PART I. PHILIPPINES
SEASONAL FISHKILL PROBLEM IN LAGUNA DE BAY

1. BACKGROUND INFORMATION

Although coastal brackish water areas remain the major source of fish and shellfish in Philippine aquaculture, in recent years the inland freshwater fish farming industry has been rapidly gaining in importance. Most of it is currently concentrated in the Laguna de Bay lake - a major freshwater body located on Luzon Island. Aquaculture is practised here in fishpens, i.e. artificial enclosures constructed of bamboo stakes and fish netting. About 8 percent of the total fish production in the Philippines is supposed to come from Laguna de Bay fishpens, this represents about 40 percent of the production through aquaculture. A major drawback of this convenient method for culturing milkfish (Chanos chanos, locally called bangus) are frequent fishkills, which vary considerably in severity from year to year. A disastrous large scale mortality of milkfish occurred in July 1975, with about 700 ha of fishpens being affected. About 5 million dead bangos were reported (about 2 million of them of marketable size). Value of fish lost was estimated at 2.8 million (Anon., 1975). This large-scale calamity caused a strong public concern and a special Coordinating Committee on the Laguna Lake Research and Development Program was established under the auspices of the National Science Development Board. This Committee comprises of the Bureau of Fisheries and Aquatic Resources (BFAR), Laguna Lake Development Authority (LLDA), National Pollution Control Commission (NPCC), University of the Philippines at Los Banos, Philippine Council for Agriculture Resources Research (PCARR) and other research institutions.

In view of the seriousness of the mass milkfish mortalities in Laguna de Bay fishpens the author concentrated his effort mostly on this phenomenon during his stay in the country. His main objective was to recommend follow-up studies which could lead to an understanding of the cause of the fishkills so that an adequate predicting and warning system signalling an approaching kill could be developed.

2. STUDY AREA (LAGUNA DE BAY LAKE)

2.1 Limnology and water quality

Rabanal et al (1964) presented the first limnological survey of the lake. This was before the introduction of fishpen culture when the lake was used only for open water fisheries. Yearly catches averaged about 80 000 metric tons (mt) of fish, as well as about 240 000 mt of shrimps, clams, aquatic plants, etc. which were used mainly in the duck-raising industry.

The total area of the lake is 90 000 ha, which makes it one of the largest lakes in Southeast Asia. It is uniformly shallow with an average depth of 2.5 m and a volume of about 2.7 × 10⁹ cubic meters.

The lake is thermally unstratified. Water temperature is high throughout the year, averaging 27.6°C and fluctuating from 22°–34°C. Under wind action, waves easily mix the water to the full depth and stir up the soft sediments. Rainfall supplies about 2 300 mm of water annually through 210 000 ha of the lake watershed. The two largest influents, the Pagsanjan and Santa Cruz rivers, flow into the east bay of the lake. The only outlet is the Pasig-Napidan river in the northern part of the west bay. This causes a general east-west circulation, however, the main water currents are created by winds, which generally cause a clockwise water movement around Talim Island (Fig. 1).

During the dry season, evaporation may exceed the inflow and the lake level can drop near or even below the sea level. When this happens, the Pasig-Napidan river reverses its direction of flow and some amount of saline and polluted water enters the lake.

Suspended organic matter, mainly plankton, was studied by Notario (1964). She presented a qualitative analysis of Laguna de Bay plankton as follows:

She found three main groups of zooplankton, namely: cladocerans, copepods and rotifers. There were significant variations in the concentration of suspended organic matter and the occurrence of algal blooms during certain periods of the year. The maximum concentrations of suspended organic matter (18.5 mg/m³)* were observed during the dry season months of May-July and the minimal (4.3 mg/m³)* occurred in the rainy season (August-October). It is noteworthy that the dry season plankton bloom was followed by a decrease in concentration of suspended organic matter. Average annual Secchi disc transparency for the lake was 40 cm. Benthic fauna of the Laguna de Bay at the same period was described by Mercene (1964).

Water quality and the recent state of lake pollution is dealt with in detail by a specialized comprehensive report prepared by SOGREAH** (1974) as well as by Thorslund (1972) and Lee (1975). These studies show that the

^* The amounts of suspended organic matter expressed in dry weights were taken as indications of the relative amounts of plankton.

^** SOGREAH is a consultancy firm which conducted the Laguna de Bay water resources development study with the support of UNDP and ADB.

Laguna de Bay water is of medium hardness with seasonal signs of salinity (specific conductance range is 450–500 μmhos/cm; but in localized areas during the sea water intrusion it can be as high as 600–3 360 μmhos/cm). The pH values range from 7.9–9.2 depending presumably on the algal cycle. Sulphate and chloride content fluctuate considerably; from 22 to over 220 mg/l SO₄ and from 125 to over 1 300 mg/l Cl. Water unaffected by sea water intrusion however, remains within the calcium-bicarbonate type. Nutrient dynamics was not studied with adequate frequency; the data obtained at 2–4 week intervals showed a substantial fluctuation in total ammonia-nitrogen (traces to 500 μg/l) and reactive phosphate (traces to 200 μg/l). Nitrate and nitrite levels varied also (0 to 500 μg/l NO₃-N₁; 0 to 20 g/l NO₂-N). Total nitrogen (Kjeldahl) concentrations were in the range of about 300 to 1 300 g/l N.

The principal sources of nutrients come to Laguna de Bay via inflow from the densely populated watershed and from the polluted Pasig-Napindan river when reverse flow occurs. Inflow from the Pasig river is often anoxic with BOD₅ values of 20–39 mg/O₂, caused by municipal and industrial effluents. The lake is in a fairly advanced stage of eutrophication, with abundant growth of blue-green algae, mostly Microcystis (as much as 1 million algae per milliliter), often forming thick compact scums in sheltered areas. Another acute problem is extensive growth of water hyacinths (Eichhornia crassipes) which cover a great deal of the lake surface. The plant is not bottom-rooted and can float freely following the water and wind currents becoming concentrated in the windward areas.

There are several ongoing eutrophication control programmes aimed at minimizing the effect of pollution in the lake. These include a proposed hydraulic control structure on the Pasig river to stop the salt water intrusion into the lake, a waste water diversion system (West shoreline interceptor), the various flood control programmes.

2.2 Fisheries, fishpen industry

Open-water fishing in Laguna de Bay as a livelihood only supported a marginal economic existence prior to introduction of fishpen culture. The 90 000 ha lake has some 10 000 fishermen, or about one for every 9 ha. There are 23 species of fish found in Laguna de Bay, belonging to 16 families and 16 genera (Pruginin 1972). The most common and dominant species are the gobies and perch (Therapondae) which have relatively low market value, plus some common carp, catfish (Arius spp), snakehead (ophicephalus striatus) and Tilapia mossambica.

Unit rate of yield in lake fish amounted to 433 kg/ha and was sustained entirely by the natural biota of the Laguna de Bay (Lee, 1975). Before the present pollution of the Pasig river, migratory species from the sea were grey mullet (Mugil sp.), milkfish (Chanos chanos) and Manila catfish (Arius manilensis). The milkfish is the only species for which fry and fingerlings can be obtained regularly from outside sources. Data presented in the SOGREAH Report (1974) showed a sharp decline in fishpens included to only about 40 000 tons (50 per cent reduction) caused presumably by recent pollution of the lake.

The Laguna Lake Development Authority (LLDA) started pen culture experiments in July 1970. The objective was to introduce milkfish, a herbivorous species reared normally in brackish water fishponds, a very popular food fish and highly valued in the Philippines, for rearing in artificial enclosures within the lake without any supplemental feeding or fertilizing. Various fish pen designs and sizes were utilized. It was shown that 5 ha pens provide for more efficient management than the larger size pens. Results of the initial studies were very encouraging (Delmendo and Gedney, 1974; LLDA 1974). A potential annual yield of 1 500 kg/ha was demonstrated during the initial experiments which was more than 3-½ times the annual open water fish catch. The product (milkfish) was more marketable than the usual open water species, and subsisted entirely on the natural food of the lake.

From 1971 up to date, the fishpen industry in the lake has boomed. The total area of fishpens in 1973 was estimated to be over 5 000 ha, with the average annual yield estimated to be 4 000 kg/ha. This is about 10 times more than the normal open water yield, and the total fishpen production nearly equals the previous total catches in the whole lake (LLDA, 1974). It was shown that the potential for expansion is 3 to 4 times more than the present extent, with a possible 27 per cent return of total investment (including cost of fishpen construction which is the major cost item).

The latest estimates (Felix, 1974) showed that the total fish pen area to date covered 7 000 ha, with an annual yield in some cases reaching as much as 10 000 kg per ha. Thus, the fishpen industry of Laguna de Bay contributes much to the protein requirements of the country.

3. DIFFICULTIES ENCOUNTERED IN FISHPEN CULTURE

In spite of the initial boom, the fishpen industry suffered some serious drawbacks which cooled down, to some extent, the original optimism. A number of problems in this kind of aquaculture became apparent with the increased use of fishpens and the expansion of cultivated lake areas. These were mainly the limited durability of the bamboo framing and netting material, frequent typhoons and floods, lack of supervision and control in awarding sites which led to congestion of fishpen units, and resulted in poor production. The above mentioned disadvantages of fishpen aquaculture can be avoided by proper technics and administrative measures and the BFAR has formulated plans and regulations to remedy these problems. (Felix, 1974). There are however, additional serious problems caused by adverse biological factors which bring about seasonal fishkills of which the most serious occurred in July 1975.

3.1 Analysis of mass milkfish mortalities in Laguna de Bay: existing information

A. There are many contradicting theories and observations on the events associated with the massive fishkills during the months July-August, prior to 1975. The earliest occurrences were primarily attributed to excess of blue-green algae mainly Microcystis. It was supposed that algal toxins produced by Microcystis blooms or emitted by bacteria associated with the algae, or clogging of the gills resulted milkfish mortality. There were also some theories on the liberation of hydrogen sulfide, a phenomenon of “black water” patches moving through the lake and killing the fish, choking of fish stocks by water hyacinths, or strictly meteorological factors (i.e. extremely high water temperatures and resulting low oxygen levels). There have been speculations about the possible effect of salt water intrusions, affecting the algae and causing them to die off; or the effect of a heavy rainfall, which is believed to be very acidic and lowers the pH value of lake water to levels intolerable to algae. Later observations by Delmendo and Gedney (1974) associated the fishkill, which killed more than 20 000 fish in May 1972 in the LLDA pilot fishpen at Looc area specificially with the peak of a Microcystis bloom. The incident occurred at a time when the algal bloom was extremely dense and the lake resembled a thick pea soup, with a prominent smell of rotting algae. In 1973, even heavier losses were encountered in fish pens in central bay, this time in June, July and August. About 90 per cent of the fish stock in the affected areas were lost. Dying or dead fish were observed in early mornings and late evenings. It should be noted, however, that no fish mortalities were ever reported from the open waters. No oxygen measurements were taken during these kills. Some hints were made about the possibility of pesticides in lake water contributing to fishkills. Clogging of fishpen netting by water hyacinths and drifting debris was observed, resulting in poor water circulation and oxygenation. A precipitation of CaCO₃ on the fishpen bottom creating a solid white crust was also reported. Overstocking, 60 000 fish per ha and over, was suspected in some cases.

B. The disastrous magnitude of the 1975 fishkills, which occurred around 29 June to 12 July, caused a great public concern attracting wide publicity and called for strong government response. The events prior to and accompanying the seasonal fishkills as observed by the Bureau of Fisheries and Aquatic Resources, can be summed up as follows:

1) On 29 June 1975 there were reports of the abnormal behaviour of bangos in fishpens along some coastal villages of Cardona town (Fig. 1). Starting 30 June until 5 July, daily reports of the mass mortalities of bangus were received by the BFAR. By 5 July, 996 hectares were reported to have been affected, and on 12 July some 50 000 bangus were reported dead in one area, Lambak, Cardona.*

2) Diurnal measurements of dissolved oxygen concentration in some affected fishpens were taken. On 7 July at Lambak, Cardona, inside one fishpen, the DO concentrations fluctuated from 4 mg/l at 1 300 to 0 mg/l at 0600 hours. Outside the pen, the range of DO at the bottom was from 3.5 to 0 mg/l. The surface water had a DO content of 7 mg/l at noon, which decreased to 0 mg/l at 6 a.m. (Fig. 2). The following morning, high bangus mortality was observed.

3) The author found in the records of the BFAR-Los Banos Laboratory some additional data which demonstrate clearly an oxygen depletion also in other affected fishpens, i.e. zero DO values at Subay, Cardona on 7 July at 0600; 1.5–1.9 mg/l DO levels at Wisdom on 9 July at 2300; 0–1.06 mg/l DO at Boor at 2200 the same day. The Secchi disc transparency readings taken inside a fishpen at Boor, Cardona (station Sampad) showed a dramatic increase between 10–11 July (from 0.54 to 1.74 m) which indicates a rapid die-off and sedimentation of the Microcystis bloom and subsequent clearing of the water column.

^* Parts of an internal report quoted with the BFAR and NSDB permission.

Figure. 1 Laguna de Bay: distribution of milkfish mortality, plankton biomass, and water circulation pattern (From BFAR data July 1975)

Figure 2. Comparison of dissolved oxygen contents inside and outside the fishpens during a 24-hour period (From BFAR data)

Similarly, analysis of water samples from affected areas performed by NPCC team showed low level DO at Malamgan, Cardona (0.7 mg/l in the surface water) on 12 July at 0620 and 1.8–2.9 mg/l DO at Taguig and Cardona (Table 1).

As a result of these alarming reports, a special interdisciplinary research group was created under the auspices of the NSDB to augment activities in the lake through integrated research, the results of which could also provide supportive information to the special task force undertaking the physical cleaning of the lake. On 23 July, an intensive 10-day monitoring programme was initiated at the location of the maximum reported milkfish mortalities in the waters east of Talim Island (central bay) to obtain relevant water quality and biological data. Monitoring was carried out jointly by LLDA, NPCC and BFAR. The fishkill was receding by that time and the oxygen conditions were almost back to normal. The minimum DO concentrations found at the LLDA Looc experimental fishpens were between 2.5–3.0 mg/l (Fig. 3) and early afternoon concentrations were back to saturation levels (over 8 mg/l).

The 10-day monitoring programme yielded valuable information on the ammonia levels in the post-kill period. On the basis of North American studies (Barica, 1973, 1975b) and preliminary observations made in the Philippines (Lenarz, personal communication, 1975) there are indications that the NH₄N level could be comparatively high immediately before and during the actual fishkill period compared to the post kill phase. From present knowledge a theoretical picture of what could happen is presented (Fig. 4). Although this is speculative at this stage for lack of actual data, this phenomenon can be worth following up.

Table 1: Dissolved Oxygen Content of Water Samples (July 11–12, 1975 NPCC data)

Figure 3. Diurnal dissolved oxygen monitoring in the post-kill period

Figure 4. Ammonia levels at three fish kill sites in Laguna de Bay (10-day monitoring by LLDA) compared with hypothetical curve for fish kill period (Barica, 1973)

In the course of preparation of this report some diurnal measurements of nutrients and DO were done by the BFAR team (Los Baños Experimental Station), in both affected * and unaffected fishpens (November 1975, Appendix 1). The data were obtained during the safe season and can be used as illustration of normal conditions in fishpens in the non-fishkill period. NH₄N ranged from 45–136 μg/1 in the affected fish pen and from 41–120 μg/1 in an unaffected one. This indicates no significant difference between the two types of pens, however, a considerable diurnal variation was noted. Soluble reactive phosphate (SRP) varied from 3–72 μg/1. Oxygen conditions during this period were normal. Predominant algal species are listed in Appendix II and show that in November 1975 there were still blue-greens dominating the water column (mostly Microcystis spp.) with diatoms being next in abundance. Fish pens varied considerably in phytoplankton composition. Mainly the unaffected fish pens (Los Baños region) differed from the rest of tested sites. Los Baños I (Antique project) contained diatoms (Coscinodiscus, Navicula, Cyclotella, Melosira). There are some unexplained differences between various parts of Laguna de Bay, as well as between individual fish pens.

3.2 Similarities of the Laguna de Bay fish mortalities with summerkills in mid-western North America

3.2.1 The cause of summer fishkills

The similarities between the symptoms, causes and consequences of the Laguna de Bay fishkills and those occurring in small prairie lakes in midwestern United States and Canadian prairie provinces are striking. Small prairie lakes, so called potholes, resemble in many ways fish pens: they are small (1–10 ha), shallow (1–3 m), unstratified, warm and with limited water exchange during summer, highly eutrophic due to input of nutrients from rich agricultural land with a regular occurrence of heavy blooms of blue-green algae, mostly Aphanizomenon flos aquae (in Canada) or Microcystis and Anabaena species (Barica, 1975a). Those which reach an algal biomass of 100 μg/l (as chlorophyll “a”) experience summerkill with a probability of about 50 per cent, those exceeding 200 μg/2 chlorophyll “a” almost always experience summerkill (more than 80 per cent probability).

^* i.e. affected by fishkill

The cause of fishkills is a sudden die-off of a heavy algal bloom and its subsequent rapid bacterial decay in the whole water column creating total oxygen depletion. What triggers the actual collapse of the bloom is not yet known; it may be a viral or bacterial attack, exhaustion of available nutrients, excretion of self-inhibiting substances, severe pre-dawn DO deficiency, or mechanical upwelling of anoxic and toxic bottom waters by wind action. However, it is noted that in many cases a new bloom of the same blue-green algae occurred within 2–3 weeks. There is also possibility that a sudden drop in water temperature and sunlight intensity can also trigger algal die-offs. The die-off of algae is followed by a rapid biochemical degradation of organic substances in the water phase surrounding the algal cells, resulting in a total oxygen depletion and liberation of ammonia, which at the high pH values of lake water can be present in a non-dissociated toxic form. This can result in a partial or total fish mortality. Low oxygen concentration is the primary cause of fish mortality (Ayles, et al in press), but there may be synergistic effects of unionized ammonia (as NH₃) and other toxic end-products of the biodegradation of algal cells. The peak of toxic (free) ammonia comes about 1–2 days after the fishkill, when fish are already dead.

3.2.2 Selection of significant environmental parameters accompanying summerkills

When research in Canada was started on the summerkill phenomenon, numerous parameters were monitored, in water and the lake sediments. The water analysis included total ammonia nitrogen, soluble reactive phosphate (SRP), total dissolved nitrogen (TDN) and phosphorus, dissolved organic nitrogen (DON), phosphorus and carbon, minimum and maximum dissolved oxygen, Secchi disc transparency and all major ions. The lake sediments were analyzed for ammonia and SRP content in interstitital water, immediate oxygen demand, biochemical oxygen demand (1 day incubation) and oxidation-reduction potential. Parameters, which showed the highest significance levels are summarized in Table 2.

From among significant parameters those posing analytical difficulties in field conditions (TDN, DON) were excluded, as well as all sediment parameters. Finally, chlorophyll “a”, ammonia, dissolved oxygen and Secchi disc transparency emerged as the most convenient, easy to measure and characterize best the summerkill process. These were used in later studies.

Table 2. Significant parameters* (x) for prediction of summerkill risk in Canadian prairie ponds (Barica 1975a)

w = winter, s = summer, y = max. chlorophyll a(s)

^* 1% level of significance for water, 5% level for the sediments and specific conductance (Snedecor and Cochran 1967).

3.2.3 Mechanism of a prairie summerkill

The technical indicators which show significant changes during the summerkills can be divided into 3 groups (Barica 1973):

1. Dissolved oxygen, pH, chlorophyll “a” and particulate C, N and P reach their maxima during the bloom phase and drop substantially during the algal collapse and fishkill phase;

2. In contrast to this, ammonia and phosphate are at the minimum level during the bloom peak, build up gradually in the pre-kill phase and increase about ten times during the fish kill. A pH drop of 1–2 units is also observed.

3. Dissolved organic N and P substances decrease during the collapse of the algal bloom, but increase again during the anoxic fishkill period.

In the recovery phase there is a stabilization of water chemistry. The mineralization processes shift into bottom water layers as the bulk of the dead plankton sinks to the bottom. Nh₄N and phosphate are released again and become available for further algal blooms and possibly new collapse.

Dissolved oxygen (DO), Secchi disc transparency, and NH₄N serve as convenient parameters to show the most significant changes during and after algal collapses (Barica, 1975b). Figure 5 presents a simplified model of the curves for these values during a summerkill. From this model the following general pattern can be drawn: the algal biomass reaches its exponential phase rapidly, with chlorophyll “a” values of 100–200 ug/1 and DO concentrations at their maximum (over 15 mg/1 as an average for the profile; occasionally over 20 mg/1 at the surface). The Secchi disc transparency is only 0.2–0.4 m in this phase, and the ammonia and SRP at near-zero levels. Then suddenly within a few days, the whole mass of algae starts sinking to the bottom of the lake, and the water clears downward from the surface, taking on an opalescent appearance presumably due to colloidal organic matter released by the bacterial breakdown of dead algal cells. Simultaneously, a substantial drop of DO concentration is observed, accompanied by a continuous increase in concentrations of NH₄N. This phase culminates when DO drops to near-zero concentrations. Then, the release of ammonia is maximal indicating anaerobic bacterial ammonification of organic matter released by the algal cells. This process is not always steady and can follow a pattern of reversals when the mass of algae initially dies to about 50 per cent of the original bloom and then recovers and continues to grow. DO drops, but not below the level required by fish. After several partial collapses of this kind, a total crash eventually takes place and anoxic conditions develop.

Figure 5. A simplified model of a summerkill mechanism. Critical levels indicate high summerkill risk. (After Barica 1973, 1975 b)

Primary productivity (as net oxygen production) and oxygen uptake of a typical summerkill lake is shown in Fig. 6, using the light and dark bottle technique (4-hour exposure) corrected for the sunlight duration and transparency (Vollenweider, 1969). Primary productivity at the peak of an Aphanizomenon flos-aquae bloom is very high (as much as 9 g/m²/day O₂). Even in the case of a partial bloom collapse which is not accompanied by substantial DO depletion primary productivity drops to negative values; i.e. the respiration processes dominate the entire water column. This is because the dead plankton is rapidly decomposed by bacteria. In the case of a partial collapse, when anoxic conditions are not reached, the process is only temporary, however, if the collapse is total and the DO content drops to near-zero concentrations in the whole water column, respiratory dominance has a longer duration. Pureley respiratory data, as given by the values of the oxygen uptake in dark bottles (Fig. 6, lower part, BOD values) show a trend analogous to the primary production, with peaks occurring at the partial or total bloom collapses.

3.3 Possible remedial actions in North American conditions

The present policy in dealing with the summerkill risk is proper classification of prairie, lakes into low and high risk categories, and to avoid stocking lakes with a high summerkill risk. This is done on the basis of basic chemical and morphometric parameters (summer chlorophyll “a”, winter ammonia, minimum DO, mean depth), as described elsewhere (Barica, 1975a).

The approach of the algal collapse (and subsequent :fishkill) in Canadian prairie lake conditions is signalled in advance by a sudden increase in algal standing crop to levels above 100 ug/1 of chlorophyll “a” and by a steady build-up of ammonia to about 500–600 ug/1 NH₄N. Color of the algal blooms change from the normal green to yellowish, and the colonies of algale start clumping together. Water between algal colonies becomes turbid and dark colored. This is a signal that a collapse can occur at any time. Daytime oxygen levels do not show any substantial change during this period and cannot be used as an indicator of approaching summerkill.

4. RECOMMENDATIONS

4.1 Rationale

A. From the existing information and documentation of the 1975 fishkill as well as field observations by the author it appears to be primarily a collapse of Microcystis bloom which causes milkfish mortalities in the Laguna de Bay fishpens. The collapse creates a nearly complete oxygen depletion in the water as a result of decomposition of dead algal cells on a massive scale. This process is further aggravated by high water temperatures resulting in low oxygen saturation levels while the initial trigger of the algal collapse is not known, the author suspects sudden change in weather conditions, namely decrease in light intensity by heavy overcast, however, other reasons, like salt water intrusion, stirring up the sediments by wind, clogging of fish pen netting thus reducing water circulation, overstocking and overcrowding of fish pens, and pre-dawn oxygen depletion in certain areas, cannot be ruled out.

Figure 6. Net production and respiration (BOD) accompanying collapse of an Aphanizomenon flos-aquae bloom in Lake 885, Erickson, Canada (From Barica, 1975b)

Investigation into the “black water” phenomenon, its movement and state of water quality (possibly decomposing patches of algal scum moving across the lake) should also be considered. The author doubts that the water hyacinths play any significant role in this process, but this needs to be verified.

B. Despite striking similarities with summerkills in Canada and mid-western U.S., no direct application of the results and conclusions on remedial activities under Philippine conditions can be made without studying the process on a scientific basis for a sufficient period of time, covering at least two years. The critical limits for Laguna de Bay must be established, as they may differ substantially from those found in Canada. Emphasis should be laid on the possibility of signalling the approaching seasonal kill in the fishpens and saving of the milkfish stock by harvesting it before the kill occurs. At the same time, the signalling parameters must be simple and fast to measure. Involvement and cooperation of researchers from a large number of agencies and individual fishpen operators to insure coverage of all critical areas of Laguna de Bay will be of utmost importance.

4.2 Selection of experimental pens

It will be necessary to select about 8–10 fishpens in the susceptible fishkill areas (central basin) and about 2 or more in less affected areas (Los Baños) as controls. It is important that exact stocking rates are known in each of the pens and stocking rates should be similar. It would be advisable for each involved agency to select its own experimental pens, so that the effort can be multiplied and more data obtained for multi-variant statistical processing and determination of significant levels of parameters. It is also important that the fishpen operators who agree to monitoring in their pens supply stocking rates and harvest figures accurately. In addition to these observation pens, which will be monitored regularly, the monitoring teams will also have to keep in close touch with other operators who would notify them quickly when observing dying fish. Spot checks can be made in non-monitored fishpens to see if the kills occur there earlier than in the selected pens.

4.3 Monitoring programme

Collecting parameters pertaining to seasonal fish mortalities in Laguna de Bay fishpens should be done in three different but interrelated disciplines:

1. meteorology and weather information in general

2. fisheries biology and aquaculture

3. water quality and limnology

4.3.1 Weather information

Basic data on air temperature, wind velocity and direction, rainfall, humidity, solar radiation and sunlight duration and cloudiness should be collected daily from a small weather station installed directly on the shore or on one of the fishpens or from the nearest weather station if this is not possible. The data should be correlated with the occurrence of fishkills to determine if weather conditions play any significant role in triggering the algal bloom collapse. Co-operation is desired with the Philippine weather authorities who should be invited to participate in the monitoring programme.

4.3.2 Fisheries biology and aquaculture

Parallel experiments in fishpens of approximately the same size and design should be performed with stocking rates ranging from an optimum (1–2 fish/m²) up to overstocked rates (6–8 fish/m²). The objective will be to determine the growth rates, as well as to study the effect on water quality in overstocked pens (metabolite build-up, oxygen uptake) and possibilities of fishkills resulting from deteriorated water quality conditions (direct ammonia toxicity, oxygen depletion). Reliable information on fish stock biomass during the season should be obtained as a possible basis for estimating oxygen uptake by fish and for determining the overall oxygen budget and availability. Some thought should perhaps be given to developing new areas of Laguna de Bay for fishpen aquaculture to avoid overcrowing and congestion in central bay.

4.3.3 Water quality sampling of selected fishpens

Monitoring of selected water quality parameters and limnological data should be considered to find the cause, and possibly means of early prediction, of seasonal milkfish mortalities in Laguna de Bay fishpens.

A. Selection of parameters. As discussed in Section 3.2.2, there are some parameters which are significant and should be measured (essential monitoring requirement) and others which can possibly bring new information or show a high significance are optional.

Note: Analyses of fixed nutrients must be performed within 4 hours after sampling to obtain meaningful results. Sample for chlorophyll “a” determination should be filtered within 8 hours through a glass-fibre or a millepore filter and stored frozen in dark until analyzed.

Any of these parameters can indicate approaching fishkill by its sudden change or continuous build-up. Processing of the data will identify which parameters are diagnostic, and at what levels they can be used for predicting seasonal fishkills. Classification of fishpens and areas of the lake will also be possible.

Selection of the essential and optional parameters to be observed will also depend on the manpower and equipment conditions of the participating agencies.

B. Standard analytical procedures. It is necessary that the participating agencies take the earliest possible steps for the standardization of analytical methods so that their data will be comparable. This is particularly important for total ammonia, dissolved organic nitrogen and chlorophyll “a” determinations, which can vary considerably according to the methods used. The author particularly recommends the use of the phenolhypochlorite method for total ammonia, and fluorometric method for chlorophyll “a” (Stainton et al, 1974). Other recognized standard methods can also be used as long as they are used by all agencies (for example, Strickland and Parsons, 1968; Golterman and Clymo, 1971 and others).

C. Laboratory and field equipment. Conventional laboratory and field equipment for the recommended monitoring programme will be sufficient. A spectrophotometer should have a wavelength range of about 300–880 nm and max. 10 nm bandpass and can be used for both nutrients and chlorophyll “a” analysis. A fluorometer (such as Turner's) would be suitable for fluorometric chlorophyll “a” determination, if decided upon. Sampling bottles (Van Dorn, Ruttner or comparable types) can be used to obtain samples from different depths of the lake. It is necessary that each agency has adequate equipment to ensure proper data acquisition.

D. Type and frequency of monitoring. Weekly spot sampling: The selected parameters should be monitored at one or two-week intervals during the period October-April (low probability of fishkill), and at least once a week during the critical period of May-September. The frequency of sampling should be increased to 2–3 times a week during the fishkill period. Recommended sites for sampling: centres of each experimental pen plus one sample outside the pens. Depths: sub-surface (10 cm), 1 m, and about 20 cm above the bottom of the pen. It is very likely that the ammonia build-up signalling approaching fishkill will be detected first in the near-bottom zone as a result of decomposition of sedimenting phytoplankton. Recommended sampling time: early morning hours, when DO concentrations are at their minimal.

Diurnal measurements. Previous studies (SOGREAH, 1974; LLDA and BFAR, 1975 data) indicated substantial diurnal fluctuation of dissolved oxygen in the Laguna de Bay. It may be the early-morning DO depletion which starts the periodic fishkill events, or a continuous decrease in early morning concentrations might signal the approaching algal collapse. It should be remembered that during an actual fishkill DO depletion persists in the total lake profile during the daylight, as there are less living algae to counterbalance bacterial oxygen uptake. There are no data available at present on diurnal nutrient fluctuation during the kill. Ammonia and SRP can be useful in predicting the seasonal fishkills.

Since the observations are time and manpower consuming, diurnal measurements should be done only in few representative fishpens (for example 2 highrisk fishkill pens in the central bay and 1 typical low-risk pen in the Los Banos region), preferably at 2–3 hour intervals over a 24-hour period, every 1–2 weeks from May to August. If notification is received from a fishpen operator of impending fishkill, 24-hour measurements should be performed on such pens at that critical time also. If it is not feasible to sample fishpens at three different depths during the diurnal studies, a sub-surface or mid-depth sampling will serve the purpose.

D. Other limnological and related studies. To better understand the mechanism of the mass fish mortality adequate attention should also be paid to the phytoplankton and zooplankton composition, its dynamics and successions. Monthly counts and species identification in selected fishpens should be sufficient. Also, information on fishpen sediments, namely their ammonia and SRP levels and oxygen uptake (immediate and biochemical) will be desirable, obtained preferably in pre-kill and post-kill phase, including again low and high risk fishpens (Los Baños and Cardona or Looc in the central basin.) These studies are already incorporated in an overall research program for Laguna de Bay and tributary rivers prepared by the NSDB task force (Appendix III).

Pesticide and heavy metal pollution should be looked at as well. Some species of algae can accumulate these substances up to levels detrimental to fish.

5. CONCLUSIONS

5.1 Whatever the similarities and analogies with the North American summerkills, no direct application or interpretation of data of the Laguna de Bay conditions is possible without an adequate and systematic research programme by the Philippine agencies concerned. Only the strategy of how to tackle the problem is directly applicable, and the experience gained in Canada can shorten the period needed for establishing prediction parameters for Laguna de Bay to about 2 to 3 fishkill seasons.

5.2 There are a number of established research institutions with skilled professionals and technicians in the field of fisheries, water pollution and limnology in the Philippines which are capable of dealing with the acute problem of seasonal fish mortalities in Laguna de Bay fishpens. Close co-operation of various agencies under the existing Coordinating Committee on Complementary Researches for the Laguna Lake Research and Development Program is providing a promising institutional framework necessary to ensure adequate monitoring of the lake. A comprehensive list of research problems set up by these agencies already exists (Appendix III). The recommendations contained in this report are meant to supplement this overall programme by dealing more specifically on fishkill prediction objectives.

5.3 It is quite obvious that pollution abatement programmes in the Laguna de Bay basin are a pre-requisite of future fisheries development in the lake, as eutrophication itself is the major cause of fish mortalities in the lake. The author is convinced that the joint effort and close co-operation of the involved agencies will bring a better understanding of the fishkill mechanism, and maybe better predicting parameters as proposed here (for example, dissolved organic nitrogen, carbon, BOD or others). Incentives for other institutions to join this programme should be encouraged and adequate instrumentation for the involved agencies provided with a high priority.

6. OTHER ACTIVITIES

6.1 Milkfish mortalities in brackish water ponds in Iloilo Province

A completely different mechanism of milkfish mortalities in brackish water coastal ponds was observed at the University of the Philippines, College of Fisheries, Leganes experimental ponds. Following heavy rainfall periods (normally in April and October during the monsoon change), bangus (milkfish) kills are common in this area. There are no data or reports available on this phenomenon except some possible explanations of the event by fish producers and fisheries biologists which are summarized in a College of Fisheries, University of the Philippines report.

They found a pH in ponds following a 8.6 cm rain dropped from 8.5 down to 4.1, with an additional decrease in water salinity (from 54.7 to 32.2 ppt). The mortalities were attributed to “acid death” which is very probable.

The consultant recommended additional studies, which would include measurement of pH salinity and dissolved oxygen in the ponds as well as determinations of sulfate, sulphide, bicarbonate-carbonate and chloride, which could identify the anion causing the depression of pH in the pond water. It was also recommended that soil testing for the above parameters (in 1:5 soil water extract) be performed at 50 cm depth intervals from the top of the dike to about 1 m below the pond bottoms to find the source of low acidity (sulphuric or perhaps hydrochloride acid). The same was suggested for rain water samples collected in the vicinity of the affected ponds and in the run-off water from the dike slopes.

Since it is usually the smaller ponds which are affected by “acid death”, the consultant feels that a minimum ratio of pondwater volume to dike surface area exists below which effect of acid material (whether originating from rain, surface run-off or leaching from soil after oxidation of reduced sulfur compounds) is critical. This factor should also be included in further studies.