas spp. 1385

Cryptophytes Chroomonas spp. 8 Hemiselmis sp. 138 Leucocryptos marina 1118 Plagioselmis prolonga 2515 Rhodomonas sp. 1 1 148

Chlorophytes Ankistrodesmus falcatus 1 Scenedesmus quadricaudus 1 cf. Kirchneriella 3um crescent 22,000

Prasinophytes Pyramimonas spp. 13 Tetraselmis spp. 1118

Euglenophyta Trachelomonas spp. 1

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Table 7 continued: Algae results for Lake Tooliorook collected during eel deaths

Lake Tooliorook Lake Tooliorook Lake Tooliorook

South east West North

30-Nov-06 30-Nov-06 30-Nov-06 Estimated Estimated Estimated Cells/mL Cells/mL Cells/mL

Cyanoprokaryota

Anabena sp. (filaments) 3105 Anabena sp. (cells) 61 179 125 Aphanizomenon sp. (filaments) 11 Aphanizomenon sp. (cells) 19 44

Cilliates Mesodinium rubrum 8 Unidentified ciliates 130

Other Unidentified amoeba 5 Unidentified bodonids 3

Unidentified heterotrophic flagellate 1

Table 8: Algae results for Lake Bolac collected during eel deaths

Lake Bolac Lake Bolac Lake Bolac South North East

13-Mar-07 13-Mar-07 13-Mar-07

Estimated Estimated Estimated Cells/mL Cells/mL Cells/mL

Total Phytoplankton 559,000 354,000 291,000 Total Diatoms 549,000 348,000 285,000

Diatoms Chaetoceros spp. 5,000 5,000 Cylindrotheca closterium 482,000 288,000 264,000 Minidiscus trioculatus 57,000 54,000 16,000 Nitzschia spp. 3,000 1,000 Pleurosigma sp. 7,000

Chlorophytes Ankistrodesmus sp. 10,000 6,000 6,000

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

4. DISCUSSION Failure to regulate blood plasma within narrow bounds has been found to be characteristic of eels experiencing ‘hypermineralisation’, also known as sea- Water quality conditions in the lakes shock (Kirsch 1972a; Shuttleworth and Freeman 1973). The four lakes fit into two categories: This sea-shock is seen in eels transferred from low- salinity freshwater to higher-salinity seawater. 1) algal-dominated This period has been shown to be temporary, with 2) aquatic macrophyte-dominated. plasma salinity stabilising, albeit at a moderately Lake Modewarre and Lake Tooliorook are higher level, once the eels adjust to seawater. The characterised by an abundance of seagrass (Ruppia blood plasma results for western Victorian eels in this sp.), low turbidity and high pH (typically greater than study, particularly those from the lakes experiencing 9—10.5). eel deaths, appear abnormal and look similar in Lake Colac and Lake Bolac are characterised by high magnitude to sea-shock eels from other studies algal concentrations, high turbidity and lower pH (Kirsch 1972a; Shuttleworth and Freeman 1973). (typically 8.5—9). The high plasma sodium and chloride results for dead The presence of seagrass is likely to play an important eel samples from Lake Bolac (March 2007) and Lake role in this difference, reducing wind-suspended Tooliorook (November/December 2006) are clear turbidity and increasing the pH of the two lakes to evidence of a failure to maintain plasma extreme values through high productivity. The high concentrations within normal bounds and are the likely productivity is likely to deplete acidic hydrogen ions, cause of the eel deaths. although the exact mechanism requires further The water salinity in the two lakes at the time of the investigation. deaths was very different. The salinity of Lake Bolac The rise in pH at Lake Bolac in the late summer— was 60 ppt (approaching twice that of seawater), autumn period suggests photosynthesis by algae in whereas Lake Tooliorook was 20 ppt (approaching the two lakes may be the cause of increased pH, two-thirds that of seawater), so clearly there is not a although this requires confirmation. Algal single threshold for all waterways. Individual factors in concentration was higher in both Lake Bolac and Colac each lake are likely to accentuate or suppress the during this period (Figure 23). effect of salinity. The most likely of these is lake water pH, but dissolved oxygen and others may also play a Salinity in all the monitored lakes was unprecedented role. As discussed in the background to fish in historical records. Salinity in Lake Tooliorook was osmoregulation, pH plays an important role in the over one-half that of seawater during the eel deaths functioning of fish gills. Wilkie and Wood (1996) have and, at Lake Bolac, it was nearly one and a half times shown that fish that can tolerate the conditions seen that of seawater. The salinity in Lake Colac is in the western Victorian lakes are rare and can only somewhat lower than the other lakes, but it doubled in survive due to specialised adaptations. the 2006—07 summer and a similar rise is expected to lead to fish deaths. The salinity in Lake Modewarre The role of pH in eel deaths in western Victorian continued to rise throughout the 2006—07 summer, lakes due to evaporation, with salinity rising fourfold between 2004 and 2007. Alkaline pH values that are toxic to fish are typically around 9 and above, but these are based on It is highly likely that similar changes occurred during freshwater conditions (Alabaster and Lloyd 1982). In past dry periods, but they went unrecorded, as did any freshwater, ammonia toxicity increases with fish deaths that occurred (EPA 2006). increasing pH, because the relative proportion of the toxic form (free unionised ammonia, NH ) is pH- Blood plasma in western Victorian eels 3 dependent. However, in more saline water, evidence In order to have normal physiological functioning, eels for increasing ammonia toxicity with increasing pH is need to maintain the salt content of their blood plasma still the focus of research and there is still much to be within a narrow range. discovered (Eddy 2005). In this study, A. australis was found to have higher Some studies have found that ammonia toxicity may plasma sodium than has been reported elsewhere in be greatest at mid-range salinities, similar to those the literature, even in waterways where eel deaths did levels in Lake Modewarre and Lake Tooliorook during not occur. However, plasma sodium (Lake Bolac March the eel deaths (Harader and Allen 1983). 2007) and chloride (Lake Tooliorook November- It is apparent from the data from the lakes studied December 2006 and Lake Bolac March 2007) were that eels died at lower salinities when pH was greater found to be elevated in the two waterways than 9. The difference between the two lakes with experiencing eel deaths, beyond even the background higher pH (Lake Modewarre and Lake Tooliorook) and elevation of A. australis plasma. those with lower pH (Lake Colac and Lake Bolac) is the

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

presence of the submerged aquatic macrophyte To keep the salinity under the tolerance thresholds for Ruppia. the fish species that exist in Lake Colac, the salinity needs to decrease to similar levels seen at the start of Potential for fish deaths in Lake Colac the 2006—07 summer. Because salinity is closely related to water volume, this effectively requires a Figure 24 shows a general model of the eel deaths in doubling of the water volume in Lake Colac before the Western District lakes. From this it can be seen that, at 2007—08 summer. Lake Modewarre and Lake Tooliorook, pH was high in the lakes and conditions became critical for eels when If rainfall remains low before the 2007—08 summer, it salinity rose by around 20 per cent per month. is predicted that Lake Colac will follow the pattern of Lake Bolac. Past patterns suggest European carp may At Lake Bolac, pH was lower but rose during summer, die early in the summer, due to rising salinity, but it is particularly once turbidity declined. With salinity rising likely that eels will die later in the second half of by more than 20—40 per cent per month and pH summer if pH has risen to around 9.5. reaching close to 9.5, a critical point was reached for eels. Climate change implications for fish in shallow Commercial fishermen estimate that there may be western Victorian lakes over 500 tonnes of fish in Lake Colac, including European carp, eels, redfin and a number of other Climate change models for south-western Victoria species (pers. comm., Western District Eel Growers predict lower rainfall, higher evaporation and greater Group). A sharp rise in salinity and pH in Lake Colac in variability (DSE 2004). The consequence of this for the 2006—07 summer took the lake close to the fish like eels with long lifespans will mean that their conditions that saw European carp die in Lake populations will become increasingly unviable in Modewarre in December 2004. Victoria’s shallow western lakes and deaths will become more frequent. The shallow western lakes may The salinity reached in Lake Colac in the summer of become unsustainable in the long term as commercial 2006—07 is unprecedented in 30 years of monitoring eel fisheries, because the time between dry periods the lake (see Leahy et al 2007). The potential for large may be insufficient for commercially viable growth and fish deaths in Lake Colac in the 2007—08 summer is harvest. heavily dependent on good rainfall in the intervening period. Fish with shorter life spans like galaxiids may be able to complete their reproductive cycle during wetter The salinity in Lake Colac at the beginning of summer phases and survive in the lakes. Stocking of 2006—07 was one-fifth that of seawater and had recreational species may also come under threat if doubled by April 2007. As a result of the increased they are not able to grow to a suitable size before salinity, turbidity decreased, causing increased pH. It is conditions become too extreme. assumed that this trend will continue in the summer of 2007—08 unless sufficient rainfall is received.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

a. Model of response in turbid algae dominated lakes (for example Lake Bolac and predicted trajectory for Lake Colac)) b. Model of response in clear macrophyte dominated lakes (for example. Lake Modewarre and Lake Tooliorook).

Figure 24: General model of eel deaths in western Victorian lakes

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

5. CONCLUSIONS The effect of drought on south-west Victoria in the last 10 years has given us an insight into the consequences EPA’s investigations during the summer of 2006—07 of climate change on Victoria’s Western District lakes. set out to answer remaining questions regarding the cause of eel deaths. With the exception of Lake Colac, the conditions in all the lakes were extremely 6. ACKNOWLEDGEMENTS unfavourable for the survival of any fish. By worldwide The authors thank the members of the Eel Death comparison, few fish are known to survive the Investigation Reference Group for their input into the conditions recorded in the lakes where deaths design of the investigation. occurred. The Eel Death Investigation Reference Group was Clear evidence of the importance of environmental chaired by Christine Forster and was comprised of stress on eel deaths has emerged. In particular, the Prof John Sherwood (Deakin University), Tom Ryan combined effect of high salinity and high pH has been (Environment Victoria), Bill Allan (commercial eel found to lead to abnormal eel blood plasma ion industry), Peter Appleford (DPI Fisheries Victoria), Dr concentrations. At Lake Modewarre, Lake Bolac and Mehdi Doroudi (DPI PIRVIC), Keith Jackson Lake Tooliorook, a pH of around 9—9.5 and a salinity (Department of Sustainability and Environment), Peter increase of over 20 per cent a month is the common Greig (Corangamite Catchment Management link in all deaths. Research over the 2006—07 summer Authority), Dr Heather Builth (Glenelg-Hopkins has shown the abnormal internal response of eels to Catchment Management Authority) and these conditions. representatives from DPI Attwood. While there are some remaining unanswered questions The investigation was greatly aided by Lachlan regarding the exact role of seagrass in elevated pH McKinnon (Audentes Pty Ltd) and Bill Allan and the and the cellular-level response of eels to high salinity members of the Western District Eel Growers Group. and pH, the patterns in eel deaths have been well established and follow predictable paths. The eel deaths since 2004 have caused large environmental, economic and social losses. The commercial value of the eels that died in the 2006—07 summer alone amounts to over $750,000. Unless water quality conditions improve before the 2007—08 summer and greater inflows occur, it is unlikely eels will be present in lakes Modewarre, Bolac and Tooliorook in the coming summer. Eels were still present in Lake Colac at the end of the 2006—07 summer, but they are likely to be one of the last large inland lake populations. There is potential for significant fish deaths in Lake Colac in the summer of 2008 if the current low rainfall level continues. The future condition of the western lakes clearly depends on a period of sustained, higher rainfall. If sufficient rainfall occurs and some of the salt is flushed from the lakes, then there is no reason the lakes cannot support a healthy fishery once again. Periodic drying and high salinity is a natural part of many lake systems in Australia. The plants and animals that inhabit the lakes should quickly recover if higher rainfall occurs. The greatest threat to the re-establishment of the Western District lakes fishery over the long term is likely to be climate change. These lakes tend to act like large rainfall gauges and sustained below-average rainfall and low run-off into lakes leads to low water levels and, eventually, drying out. The investigation into eel deaths has underlined the susceptibility of even hardy fish like eels to low rainfall conditions.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

7. REFERENCES Leahy PJ, Leonard A and Johnson R (2007) Findings of the western Victorian lakes eel death investigations Alabaster JS and Lloyd R (1982) Water quality criteria 2004-06. Science Report 1114. EPA Victoria. for freshwater fish. Second edition. London, FAO. Melbourne. http://epanote2.epa.vic.gov.au/4A256593001F8336/b ANZECC/ARMCANZ (Australian and New Zealand ycat Environment and Conservation Council/Agriculture and Resources Management Council of Australia and Lutz, (1972) Ionic patterns in the teleost. Comp. New Zealand) (2000) Australian and New Zealand Biochem. Physiol. 42A, 719—733 Guidelines for Fresh and Marine Water Quality: Volume Maetz J (1971) Fish gills: mechanisms of salt transfer in 1 The Guidelines. Paper No. 4. ANZECC/ARMCANZ, freshwater and seawater. Phil. Trans. Roy. Soc. Lond. Canberra. B. 262, 209—249 APHA (2005) Standard Methods for the examination of Martinez AS, Cutler CP, Wilson GD, Phillips C, Hazon N water and wastewater. Eaton AD, Clesceri LS, Rice EW and Cramb G (2005) Cloning and expression of three and Greenberg AE (Eds.). American Public Health aquaporin homologues from the European eel Association/American Water Works Association/Water (Anguilla anguilla): effects of seawater acclimation and Environment Association. Washington. cortisol treatment on renal expression. Biology of the Atkinson MJ and Bingman C (1997) Elemental Cell. 97, 615—627. composition of commercial sea salts. Journal of McCormick SD (2001) Endocrine control of Aquariculture and Aquatic Sciences. 8, 39—43. osmoregulation in teleost fish. American Zoologist. 41 Chan DKO, Chester Jones I, Henderson IW and Rankin (4) 781—794. JC (1967) Studies on the experimental alteration of McKenzie DJ, Shingles A and Taylor EW (2003) Sub- water and electrolyte composition of the eel (Anguilla lethal plasma ammonia accumulation and the exercise anguilla). J. Endocrin. 37, 297—317. performance of salmonids. Comparative Biochemistry De Silva SS, Gunasekera RM and Collins RO (2002) and Physiology 135, 515—526. Some morphometric and biochemical features of McKinnon LJ (2006). A Review of Eel Biology: ready-to-migrate silver and pre-migratory yellow Knowledge and Gaps. Report to EPA Victoria. 39pp. stages of the shortfin eel of south-eastern Australian Available from waters. Journal of Fish Biology 61, 915—928. www.epa.vic.gov.au/water/rivers/docs/Eel_Biology_Re DSE (Department of Sustainability and Environment) view.pdf (2004) Climate change in the Corangamite region. Olivereau M, Chambolle P and Dubourg P (1981) State Government of Victoria. Ultrastructural changes in the calcium-sensitive (PAS- Eddy FB (2005) Ammonia in estuaries and effects on positive) cells of the pars intermedia of eels kept in fish. Journal of Fish Biology. 67(6), 1495—1513. deionized water and in normal and concentrated seawater. Cell Tissue Res. 219, 9—26 EPA Victoria (2006) A review of historic western Victorian lake conditions in relation to fish deaths. Pease BC (ed.) (2004) Description of the biology and Publication 1108. Available from assessment of the fishery for adult longfinned eels in www.epa.vic.gov.au/water/rivers/eel_deaths.asp NSW. Final report to Fisheries Research and Development Corporation, Project No. 1998/127 168 Gregory J (2006) Particles in Water: Properties and pp. Processes. CRC Press. Perez-Ruzafa A, Quispe-Becerra JI, Garcia-Charton JA Harader RR and Allen GH (1983). Ammonia toxicity to and Marcos C (2004) Composition, structure and Chinook salmon parr: reduction in saline water. distribution of the ichthyoplankton in a Mediterranean Transactions of the American Fisheries Society. 112, coastal lagoon. Journal of Fish Biology. 64, 202—218. 834—837. Price RC, Gray CM and Frey FA (1997) Stontium Jobling M (1995) The environmental biology of fishes. isotopic and trace element heterogeneity in the plains Chapman and Hall. 455 pp. basalts of the Newer Volcanic Province, Victoria, Karnaky KJ, Ernst SA, Philpott CW (1976) Teleost Australia. Geochimica et Cosmochimica Acta. 61(1), 171— Chloride Cell. I. Response of Pupfish Cyprinodon 192. variegatus Gill Na/K-ATPase and Chloride Cell Fine Randall DJ and Wright PA (1987) Ammonia distribution Structure to Various High Salinity Environments. The and excretion in fish. Fish Physiology and Journal of Cell Biology. 70(1) 144—156. Biochemistry. 3(3), 107—120. Kirsch R (1972a) Plasma chloride and sodium, and Randall DJ, Wood CM, Perry SF, Bergman H, Maloiy chloride space in the European eel, Anguilla anguilla. J. GMO, Mommsen TP and Wright PA (1989) Urea Exp. Biol. 57, 113—131. excretion as a strategy for survival in a fish living in a very alkaline environment. Nature. 337, 165—166.

SCIENTIFIC INVESTIGATION INTO EEL DEATHS IN WESTERN VICTORIA

Rossi R and Cannas A (1984) Eel fishing management in a hypersaline lagoon of southern Sardinia Fisheries Research 2(44), 285—298. Sasai S, Kaneko HS and Tsukamoto K (1998) Morphological alteration in two types of gill chloride cells in Japanese eels (Anguilla japonica) during catadromous migration. Canadian Journal of Zoology 76(8), 1480—1487. Shuttleworth TJ and Freeman RFH (1973) The role of Gills in Seawater Adaptation in Anguilla dieffenbachii 1. Osmotic and Ionic Composition of the Blood and Gill Tissue. J. comp. Physiol. 86: 293—313. Tesch FW (updated by Kloppmann M) (2003) Body structure and function pp 1-72 In Tesch FW The eel. 5th Edition. Thrope, JE (Ed.) (Translated to English by White RJ). Blackwell Science. Van den Thillart G, Dufour S, Elie P, Vockaert F, Sebert P, Rankin C, Szekely C and Rijsingen I (2005) Estimation of the reproduction capacity of European eel. Report to European Union 272pp. Available from http://www.fishbiology.net/EELREP_final_report.pdf Victorian Water Quality Monitoring Network www.vicwaterdata.net Wetzel RG (2001). Limnology: Lake and River Ecosystems. Academic Press, San Diego. 1066 pp. Wilkie MP and Wood CM (1996) The Adaptations of Fish to Extremely Alkaline Environments. Comparative Biochemistry and Physiology Part B. 113(4) 665—673. Wilson R, Wright P, Munger S and Wood C (1994) Ammonia excretion in freshwater rainbow trout (Oncorhynchus mykiss) and the importance of gill boundary layer acidification: lack of evidence for + + Na /NH4 exchange. Journal of Experimental Biology, 191(1), 37—58.