Introduction A the end of August last year an intense microalgal bloom developed along the open coast at the Leigh Laboratory. This bloom was unusual in its intensity, seasonal timing, duration, geographic extent, composition and subsequent biological events. This seminar attempts to review the data about this bloom and discuss implications. It is unusual in several respects:

1. It reports on the work of many people including:
• Allison Haywood, MSc Environmental Sciences: studies of particles in open water
• Leslie Rhodes, Cawthron Institute: identification and counts of microalgae
• ARC and others: water quality sampling
• Jo Evans and others: local climate data
• Meteorological services: satellite derived data and climate indices
• Many biologists at Leigh: anomalous biological data since Sept 1992
• Fishermen and divers: reports of seawater discoloration and low underwater visibility.
2. There was no formal project. No one had the task of investigating the bloom as such, or of assigning anyone to do so. Those who involved themselves all had different aims and interests, and none included a direct study of the bloom.
3. It raises not only the standard scientific questions on the bloom itself - what was its nature, causes and consequences - but also wider questions such as:
• what is the proper scientific attitude to rare but significant events in your general area of interest, but outside your current range of concern?
• what are the proper methods for combining a wide range of data which vary widely in reliability, and scales of time, space and complexity?
• are there any general principles which would help, or are we obliged to react each time without any guidelines?
4. These questions are made more pointed because of:
• the current policies for 'user pays' and competitive research
• the contrast between professional reactions to the algal blooms of 1992 and the toxic shellfish issues of 1993.

Evidence from further afield

• Aerial survey by Alison Haywood of the western Hauraki Gulf, September 13th
• Counts and identifications from inner W. Gulf and Coromandel by L. Rhodes.
• Personal reports from fishermen and divers on water colour and underwater visibility
• Newspaper reports on water colour, etc.

• There is no routine monitoring of phytoplankton anywhere in the region, and the marine water quality parameters that are monitored are unsuitable to estimate even gross phytoplankton abundance.
• Previous studies of phytoplankton in the region have produced data of high scientific value, but they used such varied methods and standards that the data are rarely comparable.
• Chlorophyll-a concentration near the Leigh Laboratory increased tenfold at the end of August and remained high until this sampling finished at the end of October (from a mean 0.8 µg/l to mean 9.2 µg/l). This compares with 1986 mean values of 0.9 µg/l for winter and 3.1 µg/l for spring, in roughly the same area (Booth and Sonergaard, 1989)
• Over the same period, much smaller increases (approximately doubling) were recorded in the dry weight of total seston and in measurements from a standard turbidity meter (which measures scatterance rather than absorbance of light).
• From 8th October (first sample) through November, counts of phytoplankton in water samples from the same site at Leigh showed high to very high abundances of several flagellate microalgae. By the end of December samples were more or less normal for the time of year.
• The periods of high algal pigments and/or high algal counts coincided with unusually coloured seawater which was sufficiently obvious to be noticed by divers, observers in boats and from the cliff-top.
• On September 13th, an aerial survey from Auckland to Mangawhai Heads showed strongly discoloured water extending over virtually the whole western half of the Hauraki Gulf.
• On several occasions during September-November, at the entrance to the Whangateau Harbour, the incoming tide consisted of discoloured water of very low underwater visibility (1-2m), while the outgoing tide was much clearer (5-7m visibility).
• During September through November many reports from fishermen indicated that strongly discoloured seawater occurred for long periods over virtually the whole Hauraki Gulf west of a line joining Bream Head to Coromandel, with more variable occurrences further north and east.
• All reports were consistent with:
• strongly discoloured water (varying in intensity by the day) over most of the inner and western Gulf north to Leigh
• usually discoloured water but banded in extent and/or intermittent in time extending from the Hen and Chickens, through Little Barrier to Coromandel, and off western Gt Barrier.
• patches or bands of discoloured water at times at the Poor Knights, off Whangarei Heads and at Mokohinau.
• The evidence from further afield is very poor. Scattered reports from the Bay of Islands, Bay of Plenty and South Island suggest that unusual coloured water did occur but it is not clear whether this was atypical in timing or intensity.

• In early October the bloom was dominated (in terms of cell volumes) by Fibrocapsa japonica, but later the armoured form of Dictyocha speculum was co-dominant. The small coccolithophorid Gephyrocapsa oceanica was abundant throughout the bloom, but was sub-dominant in biomass.
• Previously published data for the Leigh region is hard to relate to the 1992 results because of the variety of sampling systems, taxonomic selection, and analysis. Nevertheless standard sampling and analysis (i.e. nets, hoses or dipped water, then identifications and counts) never reported these or similar species as biomass dominants. Furthermore the biomass reached in 1992 is unique apart from the summer bloom of 1983.

   Fibrocapsa was not previously reported at all and Dictyocha (while noted as present) was never regarded as being worth counting. A coccolithophorid, identified as Emiliana huxleyi (but which almost certainly was Gephyrocapsa oceanica) was reported as present each spring for a short burst.
   The 'spring blooms' previously reported seem to be variable in amount, timing and composition (between years), but mostly consist of an increase in diatom numbers together with an increase in their proportion of total phytoplankton biomass.

• Reliable daily salinities are not available for 1992, due to a failure in instrument calibration, but the available data suggests that salinity was close to or slightly below normal during July through October. Records from the Japanese research vessel Hakuho-maru which passed through the area on October 7th 1992, show normal salinities for the season, i.e. >35.0 PSU north of Leigh and <34.9 PSU to the south. There was no evidence of any stratification in the Leigh area at least until November.
• Satellite-derived data of SST anomalies, produced by the Australian Bureau of Meteorology as monthly maps, shows that the highly anomalous sea temperatures recorded at Leigh were part of a widespread pattern (centred east of New Zealand and encompassing the whole country) which lasted throughout most of 1992.
• The SST anomalies of 1992 continued a downward trend from the previous 3 years. 1989 was very arm, 1990 near average, and 1991 was very cool. This 4 year trend closely followed a trend in the Southern Oscillation Index (SOI), which signals the El Niño/La Niña changes of the topical Pacific.
• There is a clear trend relating SST anomalies at Leigh to the SOI (p=0.014) but it only explains 35% of the variance.

• The unusual features of the bloom included not just its intensity, but also its duration, its geographic extent, the species involved, their multiple dominance and their regional differences.
• The coincident and subsequent biological disruptions (see next section) included a wide range of habitats and groups, as well as similar geographic extent.
• The nearest equivalent climatic event at Leigh was the summer of 1983. The very low sea temperatures then were also accompanied by a major, widespread algal bloom and severe biological disruptions, although the season was opposite and the species involved were generally different from those of 1992 (Taylor et al, 1985)
• In both cases, while the detailed causes and limiting factors are likely to have been highly variable, the very low sea temperatures appear to have acted as a general forcing factor.
• It is widely assumed that since (a) nutrients must quickly limit the exponential increase of phytoplankton, (b) upwelling generally supplies nutrients, (c) blooms are commonly associate with upwelling or areas with known increases in nutrients, that extra nutrient supply is required to initiate and maintain a bloom. However, phytoplankton biomass off Leigh is not normally limited by nutrients (see series of papers by Taylor et al) and there was no evidence for any extra supply in 1993.
• It is also widely assumed that normal 'spring blooms' not only require a winter resupply of nutrients, but also the onset of stratification to 'hold' the phytoplankton. This has recently been challenged on the northeast coast of USA, and stratification would have been impossible at Leigh during August and September 1992.

Others: (1 previous years observation)
(i) mass mortality of juvenile scallops, Pecten novaezelandia, (appr 35-50mm) in east Great Omaha Bay, around Jan-Feb 1993
(ii) mass mortality of Dosinia ?subrosea in Omaha Bay (both sides), Jan-Feb 1993
(iii) Starfish disintegration: a low proportion of a scattered population of Ophidaster luidea? had multiple separated arms, Omaha Bay, Jan-Feb 1993.

W J Ballantine and various newspaper reports - local mass mortalities of 'shellfish'
(i) Dosinea subrosea - Te Muri, Jan 1992
(ii) grazing molluscs, kina and whelks - Waipu Cove, 5 Feb 1993

Chris Battershill: sponges (comparison with observations in 1980s)
(i) heavy mortality of sponges, especially Callyspongia (near 100%)
(ii) kelp mortality in 1983

Floor Anthoni (diving locally since 1970s), the earliest reports of:
(i) early kelp sickness and subsequent death, patterned by depth and geographic locality
(ii) sponge 'shut down' and later death and disappearance, and similar effects on other fauna.

Russ Babcock (new staff member) and Russell Cole (5+ previous years observations) (also compared to marine reserve survey 1977-78, see Ayling, 1978 and Ayling et all, 1981):
Analysis of video transects in the marine reserve  (in January 1993) showed earlier and continuing loss of kelp plants (Ecklonia radiata) from the deeper parts of the bed upwards (demonstrated by the sequential loss of canopy, stipe, holdfast and bare space). The deeper part of the kelp beds had 100% mortality and middle parts approximately 50%, throughout most of the marine reserve. Back calculation (estimating lag time) suggests a cause pre-December. Second survey in April 1993 showed continuing progressive mortality into shallower depths, although heavy recruitment has already occurred in the deeper 'vacant' spaces.
Patchy mortality recorded (in may) from Great Barrier's west coast, Little Barrier and Whangarei Heads. No sign of mortality on open coast further north to Cape Karikari, but unconfimred reports of patchy mortality at Spirits Bay. Similar events associated with cold water reported for a laminarian kelp in Japan.

P J Smith (MAF), F H Chang (NZOI) and L MacKenzie (Cawthron):
Have supplied a copy of their draft report for the Royal Society's Shellfish Toxin Workshop, entitled 'Toxic phytoplankton and algal blooms, Summer 1992-93'. This collates information for the whole country, mentioning the Hauraki Gulf events of 1992, but concentrating on events of 1993.

Department of Conservation: reports of high little blue penguin mortalities in Northland.

Maurice Miles, (North harbour Health): Orewa coughing outbreaks and sampling for Gymnodinium breve.
 
 

• Second, there is reasonable evidence to suggest that the data should not be treated as a single set. The strong tendency of the SST anomalies to follow trends in the El Niño/ Southern Oscillation patterns indicates some form of bimodality. If we adopt the simple case of treating years separately depending on whether the SOI is positive or negative, the return period for the April-September 1992 SST anomalies falls from 120 years to 18 (SD 2.1 from 13 data points.). since SOI is a crude and remote index for the purpose, it is likely that the return period could be lower still.
• These return period calculations are only useful in showing the critical importance of understanding the patterns of the phenomena. Unless we can gain some confidence on nature of the patterns we are dealing with, we can place no reliance on the calculations.
• Tf the matter dealt with bench top experiments or easily repeated observations, teplication (more data ) would be the simple answer to these problems. But these data points are years. Another 20-40 years of date would be required to distinguish even the major patterns in statistical terms. Since important decisions (including social, economic and scientific ones) must and will be made on much shorter time scales, we are obliged to consider what, if anything, can be done to assist these.
• there are possibilities which include:
• 'Pattern matching' across different fields of study. It is now being recognised that scale matching (time and/or space) with physical features is an important, indeed necessary prerequisite for many biological phenomena in the oceans (see Mann and Lazier, 1991). The use of satellite data (bothe SST and colour scanning) would be especially helpful in this respect.
• The use of 'proxy data' which can be obtained now, but refers to the past and can be used to determine, at least, pattern types. Examples include sediment core analysis.
• Connecting to existing longer-run data (for other regions, related parameters, etc.) by cross-correlations. L Paul did this some years ago connecting Leigh SST to Leigh air temperatures, and these to Auckland air temperatures and then the results to snapper recruitment.
• It follows from the above that the possibilities of direct causal connections between the aspects of the 1992-93 events, while interesting, are less important than whether there are any types of interaction. If there are, then different aspects can be used as signals for others and pattern matching becomes highly informative.
• The present analysis merely indicated that the return period of the 1992-93 events probably lies in the 10-100 year order of magnitude, and that the most reasonable guess is that it falls in the biologically critical region of 15-25 years, i.e. where such events would be important structuring disturbances.