INTRODUCTION
A trend towards a shortening of the duration of ice coverage due to global warming has been reported in many rivers, ponds and lakes around the world (Schindler et al., Reference Schindler, Beaty, Fee, Cruikshank, DeBruyn, Findlay, Linsey, Shearer, Stainton and Turner1990, Reference Schindler, Bayley and Parker1996; Magnuson et al., Reference Magnuson, Robertson, Benson, Wynne, Livingstone, Arai, Assel, Barry, Card, Kuusisto, Granin, Prowse, Stewart and Vuglinski2000; Futter, Reference Futter2003). Because ice coverage inhibits the input of solar radiation into the water, this shortening results in increased light conditions in the water beneath the ice. In ice covered waters, ice algae account for a substantial proportion of primary production (Horner & Schrader, Reference Horner and Schrader1982; Arrigo et al., Reference Arrigo, Worthen, Lizotte, Dixon and Dieckmann1997; Gosselin et al., Reference Gosselin, Levasseur, Wheeler and Booth1997; Hattori et al., Reference Hattori, Kato and Saito2001; Lizotte, Reference Lizotte2001), although the potential capacity of primary production is generally low in ice algae compared to its phytoplankton counterparts (Rivikin et al., 1989; Irwin, Reference Irwin1990; Smith & Sakshaug, Reference Smith, Sakshaug and Smith1990; Ban et al., Reference Ban, Aikawa, Hattori, Sasaki, Sampei, Kudoh, Fukuchi, Satoh and Kashino2006). The shortening of the ice coverage period therefore implies a possible decrease in the relative importance of ice algae as primary producers, and an increase in the relative importance of phytoplankton. The shortening also implies an increase of wintertime primary production in the ice covered rivers, ponds and lakes. Changes in primary producers could have pronounced effects on the structure and function of a given aquatic ecosystem (e.g. Parsons et al., Reference Parsons, Takahashi and Hargrave1984; Lalli & Parsons, Reference Lalli and Parsons1993). Accordingly, global warming can possibly change ecosystems and biological production during winter and spring in ice-covered aquatic areas.
Saroma-ko Lagoon, with a surface area of 150 km2 and a maximum depth of 19.5 m, is located on the north-east coast of Hokkaido, Japan, and is connected to the Okhotsk Sea by two narrow channels. In total, 14 rivers flow into the lagoon. This lagoon is the largest of the lakes, marshes and lagoons that are characterized by complete ice coverage during winter in Japan. During the ice coverage period, sporadic studies in the past have shown that the chlorophyll-a concentration (phytoplankton biomass) in the water was low, on the order of 0.5 mg m−3 or less, under the ice (Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Satoh et al., Reference Satoh, Yamaguchi, Watanabe, Tanimura, Fukuchi and Aruga1989; Tada et al., Reference Tada, Kurata and Nishihama1993; Niimura et al., Reference Niimura, Ishimaru and Taguchi2000); production by ice algae was far greater than that by phytoplankton (Hattori et al., Reference Hattori, Kato and Saito2001). The spring bloom has been observed in the lagoon after disappearance of ice in April (Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Nishihama et al., Reference Nishihama, Kurata and Tada1989; Tada et al., Reference Tada, Kurata and Nishihama1993). Chlorophyll-a concentrations reached 5–20 mg m−3 in the spring blooms, although the blooms lasted for only one month (Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Nishihama et al., Reference Nishihama, Kurata and Tada1989; Tada et al., Reference Tada, Kurata and Nishihama1993). Although little is known about the primary production and species composition of phytoplankton during this bloom, diatoms dominated in both the ice algae and the phytoplankton communities in the lagoon in winter (Takahashi, Reference Takahashi1981; McMinn et al., Reference McMinn, Hattori, Hirawake and Iwamoto2008).
The duration of the period of complete ice coverage has tended to decrease since 1964 in Saroma-ko Lagoon, and years with incomplete ice coverage have been frequently observed since 1990 (e.g. Shirasawa et al., Reference Shirasawa, Daibo, Fujijoshi, Ishikawa, Kawamura, Leppäranta, Saloranta and Takatsuka2001; Aquaculture Fisheries Cooperation of Lake Saroma, http://homepage3.nifty.com/saromako/). Thus, the same trend that is being reported globally for ice-coved aquatic areas has also been noted in Saroma-ko Lagoon. In winter 2009, the lagoon was only partly covered by ice, at about 64% of its surface (see Figure 1), a condition that is considered to be a result of global warming. We carried out an emergency observation of the ice-free area of the lagoon during winter and spring 2009, and noted a wintertime diatom bloom that continued through to spring. Our observations suggest that the occurrence of these types of blooms will continue as incomplete ice coverage occurs frequently. This paper, although it reports only one series of observations, provides important details of the algal bloom that should be helpful in the prediction of the influence of global warming progression on ecosystems and biological production during the winter and spring in ice-covered aquatic areas, because incomplete ice coverage is rarely observed. In addition, since Saroma-ko Lagoon has been the site of successful shellfish culture operations for scallops and oysters (e.g. Kurata et al., Reference Kurata, Hoshikawa and Nishihama1991; Nishihama, Reference Nishihama1994) this paper will also allow the prediction of possible effects of global warming on aquaculture in the ice-covered aquatic areas.
Fig. 1. MODIS image on 3 March 2009, and the location of Station 21 (solid circle; 44°09′N 143°51′E) in Saroma-ko Lagoon. Observations were carried out on 16 February, 18 March and 17 April 2009. Black and white areas in the lagoon show ice-free and ice-covered areas, respectively. The percentage contribution of ice coverage was highest on 3 March and accounted for 64% of the whole surface area in the lagoon.
MATERIALS AND METHODS
Observations were conducted at Station 21 (44°09′ N 143°51′ E) with the depth of 15 m, located near the central part in Saroma-ko Lagoon, during February to April, 2009 (Figure 1). Water samples were collected between 8:00 am and 10:00 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). The light depths were determined 30 minutes before each sampling with a quantum sensor (Alec AL30-CMP).
Chlorophyll-a (Chl a) concentrations of total and size-fractionated samples were measured with fluorometry using a Turner Designs 10-AU fluorometer according to Welschmeyer (Reference Welschmeyer1994). Size separated samples (0.2–0.5 l) were prepared by filtration through Nuclepore filters with pore sizes of 10 and 2 μm and a Whatman GF/F (about 0.7-μm pore size: total). Chl a was extracted with N, N-dimethylformamide (Suzuki & Ishimaru, Reference Suzuki and Ishimaru1990). Calibration of the fluorometer was performed with a commercially available Chl a standard (Sigma-Aldrich Co.). Chl a concentrations for the <2, 2–10 and >10 μm fractions were obtained from the differences between the total and 2 μm, 2 and 10 μm, and total and 10 μm, respectively.
Primary production (photosynthetic rate) was measured by the simulated in situ method using the 13C uptake technique (Hama et al., Reference Hama, Miyazaki, Ogawa, Iwakuma, Takahashi, Otsuki and Ichimura1983). The samples (1 l) were dispensed into two 1-l polycarbonate bottles at each light depth and enriched with 13C-NaHCO3 (99 atom% 13C; Shoko Co. Ltd.) to about 10% of the total inorganic carbon in ambient water within 30 minutes after collection of the samples, and then incubated in situ for 24 hours. Dark bottle uptake is similar to the zero-time blank in the 13C technique (Shiomoto et al., Reference Shiomoto, Ishida, Tamaki and Yamanaka1998), and thus no dark bottles were used. Bottles of seawater samples inoculated with 13C-NaHCO3 were held in the coastal surface water in Saroma-ko Lagoon, at irradiances corresponding to the depths at which the samples were taken, using black mesh screens. The temperatures in the coastal surface water were about 2°C higher than the in situ temperatures (–1.5–5°C) of the samples that were collected. Shiomoto (Reference Shiomoto2009) showed that a 2°C increase relative to the in situ temperature (0.2°C) induced increases of about 20% in photosynthetic rates for small size phytoplankton and little increase for larger phytoplankton using samples from the coastal area of Hokkaido. The primary production obtained in this study was considered to be almost equal to the in situ values. Immediately following incubation, seawater samples in bottles were directly filtered through pre-combusted (450°C for 4 hours) 47-mm Whatman GF/F filters. The particulate matter on the Whatman GF/F filters was rinsed with pre-filtered lagoon water, and the filters were immediately frozen for the later isotope analysis ashore. Before the isotope analysis, filters were treated with HCl fumes for 4 hours to remove inorganic carbon and completely dried in a vacuum desiccator. The isotopic ratios of 13C to 12C and particulate organic carbon contents were determined using a mass spectrometer (ANCA SL, CerCon). The total inorganic carbon in the water was measured with an infrared analyser (Shimadzu TOC 5000). The primary production rate was calculated according to the equation described by Hama et al. (Reference Hama, Miyazaki, Ogawa, Iwakuma, Takahashi, Otsuki and Ichimura1983). The rates in the two polycarbonate bottles at each light depth were averaged.
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 –20°C. Incident insolation (photosynthetically active radiation, PAR, 400–700 nm) was monitored every two minutes with a quantum sensor (Alec MDS MKV/L) during the incubation. For phytoplankton species examination, a 1-l seawater sample was taken and fixed with neutralized formalin (2% vol/vol). These samples were concentrated to 10 ml by siphoning off excess solution after three days of settling. Aliquots (1/20–1/10 of each sample) were examined, depending upon the cell density, using a light microscope equipped with interference and phase contrast optics. Phytoplankton were identified to species level when possible, or otherwise grouped by genera.
RESULTS
Physical and chemical conditions of the lagoon during the observation period
In the winter of 2009, the surface of the lagoon only had a maximum of 64% ice cover (Figure 1). Solar radiation during incubation was nearly equal in February and March, and an increase was observed in April (Figure 2A). The depth of the euphotic zone (1% light depth) shallowed from February (15 m) to April (7.5 m). The depth of the euphotic zone extended through the whole water column in February.
Fig. 2. Change in the daily total photosynthetic active radiation (PAR: 400–700 nm) during incubation experiments (A) and vertical profiles of temperature (B), salinity (C), nitrate (NO3; D), phosphate (PO4; E) and silicate (SiO2; F) in February (♦), March (■) and April (▴). The data of temperature and salinity were obtained at 0.5 m intervals.
Water temperatures were almost constant throughout the whole water column in February and March, and almost constant to near bottom in April (Figure 2B). The temperatures within the euphotic zone were about –1.5°C in February, about 0°C in March and about 5°C in April. Salinity was also almost constant from surface to bottom in February, whereas it increased with depth in the other months (Figure 2C). A remarkable decrease of salinity in the surface layers was found in April, probably due to a supply of inflow of snow melt water from the terrestrial environment as well as the melting water from the ice.
Nitrate was almost constant throughout the water column with concentrations of about 9 μmol l−1 in February (Figure 2D). In the other months, although nitrate was almost exhausted in the upper layers of 17% light depth within the euphotic zone, relatively high concentrations compared with the upper layers were found at the 1% light depths. The concentrations at the 1% light depth decreased from March to April. The results imply a nutrient supply from lower layers to the upper layers and less supply in April than in March. Ammonium concentrations were generally lower than 0.4 μmol l−1 within the euphotic zone (not shown). The same temporal changes during February to April as well as the same vertical profiles were observed in phosphate and silicate within the euphotic zone (Figure 2E, F). However, they were not completely exhausted in the upper layers within the euphotic zone even in March and April.
Chlorophyll-a concentrations and size distributions
Total Chl a concentrations within the euphotic zone were nearly constant in every month (Figure 3A). The Chl a concentrations increased two to three times from February (4–6 mg m−3) to March (8–11 mg m−3). Maximum values during the observation were found in March. Almost equal values as in March were observed in April (7–10 mg m−3).
Fig. 3. Vertical profiles of size-fractionated chlorophyll-a concentration (A) and chlorophyll-a size composition (B) in February, March and April. 
, <2 μm; □, 2–10 μm; ■, >10 μm. Figures in parentheses on the vertical axis show the depth in metres.
Large size phytoplankton, >10 μm fraction, dominated the total Chl a concentrations in every depth and every month (Figure 3B). The percentage contributions of the large size phytoplankton were nearly constant within the euphotic zone. The phytoplankton accounted for 70–90% of the total Chl a concentrations in February, 90–95% in March and 60–90% in April. The maximum contribution of the large size phytoplankton to total Chl a concentration was observed in March when total Chl a concentrations were maximum in this study.
Primary production
Primary production was nearly constant at the 17% light depth and above in February, whereas it decreased with depth in March and April (Figure 4A). The production rates at the 1% light depth were extraordinarily lower than those at the upper light depths in every month (<1.5 mgC m−3 day−1). The primary production values in this study ranged from 25 to 165 mgC m−3 day−1 except for the 1% light depth. The maximum value was found at the 100% light depth in March.
Fig. 4. Vertical profiles of primary production (A) and chlorophyll-a-specific primary production (B) in February, March and April. Figures in parentheses on the vertical axis show the depth in metres.
The daily primary production in the water column, calculated by the trapezoidal integration from the surface to the 1% light depth, was in the range of 298 and 786 mgC m−2 day−1 (Table 1). The maximum daily primary production was observed in March, and the maximum value was about three times higher than that in February and about two times higher than that in April.
Table 1. Daily primary production in February, March and April.
Chl a-specific primary production, an index of phytoplankton growth rate (e.g. Lalli & Parsons, Reference Lalli and Parsons1993), decreased with depth in every month, and abrupt decreases were found at the 1% light depth (Figure 4B). The Chl a-specific primary production at the 17% light depth and above increased from February (4.9–8.7 mgC mgChl a −1 day−1) to March (3.3–20.0 mgC mgChl a −1 day−1), whereas the values were nearly equal in March and April (8.7–17.0 mgC mgChl a −1 day−1). The change in Chl a-specific primary production from February to April was similar to those in total Chl a and in primary production.
Phytoplankton species
Diatoms accounted for almost 100% of all living cells counted by the light microscope at every light depth in every month. The cell numbers of diatoms increased extraordinarily from February (0.7–1.6 × 105 cells l−1) to April (1.2–4.3 × 106 cells l−1) through to March (5.6–9.7 × 105 cells l−1) (Figure 5A). Marked maximum values were found at the 100% light depth in April. The temporal variation in diatom cell numbers was not the same as those in Chl a concentration and in primary production.
Fig. 5. Vertical profiles of cell numbers of diatom (A) and percentage contributions of Thalassiosira spp. (■), Chaetoceros spp. (□) and other diatoms (
) to total diatom cell numbers (B) in February, March and April. Figures in parentheses on the vertical axis show the depth in metres.
Thalassiosira spp. were dominant at every light depth in February, accounting for 70–90% of all cell numbers of diatom (Figure 5B). Both Thalassiosira spp. and Chaetoceros spp. were dominant in March with maximum Chl a and daily primary production; the former and latter species accounted for 30–40% and 40–60% of all diatom cells, respectively. In April, Chaetoceros spp. were remarkably dominant, accounting for more 95% of all diatom cells. Succession of dominant diatom species, Thalassiosira spp. to Chaetoceros spp., was observed during February to April.
DISCUSSION
Saroma-ko Lagoon was often completely covered by ice during winter in the past. In such a situation, the spring phytoplankton bloom occurs in the ice melting season, April, in the lagoon (Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Nishihama et al., Reference Nishihama, Kurata and Tada1989; Tada et al., Reference Tada, Kurata and Nishihama1993). Chl a concentrations were between 0.5 and 1 mg m−3 in the water before ice coverage (December), and concentrations were generally less than 0.5 mg m−3 in the water below the ice (Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Satoh et al., Reference Satoh, Yamaguchi, Watanabe, Tanimura, Fukuchi and Aruga1989; Tada et al., Reference Tada, Kurata and Nishihama1993; Niimura et al., Reference Niimura, Ishimaru and Taguchi2000). The Chl a concentrations reached 5–20 mg m−3 in the spring bloom, and the bloom usually continued for about one month (Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Nishihama et al., Reference Nishihama, Kurata and Tada1989; Tada et al., Reference Tada, Kurata and Nishihama1993). In this study, the Chl a concentrations within the euphotic zone were about 5 mg m−3 in February and reached a maximum of about 10 mg m−3 in March (Figure 3A). Maximum daily primary production was observed in March (Table 1) and maximum Chl a-specific primary production (growth rate) was also observed in March (Figure 4B). Relatively high phytoplankton biomass and production were maintained in April (Figures 3A & 4B; Table 1). In addition, large size phytoplankton (Figure 3B) and diatoms contributed significantly to phytoplankton communities during observations. The results show that large size diatom bloom occurred in early spring and continued to spring in the ice-free area of the lagoon; namely the occurrence of an early spring bloom was observed.
Ice coverage and overlying snow inhibit incident solar radiation into the water below the ice, and the relative incident solar radiation reaching below the sea ice was less than 10% in Saroma-ko Lagoon (Satoh et al., Reference Satoh, Yamaguchi, Watanabe, Tanimura, Fukuchi and Aruga1989; Kishino, Reference Kishino1993; Hattori et al., Reference Hattori, Kato and Saito2001; Ikeya et al., Reference Ikeya, Kikuchi-Kawanobe and Kudoh2001). Rises in irradiance have been reported to increase phytoplankton production and biomass in the water below the ice in Antarctic lakes (Lizotte & Priscu, Reference Lizotte and Priscu1992; Lizotte et al., Reference Lizotte, Sharp and Priscu1996). Observations in Antarctica (Rivkin et al., Reference Rivkin, Putt, Alexander, Meritt and Gaudet1989) and on the Labrador Shelf (Irwin, Reference Irwin1990) have shown that phytoplankton beneath the ice have the potential capacity to increase primary production remarkably when light conditions are extraordinarily improved, i.e. dim to bright conditions. Furthermore, Ikeya et al. (Reference Ikeya, Kashino, Kudoh, Imura, Watanabe and Fukuchi2000) found the same potential capacity in Chaetoceros sp. isolated from the water in Saroma-ko Lagoon immediately after ice melting in April. An ice-free condition, i.e. improvement of light condition in the water, is considered to be a primary factor for the onset of the observed early spring bloom.
Winter blooms have been observed in sub-Arctic coastal regions without ice coverage (Keller et al., Reference Keller, Oviatt, Walker and Hawk1999; Labry et al., Reference Labry, Herbland, Delmas, Laborde, Lazure, Froidefond, Jegou and Sautour2001; Townsend & Thomas, Reference Townsend and Thomas2001; Oviatt et al., Reference Oviatt, Keller and Reed2002; Glé et al., Reference Glé, Del Amo, Beg, Sautour, Froidefond, Gohin, Maurer, Plus, Laborde and Chardy2007; Álvarez et al., Reference Álvarez, Nougueira, Acuña, López-Álvarez and Sostres2009). Light was considered to be the trigger for the onset of the winter bloom (Legendre, Reference Legendre1990; Labry et al., Reference Labry, Herbland, Delmas, Laborde, Lazure, Froidefond, Jegou and Sautour2001). Labry et al. (Reference Labry, Herbland, Delmas, Laborde, Lazure, Froidefond, Jegou and Sautour2001) showed that the winter bloom occurred when the depth-averaged irradiance (Em) received by phytoplankton in the mixed layer was above light saturation parameter (EK; Talling, Reference Talling1957; Sakshaug et al., Reference Sakshaug, Bricaud, Dandonneau, Falkowski, Kiefer, Legendre, Morel, Parslow and Takahashi1997) which is an index of the onset of light-saturation for Pmax. The early stage of the early spring bloom is considered to have occurred during February, and we calculated Em (Riley, Reference Riley1957) and EK (Sakshaug et al., Reference Sakshaug, Bricaud, Dandonneau, Falkowski, Kiefer, Legendre, Morel, Parslow and Takahashi1997) in February as follows for investigating the relationship between the two parameters:
where E0 is the total surface incident irradiance (incident PAR; μmol photons m−2 s−1), K is the vertical attenuation coefficient for PAR (m−1) and Zm is the depth of mixed layer (m) . Based on the measurement of the vertical profiles of PAR in the water, K in February was calculated to be 0.24 m−1 from the linear relationship between ln(PAR) in water and the depth. E0 was 25 mol photons m−2 day−1 and Zm was 15 m in February:
where Pmax and α are the maximum primary production and the initial slope in the relationship between light intensity and primary production, respectively. Diatoms dominated phytoplankton communities within the euphotic zone in February, March and April, and diatom species compositions were almost equal among four light depths in each month (Figure 5B). The four light depths represent four light intensities. Accordingly, the vertical profiles of Chl a-specific primary production in each month are considered to indicate the relationship between light intensity and primary production of phytoplankton community. The primary production increased to the maximum light intensity (100% light intensity: 20–30 mol photons m−2 day−1) in February, March and April, i.e. during the period of the early spring bloom (Figure 4B). The relationship between light intensity and primary production can be represented by the hyperbolic tangent equation as follows (Jassby & Platt, Reference Jassby and Platt1976):
where P and E are Chl a-specific primary production and light intensity, respectively, and Pmax and α are maximum primary production and the initial slope in the relationship between light intensity and primary production, respectively. Optimal values of Pmax and α were obtained by applying the data to Equation (3) by nonlinear least-squares fitting, using the Microsoft Excel Solver. Pmax and α in February were estimated to be 8.7 mgC mgChl a −1 day−1 and 1.33 mgC mgChl a −1 day−1 (mol photons m−2 day−1)−1, respectively. As a result, the Em and EK were calculated to be 6.8 mol quanta m−2 day−1 and 6.6 mol quanta m−2 day−1, respectively, and both of the values were almost equal. It is thus probable that light availability was satisfied for the bloom diatoms, supporting that light was a primary factor for the early spring bloom.
It is well known that the phytoplankton bloom occurs when the mixed layer depth (Zm) is shallower than the critical depth (Zcr) (Sverdrup, Reference Sverdrup1953; Parsons et al., Reference Parsons, Takahashi and Hargrave1984). Labry et al. (Reference Labry, Herbland, Delmas, Laborde, Lazure, Froidefond, Jegou and Sautour2001) showed that such a water column condition was necessary for the occurrence of the winter bloom, and hence we investigated the relationship between Zm and Zcr in February which was considered to be the initial stage of the early spring bloom in Saroma-ko Lagoon. The depth of the euphotic zone (15 m), where the net primary production is positive and there is 1% surface radiation (Parsons et al., Reference Parsons, Takahashi and Hargrave1984), was almost equal to the depth of water (16 m) in February, and the water was well mixed throughout the water column (see uniform vertical profiles of temperature and salinity in February in Figure 2B, C). Zcr is deeper than the depth of the euphotic zone (Sverdrup, Reference Sverdrup1953; Parsons et al., Reference Parsons, Takahashi and Hargrave1984), and hence the mixed layer depth was considered to be shallower than the critical depth in February, meaning that the water column condition was satisfied for occurrence of the early spring bloom. The mixed layer depth does not become deeper than the depth of water. The shallow depth of water in Saroma-ko Lagoon (19.5 m at deepest) possibly plays an important role in the occurrence of the early spring bloom after all.
The concentrations of nitrate and phosphate within the euphotic zone were about 9 and 1 μmol l−1, respectively, in February, probably at the initial stage of the bloom (Figure 2D, E). The concentrations were nearly equal to those in the water below the ice in February (before the spring bloom) in the complete ice coverage years (nitrate + nitrite: about 10 μmol l−1; phosphate: about 1.3 μmol l−1; Tada et al., Reference Tada, Kurata and Nishihama1993). The maximum Chl a concentrations in the early spring bloom, about 10 mg m−3 (Figure 3A), were nearly equal to those in the spring bloom (5–20 mg m−3; Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Nishihama et al., Reference Nishihama, Kurata and Tada1989; Tada et al., Reference Tada, Kurata and Nishihama1993). Nutrients regulate phytoplankton productivity and biomass, and Chl a concentration reflects phytoplankton productivity and biomass. Consequently, the similarity between the nutrient concentrations and maximum Chl a concentrations in this study and the previous studies circumstantial evidence that the phytoplankton (large diatoms) in the early spring bloom had the same potential growth capacity as the phytoplankton in the usual spring bloom.
This study, furthermore, revealed the bloom of large size phytoplankton and succession of dominant species in the early spring bloom in Saroma-ko Lagoon (Figures 3 & 5). Unfortunately little is known about phytoplankton species in the spring bloom in the lagoon, though Detonula confervacea, Nitzschia spp., Odontella aurita and Thalassiosira spp. existed dominantly beneath the ice in February (Watanabe et al., Reference Watanabe, Satoh and Yamaguchi1993; Ikeya et al., Reference Ikeya, Kikuchi-Kawanobe and Kudoh2001) and Chaetoceros spp. and Thalassiosira spp. dominated in the water a few days after sea-ice disappearance in April (Kashino et al., Reference Kashino, Fujimoto, Akamatsu, Koike, Satoh and Kudoh1998). On the other hand, the dominance of Thalassiosira spp. and Chaetoceros spp. was observed in the spring diatom bloom after the ice melt in Notoro-ko Lagoon which is geographically proximate to Saroma-ko Lagoon, i.e. about 25 km south-west of it, and is connected to the Okhotsk Sea (Asami & Imada, Reference Asami and Imada2001). In addition, Thalassiosira spp. and Chaetoceros spp. were recorded as dominant large size diatoms in the spring blooms in the coastal regions of Hokkaido (Nakata, Reference Nakata1984; Taguchi et al., Reference Taguchi, Satoh, Hamaoka, Ikeda, Ishikawa and Shirasawa2000), and the succession from Thalassiosira spp. to Chaetoceros spp. was also reported (Nakata, Reference Nakata1984). Saroma-ko Lagoon is connected to the Okhotsk Sea by the two channels and the lagoon water is significantly affected by the coastal water in the Okhotsk Sea (Shibanuma et al., Reference Shibanuma, Kajihara and Miyake1995), suggesting that there is analogy in the spring bloom phytoplankton between the lagoon and the Hokkaido coastal region. The same results as the previous studies of the spring bloom were obtained in this study. Consequently, there is a great possibility that the dominant diatoms in the early spring bloom are similar to those in the spring bloom in Saroma-ko Lagoon. Therefore, the same potential growth capacity and dominant diatoms show that the early spring bloom observed in this study front-loaded the spring bloom.
Our observations began from the middle of February, by which time the early spring bloom had already started. The time increase in the phytoplankton community can be estimated under the assumption that logarithmic growth is maintained as follows (e.g. Parsons et al., Reference Parsons, Takahashi and Hargrave1984):
where N1 and N2 are phytoplankton biomass (Chl a concentration) at times t1 and t2 (days), respectively, and μ is the specific phytoplankton growth rate (day−1). Biomass is usually expressed as carbon content (mgC m−3), but here the initial biomass was measured as the Chl a concentration (mg m−3). Chl a concentration was converted to the carbon content of the phytoplankton by multiplying the slope between the concentrations of Chl a and the particulate organic carbon before incubation. A significant linear regression line was obtained with a slope of 60 (P < 0.001). The value was within the range of the carbon to Chl a ratios of phytoplankton assemblages in previous culture and field studies (25–100 mgC mgChl a −1; e.g. Parsons et al., Reference Parsons, Takahashi and Hargrave1984; Furnas, Reference Furnas and Subba Rao2002). In this study, N1 as Chl a was assumed to be 0.5 mg m−3 because Chl a concentrations in the water before ice coverage (December) were reported to be 0.5–1 mg m−3 (Fukuchi et al., Reference Fukuchi, Watanabe, Tanimura, Hoshiai, Sasaki, Satoh and Yamaguchi1989; Satoh et al., Reference Satoh, Yamaguchi, Watanabe, Tanimura, Fukuchi and Aruga1989; Tada et al., Reference Tada, Kurata and Nishihama1993; Niimura et al., Reference Niimura, Ishimaru and Taguchi2000). N2 as Chl a was 5.0 mg m−3 because Chl a concentrations at the four light depths within the euphotic zone in February were averaged to be 4.95 mg m−3. N1 and N2 as Chl a were converted to 30 and 300 mgC m−3, respectively. The specific growth rate is calculated by the following equation (e.g. Parsons et al., Reference Parsons, Takahashi and Hargrave1984):
where N0 is the initial biomass of the phytoplankton and ΔN is the biomass increased by net photosynthesis per day (mgC m−3 day−1). N0 was the phytoplankton biomass in February, i.e. N2 (300 mgC m−3). ΔN (43.5 mgC m−3 day−1) was obtained by multiplying Pmax in February (8.7 mgC mgChl a −1 day−1) by the mean Chl a (5.0 mg m−3), meaning that the value was potential growing capacity of phytoplankton in February. As a result, the growth rate (μ) in February was calculated to be 0.14 day−1. Supposing that the growth rate was constant (0.14 day−1) during the onset of the early spring bloom to the day of the February observation (16 February), t2–t1, the duration of increase in Chl a concentration from 0.5 to 5.0 mg m−3, was calculated to be 16 days using Equation (4). t2 was 16 February and thus the early spring bloom was estimated to have occurred by 31 January at the latest, because the maximum growth rate during the 16 days after the onset of the bloom was used for the calculation. Accordingly, the early spring bloom was thus considered to have continued for almost three months. The duration of the early spring bloom was longer than that of the spring bloom (about one month).
Maximum phytoplankton biomass and production were observed in March (18 March) and nearly equal values as in March were found in April (17 April; Figures 3A & 4; Table 1), reflecting that the prime time of the early spring bloom lasted about one month at least. Nevertheless, nutrients, especially nitrate, were exhausted in the upper layers of the 17% light depth within the euphotic zone in March and April (Figure 2B, C & D). In contrast, relatively high concentrations were found at the 1% light depth (Figure 2B, C & D). The results imply that the nutrient supply to the upper layers from the lower layers was related to the maintenance of the continuation of the bloom. However, concentration gradients of all nutrients between the 17% and 1% light depths diminished markedly from March to April (0.71 to 0.15 μmol l−1 m−1 for nitrate, 0.0791 to 0.0036 μmol l−1 m−1 for phosphate and 1.7 to 1.1 μmol l−1 m−1 for silicate). On the other hand, markedly low salinity was found in the upper layers in April (Figure 2C). The low salinity was probably due to a supply of inflow of snow melt water from the terrestrial environment as well as melting water of ice. Consequently, nutrient supply through the rivers probably played a large part in the maintenance of the prime time of the bloom.
Progression of global warming would be expected to induce incomplete ice coverage. In winter 2009, the lagoon was only partly covered by ice (Figure 1). Thus the front-loading of the ordinary spring bloom observed in this study shows the phytoplankton state during winter and spring in the lagoon that can be expected as global warming progresses. Improvement in the light condition in the water was considered to be a primary factor responsible for the front-loading. It is not difficult to expect that the progression of global warming induces temporal and spatial expansion of the ice-free area, and hence improvement of the light condition in the water. This appears to result in front-loading of the ordinary spring bloom in areas that normally would be ice-covered and where the spring bloom normally occurs only after ice-melting. The front-loaded bloom could possibly continue through to the regular spring bloom when nutrients are supplied.
Progression of the shortening of ice coverage and frequent incomplete ice coverage means temporal and spatial expansion of the ice-free area, and hence an increase of contribution in the ice-free area to the overall primary production during winter. Based on limited data for wintertime primary production in Saroma-ko Lagoon, the primary production is higher in the ice-free area (298–786 mgC m−2 day−1 in February, March and April in this study; Table 1) than in the ice-covered area (198 mgC m−2 day−1; Hattori et al., Reference Hattori, Kato and Saito2001). The potential capacity of primary production of phytoplankton in the water below the ice has been reported to be higher than that of ice algae (Rivikin et al., 1989; Irwin, Reference Irwin1990; Smith & Sakshaug, Reference Smith, Sakshaug and Smith1990; Ban et al., Reference Ban, Aikawa, Hattori, Sasaki, Sampei, Kudoh, Fukuchi, Satoh and Kashino2006). Therefore, the progression of global warming may possibly result in increases in primary production during the winter and spring, thereby affecting the ecosystem and water quality in Saroma-ko Lagoon, as well as in other aquatic areas normally covered by ice. Continued monitoring of the ice condition and primary production in Saroma-ko Lagoon will therefore be important to confirm this possibility and for further prediction of the influence of global warming on ice-covered aquatic areas throughout the world.
ACKNOWLEDGEMENTS
We would like to express our appreciation to Aquaculture Fisheries Cooperation of Lake Saroma for the ship observation. We are very grateful to Dr M. Takahashi for his critical and helpful comments on the manuscript.
