Introduction
In a survey of primary production conducted in the subarctic North Pacific (Larrance, Reference Larrance1971) and New York Bay (Malone, Reference Malone1976), a linear function was found between the integrated value of light intensity per day and the value of primary production per phytoplankton biomass in the water column. Falkowski (Reference Falkowski1981) also found a linear relationship between the two quantities in the New York Bight and defined the value obtained as the light utilization efficiency in the water column, Ψ. According to Platt (Reference Platt1986) and Platt et al. (Reference Platt, Sathyendranath, Caverhill and Lewis1988), the values of Ψ reported previously were between 0.31 and 0.66 gC gChl-a−1 mol photons−1 m2, with only a ~2-fold fluctuation. When the results obtained from subsequent surveys of coastal, open-ocean and freshwater areas are considered, the range of Ψ appears larger than that previously identified, ranging from 0.07–1.92 gC gChl-a−1 mol photons−1 m2 (e.g. Falkowski, Reference Falkowski1981; Shiomoto, Reference Shiomoto2000; Forget et al., Reference Forget, Sathyendranath, Platt, Pommier, Vis, Kyewalynga and Hudon2007; see Table 1 for details). Most information regarding Ψ has been obtained from coastal and oceanic areas, and determination of Ψ in lakes has been scarce. Furthermore, little is known about the seasonal variation in Ψ.
Table 1. Summary of Ψ values previously reported worldwide
‘Slope’ means that the Ψ value was a slope of the regression line obtained by plotting Chl-normalized primary production against daily irradiance. ‘Each’ means that the Ψ value was calculated using each datum according to Falkowski's definition of Ψ.
a From Platt (Reference Platt1986).
b From Platt et al. (1988).
c This may be an overestimate potentially due to a systematic error in solar radiation (Platt, Reference Platt1986).
Parameter Ψ can be calculated as a ratio for individual observations or by using a linear regression for multiple observations (e.g. Falkowski, Reference Falkowski1981; see Table 1 for details). Because primary production (photosynthetic rate) does not increase continuously in proportion to light intensity, Ψ cannot be constant across a wide range of light intensities (Platt, Reference Platt1986; Platt et al., Reference Platt, Sathyendranath, Caverhill and Lewis1988). For weak or strong light, Ψ deviates from a constant value, whereas within a range of light of a certain intensity, Ψ seems to be nearly constant (Falkowski & Raven, Reference Falkowski and Raven1997; Imai et al., Reference Imai, Nojiri, Tsurushima and Saino2002; Forget et al., Reference Forget, Sathyendranath, Platt, Pommier, Vis, Kyewalynga and Hudon2007; Honda et al., Reference Honda, Sasaoka, Kawakami, Matsumoto, Watanabe and Dickey2009; Nosaka et al., Reference Nosaka, Isada, Kudo, Saito, Hattori, Tsuda and Suzuki2014). If the value of Ψ is known, primary production can be calculated from chlorophyll-a present and light intensity, an estimation that is easier to perform than the actual measurement of primary production. In this case, it is important to evaluate how reliably primary production can be estimated with a constant Ψ.
Our study area, Saroma-ko Lagoon, is a brackish lake connected to the Okhotsk Sea, and environmental factors (temperature, salinity and nutrients) fluctuate markedly throughout the year (Tada et al., Reference Tada, Kurata and Nishihama1993; Shibanuma et al., Reference Shibanuma, Kajihara and Miyake1995). Scallop aquaculture is a thriving local industry, and phytoplankton are an important source of food for the scallops. Consequently, monitoring primary production is important for determining the amount of scallop cultivation that can be sustained (aquaculture allowance). We monitored the primary production to estimate feed availability for scallops in this brackish lake. We calculated Ψ using these data and clarified its seasonal variation and the factors that cause this variation, regarding which little is known currently. Furthermore, we calculated the primary production from chlorophyll-a and light intensity assuming a certain value of Ψ. Then, we evaluated the reliability of the calculated primary production. This study contributes towards developing a simple estimation of primary production, as the commonly used method is complicated (e.g. Hama et al., Reference Hama, Miyazaki, Ogawa, Iwakuma, Takahashi, Otsuki and Ichimura1983). Examining the reliability of this simple method will also be of great benefit to fisheries in Saroma-ko Lagoon as well as in other locations.
Materials and methods
Monthly observations were made at Station 21 (water depth: ~15 m) located near the central part of Saroma-ko Lagoon from January–December from 2009–2018 (Figure 1). From January to March, the ice-forming season, observations were made in ice-free areas. In the 10-year survey, one survey each was conducted in January, February and March, and multiple surveys (4–7) were conducted each month from April–December. Water samples were collected between 10:00 am and 12:00 am from four depths corresponding to 100%, 37%, 17% and 1% light depths using an acid-cleaned Niskin X sampler hung on a non-metal wire (polyethylene). Light depths were determined 30 min before each sampling period using a quantum sensor (LI-190R, Li-Cor). Samples were brought back from the field to the laboratory under dark conditions and subsequently used for incubation experiments in the laboratory.
Fig. 1. Location of the sampling stations in Saroma-ko Lagoon. The water depths of the stations were ~14 m.
Vertical profiles of temperature and salinity were determined using an Alec memory conductivity-temperature-depth probe (ASTD 687). Nutrient concentrations were measured with a BLTEC Auto Analyzer SWAAT after storing samples at −70°C.
The sample water for measuring chlorophyll-a (Chl-a) concentration was filtered using a Whatman GF/F filter (pore size: ~0.7 μm). Chl-a was extracted with N, N-dimethylformamide (Suzuki & Ishimaru, Reference Suzuki and Ishimaru1990). Chl-a sample concentrations were measured via fluorometry using a Turner Designs 10-AU fluorometer, according to the methods of Welschmeyer (Reference Welschmeyer1994). Calibration of the fluorometer was performed using a commercially available Chl-a standard (MERK).
Primary production was measured via the simulated in situ method using the 13C uptake technique (Hama et al., Reference Hama, Miyazaki, Ogawa, Iwakuma, Takahashi, Otsuki and Ichimura1983). Samples (1 litre) were dispensed into two 1 litre polycarbonate bottles from each light depth and enriched with 13C-NaHCO3 (99 atom% 13C; Isotec) to ~10% of the total inorganic carbon in ambient water within 3–4 h of sample collection. Bottles of seawater samples inoculated with 13C-NaHCO3 were placed in a water tank installed on the roof of the Tokyo University of Agriculture building at irradiances corresponding to the depths at which the samples were taken, using black mesh screens, and incubated under sunlight for 24 h at the same water temperatures as those at which the samples were collected. Photon fluxes [photosynthetically available radiation (PAR) 400–700 nm] were monitored every 5 min during incubation using a quantum sensor (LI-190R, Li-Cor). Following incubation, seawater samples in the bottles were immediately filtered. The Whatman GF/F filters used for measuring primary production were pre-combusted at 450°C for 4 h. Before isotope analysis, the filters were treated with HCl fumes for 4 h to remove inorganic carbon and were completely dried in a vacuum desiccator. The isotopic ratios of 13C to 12C and particulate organic carbon content were determined using a mass spectrometer (ANCA SL or INTEGRA 2CN, CerCon). The total inorganic carbon in the water was measured using an infrared analyser (TOC 5000, Shimadzu). Primary production was calculated according to the equation described by Hama et al. (Reference Hama, Miyazaki, Ogawa, Iwakuma, Takahashi, Otsuki and Ichimura1983), and the values of the two polycarbonate bottles were averaged.
Primary production is affected not only by light but also by phytoplankton biomass. To compare the phytoplankton productivity, we calculated the primary production divided by the integrated amount of Chl-a within the euphotic zone and called it the Chl-normalized production, probably showing the mean of the Chl-normalized production at any depth within the euphotic zone. The Chl-normalized production in this study was also considered an index of the growth rate of phytoplankton.
Falkowski (Reference Falkowski1981) defined Ψ as follows:
$$\Psi = \displaystyle{{\hbox{\,primary}\,\hbox{\,production}} \over {\hbox{Chl}-\hbox{a} \times \hbox{PAR}}}\comma \;$$where primary production is the depth-integrated primary production within the euphotic zone (mgC m−2 day−1), Chl-a is the depth-integrated Chl-a within the euphotic zone (mg m−2) and PAR (mol photons m−2 day−1) is the daily surface irradiance. Ψ was calculated separately for each day. Equation (1) can be transformed into the following equation:
$$\hbox{Primary}\,\hbox{\,production} = \Psi \times \hbox{Chl}-\hbox{a} \times \hbox{PAR}.$$For phytoplankton species examination, a 1 litre seawater sample was taken from a depth of 6 m at Station 21 or Station 22 using a Kitahara's water bottle from April–December from 2008–2015, and fixed with neutralized formalin (2% vol/vol). Surveys were conducted 3–9 times each month. These samples were concentrated to 10 mL by siphoning off the excess solution after ~1 day of settling. Aliquots (1/20–1/10 of each sample) were examined, depending on the cell density, using a light microscope (BX41, Olympus) equipped with interference and phase contrast optics. Phytoplankton were identified at the species level when possible, or otherwise grouped by genera. The data on phytoplankton species composition were provided by the Saroma-ko Aquaculture and Fisheries Cooperative.
Results
Seasonal variations in physical and chemical environments
Large seasonal variation was observed in the mean water temperature (Figure 2A). The temperature within the euphotic zone was ~−1°C from January–March, the ice-forming season, whereas a temperature of ~5°C was observed in April, the melting season. The temperature increased from April–August and the maximum temperature, ~20°C, was observed in August. The temperature decreased from September–December and was ~2°C in December.
Fig. 2. Seasonal variations in the mean and standard deviation of the temperature within the euphotic zone (A), salinity within the euphotic zone (B), nitrate within the euphotic zone (C) and photosynthetically available radiation (PAR) during incubation (D). The mean values and standard deviations for each month were calculated using all data obtained in the 10-year observation period. For the temperature and salinity, the data at intervals of 0.5 m within the euphotic zone were used; for the nitrate, the data of four light depths within the euphotic zone were used. The numbers in parentheses represent the number of datasets. The PAR in January may have been overestimated owing to equipment malfunction.
Furthermore, there was a large seasonal variation in the mean salinity (Figure 2B). The salinity in April, the melting period, was somewhat lower than that from January–March, the ice-forming period. Salinity increased from May, reaching a maximum of 33.3 in October and decreasing from November. Lower salinity was observed in winter and spring, and higher salinity in summer and autumn. The higher and lower salinity is due to the influx of the Soya Warm Current water (temperature: 7–20°C; salinity: 33.6–34.3) (Takizawa, Reference Takizawa1982) and the East Sakhalin Current water (water temperature: <7°C; salinity: <32.0) (Takizawa, Reference Takizawa1982), respectively (Shibanuma et al., Reference Shibanuma, Kajihara and Miyake1995); cold and warm currents alternated according to season.
There was a large seasonal variation in the mean nitrate concentration (Figure 2C). The nitrate concentration was as high as ~9 μmol l−1 in January and February, the ice-forming period, but was extremely low in March, ~1 μmol l−1. Thereafter, the concentrations were lower than 2 μmol l−1 until October, and were particularly depleted from June–September. The values were high from November and were ~6 μmol l−1 in December. Nitrate concentrations were relatively high from winter to spring and low from spring to autumn. These results reflect the change in water mass.
The mean PAR tended to be lower from October–December (10–15 mol photons m−2 day−1) compared with February–September (~20–30 mol photons m−2 day−1), except for extremely high values in January (Figure 2D). A markedly high PAR was seen in January; the high value may have been observed because of a malfunction in the PAR sensor. There was a significant difference in PAR between the period from February–September and from October–December (Mann–Whitney U-test, P < 0.001). Furthermore, there was no clear seasonal variation in PAR from February–September. Although the weather on the survey day had a significant effect on PAR, it tended to be lower from late autumn to early winter.
Seasonal variations in chlorophyll-a, primary production and chlorophyll-normalized production
The seasonal variation in the mean value of the integrated amount of Chl-a within the euphotic zone is shown in Figure 3A. The obtained amount varied from 9–108 mg m−2, with a 12-fold fluctuation. Although there was only one data point each month from January–March, high values were observed from February–April, from late winter to early spring. The high values reflected the spring bloom in Saroma-ko Lagoon. Excluding the bloom months, the Chl-a amount tended to be lower in November, December and January, midwinter.
Fig. 3. Seasonal variations in the mean and standard deviation of the chlorophyll-a standing stock integrated within the euphotic zone (mg m−2) (A), primary production integrated within the euphotic zone (mgC m−2 day−1) (B), chlorophyll-normalized production within the euphotic zone (mgC mgChl-a−1 day−1) (C) and Ψ (gC gChl-a−1 mol photons−1 m2) (D). The mean values and standard deviations for each month were calculated using all data obtained in the 10-year observation period. The numbers in parentheses represent the number of datasets.
The seasonal variation in the mean phytoplankton primary production is shown in Figure 3B. The obtained primary production was 33–786 mgC m−2 day−1, with a 24-fold fluctuation. The highest value obtained over 10 years was 1600 mgC m−2 day−1 in April 2012. The seasonal variation was almost the same as that of Chl-a. However, a relatively high value was observed in September.
The seasonal variation in the mean Chl-normalized production is shown in Figure 3C. The obtained production was 3.7–15.3 gC gChl-a−1 day−1, with a 4-fold fluctuation. The production tended to be low from January–April (mean ± SD: 5.3 ± 2.6 gC gChl-a−1 day−1) and high from May–September (11.8 ± 8.0 gC gChl-a−1 day−1). The mean value (± SD) from October to December was 7.9 (5.7) gC gChl-a−1 day−1. The mean value from October–December was half of that from January–April and May–September. Significant differences were found among the means in the three periods (Kruskal–Wallis test, P < 0.05).
Seasonal variation in water-column light utilization efficiency, Ψ
The seasonal variation in Ψ is shown in Figure 3D. The monthly means ranged from 0.05–0.71 gC gChl-a−1 mol photons−1 m2, with a 14-fold fluctuation. Ψ values tended to be low from January–April (mean ± SD: 0.21 ± 0.14 gC gChl-a−1 mol photons−1 m2) and high from October–December (0.70 ± 0.36 gC gChl-a−1 mol photons−1 m2). The mean value (±SD) from May to September was 0.49 (0.30) gC gChl-a−1 mol photons−1 m2. The mean value from May–September was half of that from January–April and October–December. Significant differences were found among the means in the three periods (Kruskal–Wallis test, P < 0.05).
Seasonal variations in phytoplankton
The seasonal variation in the number of phytoplankton cells is shown in Figure 4A. The mean number of cells obtained from April–December was 6.0 × 104 to 2.7 × 106 cells l−1, with a ~45-fold fluctuation. The highest value was observed in April, and in September it was as high as in April. The lowest values were observed in November and December. From April–October, Bacillariophyceae (diatoms) accounted for more than 80% of the total cell number (Figure 4B). In November and December, when the total number of cells was small, diatoms accounted for ~40%, and Cryptophyceae accounted for another 40% of the total cell number.
Fig. 4. Seasonal variations in the mean and standard deviation of the total cell numbers of phytoplankton (A) and the mean species structure of phytoplankton (B). The mean values and standard deviations for each month were calculated using all data obtained at Station 21 and Station 22 from 2008 to 2015. The numbers in parentheses represent the number of datasets.
Determination of primary production using Ψ
If the value of Ψ is constant, it is possible to determine depth-integrated primary production by applying the depth-integrated Chl-a amount and daily solar radiation (PAR) to equation (2). Alternatively, it is possible to use a certain value of Ψ, as the representative value in a study area. In either case, it is essential to estimate the reliability of the primary production thus determined. In this study, Ψ ranged from 0.05–1.42 gC gChl-a−1 mol photons−1 m2 (average ± SD: 0.50 ± 0.33 gC gChl-a−1 mol photons−1 m2; N= 56), with a 28-fold fluctuation, indicating that Ψ is not constant in Saroma-ko Lagoon. The average is affected by remarkably high and low values; however, such an effect is not found on the median. Therefore, we used the median as the representative value in this study, which was 0.44 gC gChl-a−1 mol photons−1 m2.
Primary production was calculated by applying the median value for Ψ (0.44 gC gChl-a−1 mol photons−1 m2), the depth-integrated Chl-a within the euphotic zone and the daily surface irradiance (PAR) to equation (2). A significant correlation was found between the measured and calculated values (Figure 5A). In addition, the ratio between the calculated and measured values of primary production ranged from 0.31–8.89 (Figure 5B). The results of an outlier test revealed four data points as outliers (P < 0.05). The Ψ values excluding these four outliers ranged from 0.13–1.42 gC gChl-a−1 mol photons−1 m2; the median value was 0.45 gC gChl-a−1 mol photons−1 m2. Accordingly, the primary production was calculated again using the median value of 0.45 gC gChl-a−1 mol photons−1 m2. A significant linear relationship was also observed between the calculated and measured values (Figure 5C). The ratios of the calculated values to the measured values ranged from 0.31–3.45, with an average of 1.18 (Figure 5D). This implies that when primary production is calculated using the median value of Ψ, 0.45 gC gChl-a−1 mol photons−1 m2, the ratio of the computed primary production to the observed production is within the range of 0.3–3.5, with an average value of 1.2.
Fig. 5. Relationship between measured primary production (mgC m−2 day−1) and calculated primary production (mgC m−2 day−1) (A, C) and frequency distribution of the ratio of calculated primary production to measured primary production (B, D). All data were used in A and B, and the data excluding the four outliers were used in C and D. The straight lines in A and C were obtained by the least squares method. n: number of datasets. Mean, SD, maximum, minimum and median values of Ψ (gC gChl-a−1 mol photons−1 m2) are inserted in B and D.
Discussion
Platt (Reference Platt1986) and Platt et al. (Reference Platt, Sathyendranath, Caverhill and Lewis1988) calculated Ψ as the slope of the regression line between the production per unit biomass within the euphotic zone and the surface light. We calculated Ψ using individual data for each survey day. There are two methods for obtaining Ψ (Table 1), which have not yet been discussed. Ψ was defined as the cumulative primary production divided by the cumulative Chl-a within the euphotic zone and daily surface irradiance (Falkowski, Reference Falkowski1981). When primary production is determined with the slope of the regression line being Ψ, if the light is weak or strong, the estimated value deviates from linearity, and a correct estimation of primary production is not possible. Platt (Reference Platt1986) examined the deviation from linearity and the required correction in detail. If the target is to find the representative value of Ψ in a study area, it is important to find the slope of the regression line. In contrast, the goal of our study was to clarify the seasonal variation in Ψ in the survey area. Therefore, instead of using the slope of the regression line as Ψ, individual data were used to calculate Ψ.
In Saroma-ko Lagoon, high Ψ values were observed from October–December, whereas low values were observed between January and April (Figure 3D), implying a seasonal variation in Ψ values. The water mass, nitrate concentrations and phytoplankton species composition differed between October and November–December (Figures 2A–C & 4). In contrast, the PAR values from October–December, 10–15 mol photons m−2 day−1, were almost equal, and were lower than those in other months in which PAR values were more than 20 mol photons m−2 day−1 (Figure 2D). The high Ψ from October–December could be attributed to the low PAR. High Ψ values have been reported at low PAR values (Falkowski & Raven, Reference Falkowski and Raven1997; Shiomoto, Reference Shiomoto2000; Imai et al., Reference Imai, Nojiri, Tsurushima and Saino2002; Honda et al., Reference Honda, Sasaoka, Kawakami, Matsumoto, Watanabe and Dickey2009; Nosaka et al., Reference Nosaka, Isada, Kudo, Saito, Hattori, Tsuda and Suzuki2014).
Low Ψ values were observed between January and April, when the East Sakhalin Current water was present (Figure 3D). The same cold water was also observed in May (Figure 2A, B). The nitrate concentrations, PAR and phytoplankton species composition were almost the same between April and May (Figures 2C, D & 4). This means that these factors were not related to the low Ψ values observed from January–April. In contrast, the Chl-normalized production was low from January–April, probably because of the remarkably high Chl-a biomass caused by the spring bloom excluding January (Figure 3A, C). The low Ψ from January–April could be attributed to the low Chl-normalized production. Low Chl-normalized production values during phytoplankton blooms have been reported both in coastal (Maita & Odate, Reference Maita and Odate1988) and oceanic (Yoshie et al., Reference Yoshie, Suzuki, Kuwata, Nishioka and Saito2010) areas. The Chl-normalized production was found to be lower during the stationary phase when phytoplankton biomass was abundant (Glover, Reference Glover1980). Most of our data may reflect the stationary phase in the spring bloom.
The Ψ values in Saroma-ko Lagoon ranged from 0.05–1.42 gC gChl-a−1 mol photons−1 m2. A review of previously reported Ψ values in various areas and seasons showed that they were in the range of 0.07–1.92 gC gChl-a−1 mol photons−1 m2, excluding the value of Malone (Reference Malone1976) (Table 1). Almost the same result was found in this study, meaning that the Ψ of Saroma-ko Lagoon covered the Ψ of the entire hydrosphere, although the environmental factors and the phytoplankton species composition in Saroma-ko Lagoon might differ from those of other locations. The environmental factors in the lagoon showed large seasonal variations (Figure 2). The environmental factors in winter–spring and in summer–autumn were similar to those in the subarctic and subtropical regions, respectively. Accordingly, we were able to collect Ψ across a wide range of environmental conditions during the survey at Saroma-ko Lagoon. However, if Ψ is unique to the study area, such a result should not be obtained in this study. Furthermore, the representative value in Saroma-ko Lagoon (0.45 gC gChl-a−1 mol photons−1 m2) is nearly equal to the Ψ values obtained as the slope for the regression lines by plotting the Chl-normalized production against daily irradiance in various regions (mostly 0.3–0.7 gC gChl-a−1 mol photons−1 m2; Table 1). The Ψ values obtained as slopes are considered representative Ψ values of these regions. Consequently, it is possible that Ψ has no regional characteristics and fluctuates depending on the environment, especially light intensity and phytoplankton growth conditions. In addition, by combining the previously reported data and that obtained in Saroma-ko Lagoon, Ψ values in the entire hydrosphere would range from 0.05–2 gC gChl-a−1 mol photons−1 m2. This indicates that primary production could be estimated from the in situ light (<50 mol photons m−2 day−1; see below) and Chl-a 0.3–3.5 times the actual value by using the Ψ value of 0.45 gC gChl-a−1 mol photons−1 m2 regardless of location. The median Ψ value in Saroma-ko Lagoon (0.45 gC gChl-a−1 mol photons−1 m2) may be representative of the entire hydrosphere. This value matches closely with the global Ψ value (0.44 gC gChl-a−1 mol photons−1 m2) proposed by Behrenfeld & Falkowski (Reference Behrenfeld and Falkowski1997). Using the value obtained in Saroma-ko Lagoon, the primary production in various areas can easily be obtained because there is no need for complicated determination of primary production using isotopes. Therefore, it is necessary to verify whether the estimated values in various areas can be obtained with the same accuracy as that in Saroma-ko Lagoon.
There were four outliers in the ratio of calculated to measured primary production. There were substantial differences in physical, chemical and biological factors between the four survey days when the outliers were obtained (Table 2). In contrast, on the four outlying survey days, the PAR exceeded 50 mol photons m−2 day−1, which was a common feature among the four days. Consequently, in Saroma-ko Lagoon, when daily PAR values are less than 50 mol photons m−2 day−1, it is possible to determine the primary production at a minimum of 0.3 times and a maximum of 3.5 times that of the measured value by applying the Ψ of 0.45 gC gChl-a−1 mol photons−1 m2 to equation (2). This value is useful for monitoring the primary production in Saroma-ko Lagoon and hence, for the sustainable development of fisheries.
Table 2. Physical, chemical and biological characteristics at the observation date when four outliers were found
ND, no data.
Data for temperature, salinity and nitrate were obtained within the euphotic zone. Data for PAR were obtained at the surface. Data for production and Chl-a were the amounts integrated within the euphotic zone. Data for dominant phytoplankton were obtained at 6 m.
Scallop aquaculture is important in Saroma-ko Lagoon. To avoid excessive aquaculture and maintain good water quality, scallop aquaculture is performed by establishing an aquaculture allowance. The most important factor in determining the aquaculture allowance is phytoplankton primary production because phytoplankton are important for the scallop diet. Accordingly, it is essential to constantly monitor the primary production to understand the feeding conditions of scallops. The results of this study greatly contribute not only to sustainable scallop aquaculture in Saroma-ko Lagoon, but also to the aquaculture of aquatic organisms feeding on phytoplankton in other areas.
Acknowledgements
We would like to thank the Saroma-ko Aquaculture and Fisheries Cooperative for their assistance in collecting samples and for funding this study. Furthermore, we are grateful to the cooperative for providing data on phytoplankton species.
