Environmental and biological aspects of the mass mortality of Pilchards (Autumn 1995) in Western Australia

Executive Summary

The first pilchard deaths were reported in South Australia during March 1995 in the eastern region of the Great Australian Bight. Subsequently, dead pilchards were found both east and west of this point source moving in a ‘bushfire-like' front at an average of 30 km/day. By early May the fronts had reached Albany in Western Australia (WA) and Bass Strait in Victoria. They continued moving up both coasts, reaching Carnarvon (WA) and Noosa Heads (Queensland) by the end of June, thereby affecting pilchards throughout their entire Australian distribution (6700 km). In early June, a similar pattern of deaths began to occur in the north of New Zealand.

The pattern of deaths at all locations was similar with only adult pilchards (> 10 cm, 13 cm in WA) affected. No other species or even juvenile pilchards were found dead; moreover, neither predators or scavengers died as a result of their consumption. Fatalities lasted for only a few days at any one location but the intensity did not appear to diminish with time or distance from the origin. Subsequently, no further deaths were observed.

Originally it was thought an upwelling event that occurred in South Australia during early March could have been the cause by lowering water temperatures and/or giving rise to a toxic phytoplankton bloom. The list of possibilities was later expanded to include clogging of gills by non-toxic phytoplankton, the impact of stress from altered environments, and finally the possibility that the deaths were caused by some exotic pathogen. To examine these theories numerous biological and environmental samples were collected, particularly of dead and dying fish, along with healthy fish taken before the front and unaffected fish behind the front. Samples of phytoplankton and environmental measurements were also obtained. Similar investigations were conducted at many institutions around the country with a Pilchard Mortality Task Force created under the auspices of the Consultative Committee for Exotic Animal Diseases (CCEAD) which organised a number of teleconferences to help disseminate information and coordinate activities.

Affected pilchards, aside from being dead, were in good condition with many having advanced gonad stages. Thus, in Esperance, significant levels of egg production were found in an area where many fish were dying. The dead fish usually had open mouths and gills that were pale in colour. Histopathology showed that the gills had epithelial hyperplasia (multiplication of the surface cells), synechiae of the secondary gill lamellae tips (they had stuck together), sloughing of epithelial cells and oedema. Thus, the cause of death was asphyxiation.

Identical patterns of damage were seen at all locations where the deaths occurred indicating a common cause. Consequently, unless a single theory could explain the deaths in all areas, it was unlikely to be correct. Possible causes of the gill damage include contact with toxic and non-toxic phytoplankton. It was initially suggested (incorrectly) that adult pilchards were the only phytoplankton feeder in the region and may therefore have been the only affected group. Alternatively, this damage could have arisen from the pathogenic effect of some specific virus or bacterium to which only adult pilchards were susceptible.

The data collected in WA clearly showed that the pilchard mortalities had no relationship with phytoplankton. Toxic phytoplankton were not involved in the deaths; few were seen in the phytoplankton samples and no toxic substances were found in any affected fish. There were no blooms of non-toxic phytoplankton anywhere along the WA coast during the times when the mortalities were occurring. The composition of the phytoplankton that was present varied greatly between sites independently of whether dead and dying pilchards were in the region. Most sites where deaths were recorded had very low densities of phytoplankton. Furthermore, examination of the affected gills by SEM showed no evidence of clogging or mechanical damage consistent with phytoplankton being responsible.

The stomach contents of the affected pilchards varied greatly. In addition, many had empty stomachs. The slight difference in feeding capabilities between adult and juvenile pilchards is insufficient to support the total lack of impact on juveniles. Similarly, the lack of an effect seen on the other coexistent, filter-feeding clupeids (e.g. anchovies, Sardinella) is a further rejection of the hypothesis that this phenomenon was somehow caused by phytoplankton blooms. Finally, the passage of the front of dying fish moved in the opposite direction to the prevailing currents and continued even after several severe storms. None of these observations are consistent with phytoplankton being involved.

The gill damage seen in the dead pilchards collected in WA was always associated with the presence of a Herpesvirus. This virus was not present in fish sampled ahead of the deaths, and was not found in samples of survivors after mortalities stopped. Work at the CSIRO animal health laboratories at Geelong, the New South Wales Agriculture Department and the New Zealand National Institute of Water and Atmosphere (NIWA) also linked the presence of the virus with the deaths. An amoeba was associated with gills of many dead fish, but this was usually in insufficient numbers to account for the damage seen, and was not always present in fish with gill damage. Thus the virus was the only consistent factor in all of the kills in WA, Australia and New Zealand. Furthermore, the rate of passage of the fronts was within the limits of daily movement rates of adult pilchards

The interim task force report (Anon., 1995) reported that the virus could have been endemic and triggered by stress caused from changes in temperature or the onset of spawning. Neither of these ‘causes', however, have any support. Firstly, water temperatures in WA were not abnormal: they were mostly in the 18 - 21°C range, with no cold-water intrusions on to the shelf or subsurface currents of cold water as mentioned in some reports to support any ‘stress' theory. Secondly, as the spawning season for pilchards varies greatly between regions, affected fish were at all stages of the spawning cycle.

The more likely alternative hypothesis is that the virus was recently introduced into Australia. Whilst involvement of the Herpesvirus in the mortalities has not been conclusively demonstrated by infection trials, the association of the virus with the gill damage, the severity of the impact on the population and the bushfire-like passage of the front are all consistent with a novel pathogen infecting a naive population. Furthermore, once infection had passed an area, reinfection did not occur, suggesting that the surviving fish were resistant to infection. No direct evidence was obtained to determine the possible method for any introduction.

The impact of the deaths on the stocks of pilchards was also determined. Counts of dead pilchards were made both on the sea surface and the bottom at a number of locations in WA. In addition, biomass estimates were calculated using the daily egg production techniques which showed that the impact to the pilchard stocks in WA was the death of approximately 10 - 15%, with this still representing many thousands of tonnes. Nonetheless, all WA pilchard fisheries recommenced fishing a few weeks after the deaths had ceased and have since not registered any major unexpected changes in catch rates. No impacts have been seen in other species that utilise pilchards and consequently no longer term effects are envisaged.

The major conclusions drawn from the investigations conducted in WA are:

1. Pilchards were affected over their entire range.
2. Approximately 10 - 15% of the stocks were killed.
3. There was no involvement of phytoplankton either directly or indirectly in the deaths of the pilchards.
4. There were no large or small-scale environmental anomalies which could have affected the pilchards over any part of their range in WA.
5. The only consistent factor in the deaths at all locations was the presence of a previously undescribed Herpesvirus in the gills of the affected pilchards.
6. The pattern and severity of the impact suggest that the Herpesvirus was not a latent infection.

Thus:

Conclusion

The most likely cause of the massive mortalities of pilchards in Australia during early 1995 was from a novel Herpesvirus to which the Australian pilchard population was naive and whose origin was, therefore, most likely to be exotic.