Summer flow event induces a cyanobacterial bloom in a seasonal Western Australian estuary

CSIRO PUBLISHING

www.publish.csiro.au/journals/mfr

Marine and Freshwater Research, 2003, 54, 139–151

Summer flow event induces a cyanobacterial bloom in a seasonal Western Australian estuary Barbara J. RobsonA,B,D and David P. HamiltonA,C A Centre

for Water Research, University of Western Australia, Crawley, WA 6009, Australia. address: CSIRO Land and Water, GPO Box 1666, Canberra, ACT 2601, Australia. C Present address: Department of Biological Sciences, University of Waikato, PO Box 3105, Hamilton, New Zealand. D Corresponding author; email: [email protected]

B Present

Abstract. In January 2000, record rainfall led to the first recorded bloom of Microcystis aeruginosa in the Swan River estuary. A simple model is used to examine the bloom dynamics and the unusual conditions that produced it. Laboratory trials were conducted to determine the response to salinity of M. aeruginosa, while other parameters for the model were obtained from the literature. Growth was found to be optimal at salinities up to 4, and declined to zero at 25. The unseasonable summer rainfall flushed brackish and marine water from the estuary and produced a surface mixed layer with low salinity. The model simulations show that the hydrological conditions, in combination with high concentrations of inorganic nutrients (dissolved inorganic nitrogen >1.2 mg L−1 , filterable reactive phosphorus >0.02 mg L−1 ) in river inflows, high water temperature and high daily insolation, promoted rapid phytoplankton growth, favouring dominance by M. aeruginosa. Doubling rates during the bloom were around 0.35 day−1 and cell counts exceeded 105 cells mL−1 within three weeks of the inflow event. Although this doubling rate ultimately determined the total bloom biomass, local concentrations were strongly influenced by physical processes that concentrated M. aeruginosa cells both vertically and horizontally, and advected a seed population from the upper estuary into the lower basin. Extra keywords: advection, cyanobacteria, Microcystis aeruginosa, model, nuisance algae. Introduction Freshwater cyanobacteria are not generally associated with estuarine environments because of their limited tolerance to salinity (Kirst 1990) and, in freshwater reaches, because of constraints due to turbulence (Cloern 1996) or high rates of flushing (Chan and Hamilton 2001). Further, cyanobacteria that can fix atmospheric nitrogen may be unable to do so in estuarine and marine environments because of deficiencies in organic matter (Paerl et al. 1987) or relatively high levels of turbulence (Paerl 1985). Nonetheless, blooms of the cyanobacterium Microcystis aeruginosa have (albeit rarely) occurred in estuaries, with cases reported in the Patos Lagoon, Brazil (Yunes et al. 1996), which is flushed seasonally with freshwater, and in the Potomac River (Jones et al. 1992), a tidal freshwater system. In many freshwater environments, M. aeruginosa is particularly problematic because of its ability to accumulate and form blooms at the water surface (Reynolds 1997) and to produce toxins (microcystin) (Baker and Humpage 1994; Jones and Orr 1994). Bloom formation by M. aeruginosa is closely related to induction of buoyancy through gas vesicles (Walsby et al. 1997) and reductions in carbohydrate content following darkness or low light exposure (Kromkamp and Walsby 1990; © CSIRO 2003

Wallace and Hamilton 1999). Ganf and Oliver (1982), among others, have suggested that buoyancy regulation is a means to overcome the growth impediments of vertical separation of light and nutrients, although there is conjecture about whether this is an adaptation that is specific for nutrient assimilation (Bormans et al. 1999). In the Swan River estuary in Western Australia, cyanobacterial cell densities have rarely exceeded 5000 cells mL−1 (Hosja and Deeley 1994). It is believed that the higher densities have been due to rainfall events transporting cyanobacteria from adjoining wetlands into the estuary, rather than to growth per se within the estuary (John 1994). Freshwater cyanobacterial cells that enter the estuary are likely to be adversely affected by high salinity (>20) in summer and autumn (Kurup et al. 1998). Further, growing conditions are suboptimal in winter and spring, when water temperature and ambient light levels are relatively low, and freshwater discharge and flushing rates are high (Chan and Hamilton 2001). By contrast, an upstream tributary of the Swan River estuary, the Canning River, has experienced frequent and severe blooms of the freshwater cyanobacterium Anabaena circinalis since the early 1990s (Hosja and Deeley 1994; 10.1071/MF02090

1323-1650/03/020139

140

Marine and Freshwater Research

Vincent 2001). The intrusion of saline water up the Canning River is prevented by a weir, which also reduces turbulence. Recurrent toxic cyanobacterial blooms were also a major problem in the Peel-Harvey Estuary, 80 km south of the Swan River, throughout the 1980s (Lukatelich and McComb 1986). The species responsible for these blooms was Nodularia spumigena, a cyanobacterium that is tolerant of brackish conditions. The Peel-Harvey Estuary is shallow (mean depth ∼1 m) and, until 1994, had a narrow connection with the ocean that strongly restricted tidal flushing. Freshwater conditions persisted well into spring and facilitated the initiation of Nodularia blooms (Hearn and Robson 2000, 2001). Remedial measures completed in 1994 created an artificial channel between the estuary and the ocean and have evidently prevented the recurrence of Nodularia blooms by increasing the tidal prism and reducing the duration of freshwater conditions (D. A. Lord and Associates 1998). It could be speculated that only an unusual set of environmental conditions could initiate a freshwater cyanobacterial bloom in the Swan River estuary. In this paper we describe these conditions, the result of which triggered a large bloom of M. aeruginosa in February 2000. Cell densities peaked at more than 100 000 cells mL−1 (more than 1 × 106 cells mL−1 at some sites) three weeks after heavy summer rainfall occurred throughout the catchment, flushing the surface layer of the estuary with fresh water. At this time the Swan River was closed from recreational activity and fishing for 10 days because tests showed that the bloom was toxic (Orr et al. 2000). Little quantitative information is available on the effect of salinity on doubling rates of M. aeruginosa. Otsuka et al. (1999) reported net growth at salinities up to 7, whereas Orr et al. (2000) found visible signs of stress and increased lysis at 10, but not at 5, in M. aeruginosa cells isolated from the bloom in the Swan River estuary. In view of the potential importance of relatively small changes in salinity on the growth of M. aeruginosa in the Swan River estuary, one of the objectives of this study was to contribute more detailed information on the effects of salinity on M. aeruginosa. This information was used to assess quantitatively the role of salinity in the development and subsequent decline of the bloom that occurred in the Swan River estuary and to test the hypothesis that salinity was a key controlling variable in the bloom dynamics. Materials and methods Study site The Swan River estuary (Fig. 1) is located on the Swan Coastal Plain, Western Australia. It flows through the city of Perth and covers an area of about 52 km2 , with a maximum depth of 21 m, mean depth of 6 m and a catchment area of 121 000 km2 (Viney and Sivapalan 2001). The estuary is subject to nutrient inputs associated with inorganic suspended sediments (mostly through the Avon River), fertilizer inputs from agriculture, viticulture, public parks and domestic gardens, and, in some rural catchments, domesticated animal wastes (Peters and Donohue

B. J. Robson and D. P. Hamilton

Beach observation sites Boat observation sites Routine sampling sites

Success Hill Ron Courtney Island

Swan River

Transect pathway for salinity and oxygen profiles

Maylands

Perth The Narrows

Indian Ocean

Blackwell Reach

Nile St

St Johns

AREA A AREA B

Lower Estuary

AREA C AREA D

Upper Estuary

N 0

5 kilometres

Armstrong Spit

Canning River

Fremantle

Fig. 1. The Swan River estuary, showing observation sites during the bloom.

2001). There is also a large pool of nutrients that could be mobilized from the bottom sediments of the estuary (Douglas et al. 1997). Mean annual rainfall varies from ∼870 mm in Perth to 10) from the upper estuary. In the lower estuary it produced a surface water layer of ∼4 m with relatively low salinity (S < 7) (Fig. 5d, e).

35 5 Jan. 2000 30 10 Jan. 2000

25 17 Jan. 2000

20

24 Jan. 2000

31 Jan. 2000

15

16 Feb. 2000

10

21 Feb. 2000

5

28 Feb. 2000 0

5

10

15

20

Distance upstream from Narrows (km)

Fig. 5. Vertical cross-section of observed salinity in the estuary from Fremantle (−5 km) to Guildford (20 km) during the period of interest, interpolated between available data profiles shown as crosses at sampling stations along the estuary.

Flow event induces estuarine cyanobacterial bloom

7000

Area A Area B

6000

Area C

5000

Area D

4000

Trend

3000 2000 1000 0 5 Feb.

Water quality

10 Feb.

15 Feb.

20 Feb.

25 Feb.

Fig. 6. Cyanobacteria cell counts in the estuary from 7 to 23 February 2000. Observations are grouped by area of the estuary (see Fig. 1) and exponential regressions are shown for the growth period (r 2 = 0.61, P < 0.05) and the subsequent period of decline (r 2 = 0.21, P < 0.05). Outliers (a data point in Area A on 7 February where no cyanobacteria were observed and a single observation of cyanobacteria in excess of 1.1 × 1012 cells m−2 near the shore in Area B on 12 February) have been excluded from the regression analyses. Surface cyanobacterial density (cells mL⫺1)

Regular monitoring at eight stations (Fig. 1) found no cyanobacterial cells on 5 January. Cell counts on 17 January (before the main rainfall event) were zero in the lower reaches, but were 1212 cells mL−1 at Ron Courtney Island, 1525 cells mL−1 at Kingsley Drive and 11 767 cells mL−1 at Success Hill, that is, at stations in the upper estuary. Diatoms and dinoflagellates still dominated the assemblage in the lower reaches at this time, while chlorophytes dominated the upper reaches, with cell counts ranging from 31 765 cells mL−1 at Success Hill to 58 782 cells mL−1 at Ron Courtney Island. The next sampling day was 31 January (after the major inflow event), at which time cell counts for all groups were relatively low ( 0.4 mg L−1 and [NO3 ] > 2 mg L−1 ) immediately following the major inflow event. At this time water temperatures were high (>25◦ C), river inflows had flushed the upper estuary and created a surface intrusion overlying marine water in the lower estuary (Fig. 5), and irradiance was high. On 25 January cyanobacterial concentrations at 10 estuary sampling stations ranged from 0 to 60 cells mL−1 . Fig. 6 shows the mean daily cyanophyte cell counts from depthintegrated samples taken during the period of relatively intense sampling (6–23 February) in each of the areas denoted in Fig. 1. Cell densities from morning surface samples are shown in Fig. 7. Cell counts of M. aeruginosa peaked on 16 February, about three weeks after the maximum river discharge. Counts exceeded 1 × 106 cells mL−1 at some stations but declined rapidly after 17 February. Exponential trend lines applied to the integrated measurements correspond to an average growth rate of 0.35 day−1 (r 2 = 0.61, P < 0.01) during the growth phase of the bloom (7–16 February). During the period of bloom decline (17–23 February), depth-integrated M. aeruginosa cell counts fell at a rate of 0.25 day−1 . The lower r 2 value (0.24, P < 0.05) of the fit during this phase largely reflects the spatial variability of the bloom, with more rapid decline in the lower estuary. Qualitatively, Fig. 6 suggests that the decline in M. aeruginosa was considerably more rapid near the mouth of the estuary

143

8000 Cyanobacteria (cells mL–1)

By mid-February, river discharge had decreased substantially (Fig. 4). Salinity had increased to above 10 in most of the estuary, with re-entry of water of marine origin, driven as a gravity current under the influence of reduced resistance from discharge of freshwater and the passage of a low barometric pressure front that elevated ocean water levels. Barometric forcing strongly influences the incursion of marine water into the Swan River estuary and can displace the location of the salt wedge by around 10 km in one day compared with a typical tidal excursion of 2–4 km (Hamilton et al. 2001).

Marine and Freshwater Research

1 × 109 1 × 108 1 × 107 1 × 106 1 × 105 1 × 104 1 × 103 1 × 102 1 × 101 1 × 100 5 Feb.

9 Feb.

13 Feb.

17 Feb. 21 Feb.

25 Feb.

Fig. 7. Surface cyanobacteria density at random locations (beach and boat observation sites indicated in Fig. 1) in the Swan River estuary from 10 to 25 February 2000.

(Area A in Fig. 1) than in the middle reaches (Areas B and C). This decline is consistent with increasing salinity, because the salt wedge propagated upstream from the mouth of the estuary during the period of bloom decline (Fig. 5). Salinity response In the laboratory experiments, M. aeruginosa grew optimally at salinities up to 4, above which growth rate declined to zero at salinity ∼25 (Fig. 8). These results imply a maximum doubling rate, µmax , of 1.2 day−1 . This is significantly higher than values previously reported in the literature (e.g. 0.48 day−1 , Reynolds et al. 1981; 0.91 day−1 , Ward and Wetzel 1980). Robarts and Zohary (1987) cite a range of growth rates from 0.25 (Krüger and Eloff 1978) to 0.81 day−1 (Van der Westehuizen and Eloff 1985). Reynolds (1984), however,

144

Marine and Freshwater Research

B. J. Robson and D. P. Hamilton

0.7

trial 1 trial 2 trial 3 u = 1.2f (S) 0.2/f (S)

1 0.8

Microcystis

0.6 freshwater diatoms

0.5 DIN (mg L−1))

Net doubling rate (day⫺1)

Highest doubling rate at 20ºC

(a)

1.2

0.6 0.4

0.4

marine diatoms

0.3 0.2

0.2 chlorophytes

0.1

0 0

0.2 0

5

10

15 20 Salinity

25

30

dinoflagellates

0

5

10

15 Salinity

35

(b)

Modelling Succession Equation 2 was applied to determine which phytoplankton groups characterized inTable 1 had the greatest doubling rates at salinities ranging from 0 to 30 and at DIN concentrations of 0–0.7 mg L−1 . (Above 0.7 mg L−1 , increasing nitrogen concentration has little effect on which group dominates.) It was assumed that phosphorus was replete and light was optimal. Results for temperatures of 20 and 30◦ C are shown in Fig. 9. Water temperatures in the Swan River estuary in summer generally approach 30◦ C; however, in most years salinity exceeds 30 through most of the estuary (Kurup et al. 1998). Hence, in a ‘typical’ summer, these conditions indicate dominance of marine diatoms and dinoflagellates (see Fig. 9b). In winter, salinities are lower but temperatures are also lower, and this prevents substantial growth of M. aeruginosa. In this case conditions approximate the left side of Fig. 9a, supporting the growth of chlorophytes except when DIN concentrations are very low. Freshwater diatoms also feature in the usual winter phytoplankton succession, primarily because

25

30

Highest doubling rate at 30ºC 0.7

Fig. 8. Effect of salinity on net doubling rate of M. aeruginosa in three laboratory trials, with modelled salinity response. Salinity parameter values corresponding to these laboratory observations are given in Table 2.

0.6

dinoflagellates

0.5 DIN (mg L−1))

points out that phytoplankton growth rates decrease with increasing cell or colony size. Microcystis aeruginosa generally occurs as solitary cells in laboratory studies, but the value reported by Reynolds et al. (1981) is for the colonial form, which might explain the difference from the growth rate observed in the present study. Another factor that might play a role in the different doubling rates obtained by different studies is the technique used to measure growth. In this study, fluorescence (and hence chlorophyll a concentration) was measured. Van der Westehuizen and Eloff (1985), for example, measured turbidity.

20

Microcystis 0.4 chlorophytes marine diatoms

0.3 day 1–13

0.2 0.1

day 43–57 M. aeruginosa

0 0

5

none

10

15 Salinity

20

25

30

Fig. 9. Phytoplankton groups expected to be fastest growing over a range of salinities and DIN concentrations at (a) 20◦ C and (b) 30◦ C. Approximate observed salinities and nitrogen concentrations within the estuary during the period of interest are shown as dotted outlines in (b).

of their advection into the estuary from tributaries (Chan and Hamilton 2001). In February 2000, water temperatures approached 30◦ C and salinities fell below 5 through much of the estuary. These conditions were ideal for proliferation of M. aeruginosa (see Fig. 9b). It is possible that another cyanophyte, such as Anabaena flos-aquae, could have developed similarly in the estuary. Like M. aeruginosa, Anabaena favours freshwater, phosphorus-rich conditions and can form buoyant scums. Anabaena has a similar specific growth rate to M. aeruginosa under optimal conditions (0.99 day−1 ; Lee and Rhee 1999). Unlike M. aeruginosa, however, Anabaena shows optimal growth in the range 12.5–22◦ C, with growth severely constrained at 30◦ C (Lehtimaki et al. 1997; Rapala et al. 1997), explaining dominance by M. aeruginosa in the Swan River in February 2000. Growth of Microcystis aeruginosa Salinity strongly inhibited growth of M. aeruginosa [i.e. f (S) < 0.5] at the start of January, but was near optimal

Flow event induces estuarine cyanobacterial bloom

Marine and Freshwater Research

145

Table 1. Parameter values used to model phytoplankton growth and loss rates Parameter

Description

Microcystis aeruginosa

Freshwater diatoms

Marine diatoms

Chlorophytes

Dinoflagellates

µmax KP

Maximal specific growth rate at 20◦ C (day−1 ) Half-saturation coefficient for phosphorus uptake (mg L−1 ) Half-saturation coefficient for nitrogen uptake (mg L−1 ) Optimal salinity Parameter for salinity response calculation (Griffin et al. 2001) Maximal salinity for salinity response calculation Standard temperature (◦ C) (defines the slope of the function in Fig. 3) Optimal temperature (◦ C) Maximal temperature (◦ C) Arrhenius coefficient for temperature function

– 0.006D

2.10A –

1.44A –

1.90B –

0.66C –

0.03E

0.06F

0.022G

0.05H

0.03I

– –

3J 8I

20J 8I

9K 8K

25L 3L

– 20

18J 15

18J 19

12K 20

29M 22

30N 35 1.08V

20P 28T 1.04W

26Q 32U 1.04X

27R 27 1.14Y

29S 34 1.09Z

KN Sopt β Smax Ts Topt Tmax θ

Maximal growth rates for miscellaneous diatoms in laboratory studies range from 0.4 day−1 (Wheeler et al. 1974) to 5 day−1 (Eppley et al. 1971; Furnas 1991). For Scenedesmus spp., a dominant marine diatom group in the Swan River estuary, maximal doubling rates at 24–25◦ C range from 1.32 to 2.84 day−1 (Reynolds 1984). The values given were calibrated within this range, comparing results from an expanded three-dimensional version of the model presented here with field data for the period 1995–98 by Chan et al. (2002). B The literature range (Chan et al. 2002) for µ max of Chlamydomonas, the dominant chlorophyte in the Swan River estuary, is 0.5 (Wheeler et al. 1974) to 3.8 day−1 (Jorgensen 1979). Reynolds (1984) gives 1.90 day−1 , the value used here. C Laboratory measurements of dinoflagellate growth rates range from 0.3 day−1 (Bjornsen and Kuparinen 1991) to 0.7 day−1 (Eppley et al. 1971; Chang and Carpenter 1991). The value given was obtained by Chan et al. (2002) through model calibration within this range. D From Holm and Armstrong (1981). E No literature values for K for M. aeruginosa were found, so a typical value for cyanobacteria was used. N F Bowie et al. (1985) give a literature range of 0.015–0.923 mg N L−1 for the K of unspecified diatoms. Model calibration (Chan et al. 2002) for the N Swan River estuary gave a value of 0.06 day−1 . G Literature values for K of marine diatom species found in the Swan River estuary range from 0.006 (Eppley et al. 1969) to 0.03 mg N L−1 (Bierman N et al. 1980). H Literature values for K of chlorophytes range from 0.001 (Scavia and Park 1976) to 0.15 mg N L−1 (Di Toro et al. 1971). N I Eppley et al. (1969) give a range of 0.015–0.053 mg N L−1 for the K of Gymnodinium spp., a common dinoflagellate in the Swan River estuary, and N 0.013–0.077 mg N L−1 for the KN of Gonyaulax spp., which is also found in the estuary. Chan et al. (2002) used a calibrated value of 0.06 mg N L−1 . J Diatom blooms in the Swan River estuary have been observed over a range of salinities from 4 to 38, with a peak at the extreme salinities and a decline in bloom frequency at intermediate salinities (7–12) (Chan and Hamilton 2001). These peaks reflect the dominance of two groups, estuarine (>12) and freshwater (