Introduction
Despite the accelerating reduction of Arctic sea ice over the past decades (Budikova Reference Budikova2009), ice algae are usually considered indispensable in biogeochemical cycles in the Arctic Ocean (Gosselin and others Reference Gosselin, Levasseur, Wheeler, Horner and Booth1997; Søreide and others Reference Søreide, Hop, Carroll, Falk-Petersen and Hegseth2006). They used to contribute 4% ~ 26% of the total primary production in seasonally ice covered waters and more than 50% in the permanently ice covered central Arctic (Gosselin and others Reference Gosselin, Levasseur, Wheeler, Horner and Booth1997; Sakshaug Reference Sakshaug, Stein and Macdonald2004). In a recent study of west Greenland, ice algae were found to contribute 30% of total primary production during the sea ice season and less than 1% annually (Mikkelsen and others Reference Mikkelsen, Rysgaard and Glud2008). Particulate ice algal production during the transect from the Chukchi Sea to the Nansen Basin via the North Pole in July/August 1994 ranged from 0.5 to 310 mg C m−2 day−1 and showed maximum rates in the central Arctic Ocean (Gosselin and others Reference Gosselin, Levasseur, Wheeler, Horner and Booth1997). They were major contributors to primary production and algal biomass in the Chukchi and Beaufort Seas during May/June 2002 with significantly higher primary productivity (0.1–23.0 mg C m−2 h−1) and pigment content (0.2–304.3 mg pigments m−2) than water column parameters (0.2–1.0 mg pigments m−3; < 0.1–0.4 mg C m−2 h−1) (Gradinger Reference Gradinger2009). They are also a main food source for sympagic fauna (Werner Reference Werner1997; Poltermann Reference Poltermann2001), which might supply and extend the restricted grazing season for Arctic zooplankton (Michel and others Reference Michel, Legendre, Ingram, Gosselin and Levasseur1996), and serve as food for the benthos as large aggregates when the ice begins to melt in spring (McMahon and others Reference McMahon, Ambrose, Johnson, Sun, Lopez, Clough and Carroll2006; Renaud and others Reference Renaud, Riedel, Michel, Morata, Gosselin, Juul-Pedersen and Chiuchiolo2006). Additionally, ice algal communities are a major contributor of the Arctic's biodiversity. For example, at least 251 species were recorded in first year ice from the Chukchi Sea in early June 1998 (Quilfeldt and others Reference Quilfeldt, and Clough2003).
The central Arctic Ocean now is changing substantially with an expected reduction in ice cover and probable changes of the polar flora and fauna due to global climate change. A reduction in ice thickness of the southern Chukchi Sea was observed of approximately 0.5–1.0m over the past two decades (Shirasawa and others Reference Shirasawa, Eicken, Tateyama, Takatsuka and Kawamura2009). The ice extent observed in summer 2003 was almost as low as the record low levels of the ice extent in the summer of 2002 since being monitored by satellite from 1979 (Lindsay and Zhang Reference Lindsay and Zhang2005). Annual ice algal production of 2003 at Barrow, Alaska (2.0 g C m−2 year−1) was lower than three decades ago by a factor of two to three (Lee and others Reference Lee, Whitledge and Kang2008). Changes in species composition were also observed. Unusual coccolithophorid blooms, the more temperate species, occurred since 1997 in the Bering Sea (Iida and others Reference Iida, Saitoh, Miyamura, Toratani, Fukushima and Shiga2002) may replace the previously occurring summer flagellate community (Schumacher and others Reference Schumacher, Bond, Brodeur, Livingston, Napp, Stabeno, Hempei and Sherman2003). Emiliania huxleyi was even reported to intrude and dominate the high Arctic Ocean in summer 2003 (Hegseth and Sundfjord Reference Hegseth and Sundfjord2008). Greater overall primary production was expected in a warmer Arctic with less sea ice and earlier ice melting, and resulted from greater increase in the pelagic production than the reduction in ice algal production (Wassmann and others Reference Wassmann, Slagstad, Wexels Riser and Reigstad2006; Lavoie and others Reference Lavoie, MacDonald and Denman2008). This was proved by a recent study that the annual Arctic primary production has increased yearly by an average of 27.5 TgC.yr−1 since 2003 and by 35 TgC.yr−1 between 2006 and 2007 (Arrigo and others Reference Arrigo, van Dijken and Pabi2008). As the ongoing trend persists, the pelagic food web is expected to dominate over a geographically larger area (Bluhm and Gradinger Reference Bluhm and Gradinger2008).
Most studies of algal communities in the ice-covered Arctic waters focus on coastal and shelf regions. The Canada Basin (maximum depth > = 3,800 m), the largest and geographically most isolated Arctic basin (Swift and others Reference Swift, Jones, Aagaard, Carmack, Hingston, Macdonald and Mclaughlin1997), is one of the least known areas in the Arctic Ocean (Lee and Whitledge Reference Lee and Whitledge2005). Our knowledge about the ice algal community in this area is still sparse and can hardly be compared due to different measurements and variable studying seasons. The ice algal biomass of the Canada Basin were measured between 0.4 μg C L−1 to 5.3 μg C L−1 for the multi-year ice from autumn 1997 to spring 1998, and < 0.1 μg C L−1 to 1.7 μg C L−1 for the first year ice in autumn and winter of 1997 (Melnikov and others Reference Melnikov, Kolosova, Welch and Zhitina2002). The annual primary production and new production of phytoplankton in the Canada Basin in 2002 was 5 and 1g C m−2 year−1, respectively (Lee and Whitledge Reference Lee and Whitledge2005). Chl a concentration in off-shore pack ice of the Beaufort Gyre ranged from 0.1–1.7 mg Chl a m−2 in summers 2002 and 2003 which was similar to estimates for multi-year ice of the transpolar drift and ~2 orders of magnitude below coastal fast first year ice estimates (Gradinger and others Reference Gradinger, Meiners, Plumley, Zhang and Bluhm2005). Until now, we are still far from having a comprehensive perception of the algal community in the Canada Basin.
Our present research, conducted during the second Chinese National Arctic Expedition (CHINARE-2003) in the late summer of 2003 and based on R/V Xuelong, focused on the algal communities in sea ice and the under ice water column in the pack ice zone of the Canada Basin. This information provides baseline data to evaluate future changes in the algal communities driven by changing ice regimes and any loss of the multi-year ice cover (Serreze and others Reference Serreze, Maslanik, Scambos, Fetterer, Stroeve, Knowles, Fowler, Drobot, Barry and Haran2003).
Materials and methods
Samples were collected in the Canada Basin of the Arctic Ocean between 11 August and 6 September 2003 (Fig. 1). A MARK II type ice corer (9 cm in inner diameter) was used to collect one ice core each from seven pack ice stations (Table 1; due to loss of the on-the-spot record, the ice cores longer than 1.5 m were considered as multi-year ice). The upper half of each ice core was drilled out first to obtain brine, and the lower half was drilled through later. The whole process for each ice core was completed within 10 minutes. Ice cores were immediately sectioned into 10- or 20- cm segments, transferred into 1-L plastic containers, and then melted in particle free seawater (pre-filtered through 0.2 μm Whatman nuclepore polycarbonate filters to minimise osmotic stress during the melting process) at 4°C in the dark (Garrison and Buck Reference Garrison and Buck1986). The underlying water samples were collected with a 2.5 L Niskin type water sampler, directly through the drilled ice holes. At most stations, water samples were only collected in the ice-water interface (0 m depth). For station SI4, water was sampled at 0 and 10 m depths, while 0, 5 and 10 m depth waters were sampled for station SI5. For species identification and cell counts, all the samples were fixed with glutaraldehyde (final concentration 1.5–2%) and stored at 4°C in the dark before analysis.
The investigated area in the Canada Basin of the Arctic Ocean during the second Chinese National Arctic Expedition from 11 August to 6 September 2003.
Sampling information of seven ice stations in the Canada Basin during the Second Chinese National Arctic Research Expedition
Samples were settled according to the methods described by Utermöhl (Reference Utermöhl1958) before counting. Algal cells were counted and identified in a 0.1 ml phytoplankton counting chamber using a Nikon 80i microscope with phase contrast illumination at a magnification of 200–400×. Identification of algal species was based on morphology and cell size (Poulin and Cardinal Reference Poulin and Cardinal1982a; Poulin and Cardinal Reference Poulin and Cardinal1982b; Poulin and Cardinal Reference Poulin and Cardinal1983; Medlin and Hasle Reference Medlin and Hasle1990; Medlin and Priddle Reference Medlin and Priddle1990; Hegseth Reference Hegseth1992; Legendre and others Reference Legendre, Ackley, Dieckmann, Gulliksen, Horner, Hoshiai, Melnikov, Reeburgh, Spindler and Sullivan1992; Hasle and Syvertsen Reference Hasle, Syvertsen and Tomas1996). At least three horizontal transects across the bottom of the chamber (one in the middle and the other two at the two sides) and 300 cells with intact plasma were counted for each sample. More than 10 algal cells for each taxon were measured and cell numbers were transformed into carbon biomass according to the formula of Menden-Deuer and Lessard (Reference Menden-Deuer and Lessard2000). Shannon-Wiener diversity index (Shannon and Wiener Reference Shannon and Wiener1963) with a log2 basis and species number were used to measure the biodiversity of the algal communities. SPSS v. 16.0 with the Euclidean distance and the nearest neighbour method was applied to conduct cluster analysis of algal communities, treated as a matrix composed of species and specific abundances, to determine their relationships and similarity. T-test and correlation analysis were included.
Results
Algal species in the ice cores and the underlying water column
102 and 78 algal species were identified in the ice cores (Table 2) and the underlying water column (Table 3) respectively of the seven sea ice stations, with diatoms dominating in both habitats (95 species in ice; 75 species in water). A small fraction (< 1% in total abundance of each sample) was composed of dinoflagellates, chrysophytes and chlorophytes. Many algae could only be identified to genus or higher taxa, such as Nitzschia spp. and Navicula spp.
Algal abundance (×103 cells L−1) of every species present in the sea ice samples collected from seven ice cores (“_” characterised as the abundance < 10 cells L−1 or absent)
Algal abundance (×103 cells L−1) of every species present in the water samples collected in seven ice stations (“_” characterised as the abundance < 10 cells L−1 or absent)
The micro-sized ice flora of the Canada Basin was predominately diatoms (Table 2), including Cylindrotheca closterium, Fragilariopsis cylindrus, Nitzschia frigida, Nitzschia neofrigida, Nitzschia polaris, Fragilariopsis curta, Thalassionema nitzschioides, Nitzschia arctica, Nitzschia promare, Pinnularia quadratarea, and Navicula directa (from most abundant to less abundant) (Fig. 2).
The vertical distribution of algal biomass in the ice cores and the underlying water.
Pennate diatoms also dominated the underlying water column (Table 3), such as C. closterium, F. cylindrus, F. curta, N. frigida, Navicula pelagica, N. neofrigida, Synedropsis hyperborea, Pseudonitzschia delicatissima, N. polaris, Stauroneis radissonii, and N. promore (from most abundant to less abundant), most of which were also found as the dominant species in the ice cores. All samples contained numbers of Chaetoceros spores and unidentified dinoflagellates spores. The algal spores together occupied more than 1% of abundance in all the ice cores and most of the water samples, ranging from 0.48 × 103 cells L−1 to 1.48 × 104 cells L−1 (Tables 2 and 3).
Algal biomass and abundance
The ice algal biomass ranged from 1.3 to 89.5 μg C L−1, with the lowest value found at 216–226 cm in the ice core from SI1 and the highest in the bottom 10 cm of the ice core from SI4 (Fig. 2). The total biomass of the SI6 ice core was less than other ice cores (P < 0.01) (Fig. 2). The algal abundance in the ice cores ranged from 1.40 × 104 to 8.73 × 105 cells L−1, with the minimum abundance located at 216–226 cm in the ice core of SI1 and the maximum at 130–155 cm in the ice core of SI5 (Fig. 3).
The vertical distribution of total algal abundances containing that of the dominant species in the ice cores and the underlying water.
The ice algal biomass differed greatly among different stations and within each ice core (Figs. 2, 4a). The ice algal maximal biomass was found at various layers of different stations. Of the seven ice cores, three (SI3, SI4 and SI6) had an ice algal biomass bottom maximum, two (SI1, SI2) had the maximum biomass in the sub-surface segment, one (SI7) had the biomass maximum in the sub-bottom layer, and one (SI5) in the centre.
The horizontal distribution of (a) algal biomass (b) number of algal species and (c) Shannon-Wiener index at 0-m depth of the water column, the bottom and surface layers of the ice cores, and the layers with maximum biomass and species number.
The algal biomass of all the 10 cm bottom samples was one order of magnitude higher than the underlying 0 m samples (Fig. 4a; P < 0.05). The algal biomass in the water column ranged from 0.56 to 7.63 μg C L−1, with the lowest and the highest in the ice-water interface at SI5 and SI3, respectively (Fig. 2). The phytoplankton abundance ranged from 1.4 × 104 to 9.4 × 104 cells L−1, with the minimum and the maximum value found at the ice-water interface of SI7 and SI3, respectively (Fig. 3). A discrepancy in cell volume among different species caused the different samples with maximum ice algal biomass and abundance and minimum phytoplankton biomass and abundance (Figs. 2, 3).
Abundances for each predominant species varied among stations and layers, and the species composition diverged among stations and between ice and water, either (Fig. 3). Abundant cells of C. closterium, F. cylindrus, N. frigida, N. neofrigida and the spores were found at every station. Large number of dinoflagellate and Chaetoceros spores dominated especially in the near-surface layers of the ice cores, in the water column and in every layers of SI5. They accounted for over 40% in the surface layers in the ice cores of SI5, SI6 and SI7 (Fig. 3).
Biodiversity and composition of the algal community
The number of microalgal species showed indented vertical distribution in every station (Fig. 5) and diverged in the horizontal direction among different stations (Fig. 4b). The Shannon-Wiener index ranged from 1.40 to 4.88 and exhibited similar distribution pattern with less varying extent (STDEV of the Shannon-Wiener indices = 0.68; STDEV of the species numbers = 14.50). They were positively correlated (r = 0.56, P < 0.01) (Fig. 4b, c, 5). An ice sample of the 10–30-cm ice segment of SI6 had both the highest number of algal species (71 taxa) and the highest Shannon-Wiener index (4.88). The second highest, 70 taxa and 4.78, were for phytoplankton at 0-m depth of SI3. The number of algal species and the diversity index in water samples, with much less biomass (P < 0.01) (Figs. 2, 4a), were comparative with the ice samples (P = 0.45 and 0.50 > 0.05) (Fig. 4b, c, 5). Biomass was correlated with species number (r = 0.27, P < 0.05), but was not correlated with Shannon-Wiener index (r = 0.059, P > 0.05). So did abundances (r = 0.41, P < 0.01 and r = -0.024, P > 0.05, respectively) (Fig. 6).
The vertical distribution of the number of algal species and the Shannon-Wiener indices in the ice cores and the underlying water.
(a) Correlation of biomass with number of algal species, (b) correlation of biomass with Shannon-Wiener index, (c) correlation of abundances with number of algal species and (d) correlation of abundances with Shannon-Wiener index.
The cluster analysis result displayed the heterogeneity of the algal community concerning the species composition and the specific abundances (Fig. 7). Among 60 samples, there were 29 samples classified into the primary group, including nearly all of the water samples except two (one collected from SI5 at 5m and the other from the ice-water interface of SI3), which had been characterised by higher species richness and algal biomass than other water samples and were clustered into the secondary group. Most of the algal communities near and at the sea ice surface were also grouped into this assemblage. None of the ice algal communities in the bottom 10 cm shared a same category and none of a single ice core had all its sub-samples clustered into one group (Fig. 7).
Dendrogram based on similarity among algal communities concerning species composition and specific abundances.
Discussion
Diatoms, especially pennate diatoms dominated in the microalgal community in the sea ice and the underlying water in the Canada Basin at the end of the summer 2003. 102 algal species were found in this study, of which 95 species were diatom, confirming that ice algal community was an important component of the Arctic's biodiversity and diatom was the most important contributor in terms of abundance, productivity and number of species (Melnikov and others Reference Melnikov, Kolosova, Welch and Zhitina2002; Quilfeldt and others Reference Quilfeldt, and Clough2003). Flagellates were also reported to substantially contribute to the structure and dynamics of the ice biota community (Ikävalko and Gradinger Reference Ikävalko and Gradinger1997; Thomsen and Ikävalko Reference Thomsen and Ikävalko1997), which may be under-represented in this study since they were difficult to see in the count samples and their cells were fragile to fixatives because of lacking a hard shell. The total number of algal species in this study ranked in the middle among the investigations up to now of the Arctic Ocean including coastal areas despite seasonal variability (Quilfeldt and others Reference Quilfeldt, and Clough2003), with the average Shannon-Wiener index of 3.58 ± 0.68. The correlation coefficient between abundances and species numbers in the Chukchi Sea (Quillfeldt and others Reference Quilfeldt, and Clough2003) (r = 0.83, P < 0.01), one of the few papers listing ice algal species by layer, was higher than our result (r = 0.41, P < 0.01). The irrelevance between the diversity index and biomass (or abundance) and the relevance between the number of algal species and biomass (or abundance) indicated the unevenness of the microalgal community.
Most of the dominant species found in this study were characteristic ice algal species in the Arctic Ocean, including N. frigida, N. neofrigida, F. cylindrus, N. promare, F. oceanica, N. pelagica (Melnikov and others Reference Melnikov, Kolosova, Welch and Zhitina2002; Quilfeldt and others Reference Quilfeldt, and Clough2003; Wassmann and others Reference Wassmann, Slagstad, Wexels Riser and Reigstad2006; McMinn and others Reference McMinn, Hattori, Hirawake and Iwamoto2008). Navicula spp., Nitzschia spp., Fragilariopsis spp., and Pseudonitzschia spp., the most frequently documented pennate genera in the Arctic pack ice (Quillfeldt Reference Quillfeldt1997), also contributed to a great fraction of the algae in this study (> 1% of total abundance in most samples). The existence of Chaetoceros spp., Thalassiosira spp., which are considered true planktonic forms (Gran Reference Gran and Nansen1904), with abundant Chaetoceros spores in the ice samples signaled the ocean origin of the sea ice in the Canada Basin (Quilfeldt and others Reference Quilfeldt, and Clough2003). The appearance of the great number of spores should be attributable to the common survival mechanism for algae in the absence of light (Eilertsen and others Reference Eilertsen, Sandberg and Tøllefsen1995), and could be a good strategy for the formation of a new bloom. Most of the spores here came from dinoflagellates and Chaetoceros spp., which were found to form resting spores regularly (Wassmann and others Reference Wassmann, Slagstad, Wexels Riser and Reigstad2006). The species composition of the algal spores was in accordance with the general algal species renewal in the Arctic sea ice. One example was the sea ice algal community in Kobbefjord, West Greenland dominated by flagellates in December-February, followed by C. simplex in March and pennate diatoms in May (Mikkelsen and others Reference Mikkelsen, Rysgaard and Glud2008).
Our survey, from 11 August to 6 September, was at the end of summer melting and the beginning of autumn freeze-up for the high Arctic, since sea ice starts to melt when the underlying water temperature is over -1.7 °C and ends by August in the Arctic (Andersen and others Reference Andersen, Koc and Moros2004). Due to the reported severe melting in the Arctic in 2003 summer (Lindsay and Zhang Reference Lindsay and Zhang2005; Arrigo and others Reference Arrigo, van Dijken and Pabi2008; Shirasawa and others Reference Shirasawa, Eicken, Tateyama, Takatsuka and Kawamura2009), the prevailing amount of pennate diatoms in the water samples of the present study should be mostly released from the sea ice. Besides, their existence indicates the possible effect of ice algal inocula in the Arctic summer. For example, N. frigida, one of the dominant species, can remain in plankton for some time after being released from sea ice (Syvertsen Reference Syvertsen1991).
Compared to Melnikov and others (Reference Melnikov, Kolosova, Welch and Zhitina2002) we observed higher values for ice algal biomass (89.49 μg C l−1 compared to 19 μg C l−1) and more species. This discrepancy seems to be puzzling here due to the lack of some environmental knowledge like concentrations of nutrient, although the seasonal change might partly account for it. With similar algal species and much lower algal biomass than the above ice samples, the algae in the under-ice water was probably limited by lack of nutrients in the central Arctic since the deep water of the Canada Basin experiences very slow exchange and tidal currents (Kowalik and Proshutinsky Reference Kowalik, Proshutinsky, Johannessen, Muench and Overland1994).
The maximum ice algal biomass of each ice core in the Canada Basin of 2003 summer was found in various parts of the ice cores, either at the bottom, near the bottom, in the centre, or near the surface. Algae in the Arctic pack ice are more distributed throughout the ice during winter, becoming concentrated in the bottom sections in spring, and continuing to grow and accumulate in the bottom layer in summer (Quilfeldt and others Reference Quilfeldt, and Clough2003). This was also the case in the Canada Basin in 1997–1998 (Melnikov and others Reference Melnikov, Kolosova, Welch and Zhitina2002). Although many studies concentrated on the bottom communities of summer Arctic sea ice algae, more attention was paid to the internal maxima (Gradinger Reference Gradinger1999; Quilfeldt and others Reference Quilfeldt, and Clough2003). Our findings agree with the proposal by Gradinger (Reference Gradinger1999) that all layers of the sea ice must be studied to avoid underestimation of algal biomass and production.
The cluster analysis result indicated most of the algal communities near or at the ice surface were more similar to the phytoplankton communities than other ice algal communities. The result reflected the spatial heterogeneity of the sea ice algal communities concerning their species composition and specific abundances, which was shown by the distribution patterns of biomass, abundances, number of species and diversity index too. The spatial heterogeneity was found to be characteristic of the sea ice algae in the Arctic Ocean (Gosselin and others Reference Gosselin, Levasseur, Wheeler, Horner and Booth1997; Rysgaard and others Reference Rysgaard, Kühl, Glud and Hansen2001) and is partly due to the variable horizontal and vertical structure of pack ice arising from the continual movement and shifting (Andersen Reference Andersen and Herman1989). Variable salinity and nutrient concentrations in brine were also responsible for this characteristic, which need to be studied more finely than the currently and broadly used cutting and melting method. In addition, light, the main factor limiting primary production in the Arctic, was another possible interacting cause and was mostly correlated with ice thickness and snow depths (Lazzara and others Reference Lazzara, Nardello, Ermanni, Mangoni and Saggiomo2007).
As over-melting of sea ice carried on in 2003 summer in the Canada Basin of the Arctic Ocean indicated by low salinity throughout the whole ice core and the dominance of bacteria and heterotrophic flagellates in the microbial food web (He and others Reference He, Cai, Jiang, Chen and Yu2005), our survey investigated in detail the composition, biomass and structure of microalgal communities through seven ice cores and the underlying water columns in the pack ice zone, which supplements our limited knowledge in the central Arctic Ocean and provides a background of the sea ice algal communities as the global climate change maintains.
Acknowledgements
This research was supported by the National Natural Science Foundation of China (Nos. 40576002 and 40776003); the Science and Technology Commission of Shanghai Municipality (No. 052307053); the Polar Youth Foundation for Innovation (No. JDQ200802), the Polar Strategic Research Foundation (No. 2008209), and the LMEB Open Research Foundation (No. LMEB200902). And we would like to dedicate our big thanks to the two reviewers whose advice greatly helped improve and complement this paper.
