Summer Phytoplankton Community Structure in the Kuroshio Current-Related Upwelling Northeast of Taiwan

Observations of chlorophyll a concentration and phytoplankton assem­ blage were conducted during the ·summer of 1990 on 19 stations along two par­ allel transects across the Kuroshio Current front off the northeastern coast of Taiwan. The 0-lOOm seawater chlorophyll a content varied between 8.63 and 41.SOmg·m-2 in August and between 8.88 and 94.70mg·m-2 in September. High nitrate content, low temperature, high salinity and high chlorophyll a concentrations were observed at the stations along the front. The vertical dis­ tribution of chlorophyll a was characterized by a shallower subsurface maxi­ mum of between 0 and 25m in the stations associated with the upwelling, while those of the non-upwelling stations were between 50-75m. Diatoms, Skele­ tonema costatum in August and Thalassionema nitzschioids in Septem­ ber, dominated the upwelled surface water because significant amounts of nutrients are supplied to surface or near-surface waters.


INTRODUCTION
The response of phytoplankton in upwelled waters can be conceptualized in various ways. Increases in production (Barber and Smith, 1981;Guillen et al., 1971;Reid et al., 1970) and/or standing crop (Guillen et al., 1971;Reid et al., 1970) in the nutrient-rich upwelling waters were observed either immediately or with a lag period·due to the variation of upward speed and/or the chelation of trace metals of the upwelled water. Rapid vertical speed may lead to decreased production. The speed of water drifting away from the upwelling zone will, thus, in part determine the richness and the positioning of the productive belt.
Increases in the netphytoplankton fraction are closely coupled with the occurrence of coastal upwelling. In both temperate (Gilmartin, 1964;Anderson, 1965) and tropical (Saijq and Takesue, 1965;Malone, 1971 a) marine environments, the nannophytoplankton are usu ally responsible for 80% to '100% of the observed phytoplankton productivity and standing crop. Netphytoplankton, on the other hand, exceeds nannophytoplankton in productivity and standing crop during periods of strong upwelling (Malone, 1971 b). This shift in phytoplank ton assemblages during upwelling was suggested to be caused by the retention of larger   TAO, Vol.3, No.3, Sep. 1992 particles in strong vertical advection, and by the increases in ambient N03-N concentrations. The relative importance of these two size groups of phytoplankton thus can be an effective tool to describe an upwelling process.
Large populations of phytoplankton dominated by diatoms were observed in many upwelled waters (Tant, 1976;Yoder et al., 1981;Takahashi and Kishi, 1984). In Peruvian upwelling area, a low diversity of phytoplankton population was illustrated by the observation that eight species of diatom comprised 83% of the population (Ryther et al., 1970).
The Kuroshio Current originates from east of the Philippines and flows north along the east coast of Taiwan. After leaving Taiwan, the Kuroshio Current turns northeast along the edge of the East China Sea and frequently intrudes over the continental shelf (Chem et al., 1990). In contrast to the oligotrophic Kuroshio Curr ent surface water, a rather constant nutrient front is observed near the shelf break (Liu et al., 1988). Like other upwelling systems, this topographic process brings nutrient-rich water onto the shelf, and may promote primary production. The shallow water on the continental shelf off northeastern Taiwan has been a major fishing ground. The physical (Fan, 1980;Liu, 1983;Liu and Pai, 1987;Chem et al., 1990) and chemical (Liu et al., 1988;Wong et al., 1991) aspects of the upwelling process have been described. There is, however, no paper published on attempts to establish the biological aspects.
The present study investigated the chlorophyll a distribution and characteristics of the phytoplankton assemblages in the upwelled waters of the Kuroshio Current and the East China Sea. The relative importance of nannophytoplankton versus netphytoplankton in contributing to the phytoplankton standing crop, the distribution of chlorophyll a concentration, and the characteristics of the phytoplankton community structure were used as indicators to describe the upwelling phenomena.

Cruises
The present research was based on data collected during two cruises (cruise 248: August 1-8, 1990 and cruise 254: September 17-22, 1990) on the RN Ocean Researcher I of the National Science Council, Republic of China. Observations were made on two parallel transects, A and B (Figure 1), across the Kuroshio Current front off the northeastern coast of Taiwan. Transect A had nine sampling stations and B had ten. Adjacent stations in each transect were 28.3km apart in distance. At the 200-m isopleth, where the shelf front is located, one additional station was added in each transect. They were station 542A between stations 5422 and 5323 on the transect B and station 512A between stations 5121 and 5022 on the transect A. The Kuroshio main current flows through stations 5125, 5030, 4531 in transect B and stations 4523, 4424 in transect A.

Sample Collection and Analysis!
A rosette multi-sampler was lowered to collect water samples from 3, · 10, 25, 50, 75 and lOOm. Surface water (Om) was collected with a bucket. One half liter of seawater from each depth was preserved with neutralized formalin for species identification. Two 1-1 samples were taken for chlorophyll a measurement with several drops of M gC03 added. In order to understand the relative importance of phytoplankton of different size-groups in contributing to standing crop, phytophmkton were divided into two size groups based on their retention by fine-mesh nets. The water samples for chlorophyll a measurement were sieved through two stacked filters, a 10-µm mesh cloth (Nytex, Switzerland) on top of a Whatman GF/C glass fibre filter (Yamamoto et al., 1988). Those phytoplankton retained on the cloth screen were regarded as the netphytoplankton. Those on the GF/C filter were regarded as the nannophytoplankton. The filtration procedure was finished on board and samples were frozen at -20°C in darkness until analysis. The use of GF/C glass-fiber filters leads to no more than 10% underestimation of nannophytoplankton chlorophyll a (Venrick et al., 1987). Chlorophyll a and phaeopigments concentration were measured according to the pro cedures described by Strickland and Parsons (1972). The fluorescence was measured with a Hitachi F3000 fluorescence spectrometer. Each observation was duplicated and the mean of the replicates was reported. Integrated water column chlorophyll a ( C hl -a) concentration (mg·m-2 ) to lOOm was estimated by the following equation: where Ci and Ci+1 are the chlorophyll a concentrations (mg·m-3 ) and Di and Di+l are the depths of the upper and "lower limit of sample i.
Species identification and cell count were done on the surface samples obtained in all stations on the B transect. At least 500 cells per sample were examined. This sorting pro cedure ensured the bias of the community diversity statistics estimation, H', to be negligible (Mcintire and Overton, 1971). The filamentous blue-green algae, Trichodesmium spp.
were counted by their filament numbers.
Water temperature, salinity and nutrient data were adopted from the Chemical Oceanog raphy Data Bank of the National Science Council R. V. Ocean Researcher I Regional In strument Center. Water samples collected concurrently with those for the chlorophyll a measurements were used to determine the nitrate concentrations (Gong and Liu, 1991). Hy drographic results of temperature, salinity and nitrate along the transects A and B (Figures 2  and 3) were adopted partially from Gong and Liu (1991).

Data analysis
Series of diversity numbers presented by Hill (1973) were used to describe the phyto plankton assemblages and were calculated based on the observed number of species in each sample. They are: where S is the total number of species, where H' is Shannon's index and defi ned as: Pi is the proportional abundances of the ith species at the sample, where,\ is Simpson's index and defined as= � Pi2 These diversity numbers, which are in units of number of species, measure the "effective number" (Hill, 1973) of species present in a sample. No is the number of all species in the sample; N 1 is the :number of abundant species in the sample; and N 2 is the number of very abundant species.
The averaged linkage method of cluster analysis from SAS package program was used to compare. the species composition between sampling stations.
The ''aging index of upwelling" (AIU) proposed by Takahashi et al. (1986) was used to evaluate the aging status of upwelled water mass and is defined as: where Xis concentrations of chlorophyll a (mg·m-3 ) and Y is the concentrations of nitrate plus nitrite (µM) in seawater. AIU has a range of between 0 and 1. In newly upwelled water where there are insignificant amounts of phytoplankton, AIU will be 0. In old upwelled water where there are high levels of chlorophyll and low levels of nutrients due to active phytoplankton uptake, AIU will be 1. In the present study, AIU was calculated based on data from surface samples.

RESULTS
The 0-lOOm seawater chlorophyll a contents (netphytoplankton plus nanophytoplankton) varied between 8.63 and 41.80 mg·m-2 in August and between 8.88 and 94.79mg·m-2  in September. High nitrate content, low temperature and high salinity wer.e observed in both cruises at stations 5422, 542A, 5323 of transect B and stations 5220, 5121, 512A of transect A (Figures 2 and 3). This indicated active upwelling in those stations. High concentrations of chlorophyll a were also observed at these stations (Figures 2 and 3). Stations having surface N0 3 -N concentrations higher than 0.6µM had surface chlorophyll a concentrations higher than 0.5mg·m-3 (Figure 4) during September.  The vertical distribution of chlorophyll a in September was characterized by a subsurface maximum between 0-25m in stations associated with the upwelling, and between 50-75m in other stations ( Figure 5). The depths of nannophytoplankton chlorophyll a maxima always coincided that of netphytoplankton ( Figure 5). Comparison of the mean squares and the ranges of the chlorophyll a concentrations of nannophytoplankton and netphytoplankton (Table 1) indicated that the variations in netphytoplankton were either higher than or similar to those of nannophytoplankton. The August netphytoplankton and nannophytoplankton chlorophyll a concentrations ranged from 0.005 to 0.882mg·m-3 and 0.003 to 0.641mg·m-3 respectively.
In September, they were between 0.005 and 0.796mg·m-3 as well as between 0.007 and 0.930mg·m-3 , respectively.
Cluster analysis on the surface phytoplankton species compositions between the eight stations on the B transect of the August cruise indicated that the phytoplankton which appeared in stations 5422 and 542A were similar to those found in stations 5521 and 5030 ( Figure 6) which were on the shelf and on the main axis of the Kuroshio Current respectively ( Figure   1). These stations were grouped as one cluster. The species composition of the surface phytoplankton found in station 5323 was between the cluster mentioned above and the shallowest station of 6020. This indicated the mixing of the Kuroshio Curr ent water and the East China Sea coastal water in the area around stations 542A and 5323. High diatom (5.90 x 103 and 14.80 x 10 3 cells·I-1 , respectively) and low Trichodesmium   concentrations (0.17 and 0 fi laments· l-1 respectively) were also observed at these two stations ( Table 2). The remaining stations, in contrast, had diatom concentrations of between 0.37 x 10-3 and 1.01 x 10-3 cells·l-1 and Trichodesmium concentrations of between 0.24 and 36.44 filaments·I-1 , with the exception of station 6020. A high diatom concentration of 16.32 x 10-3 cells·l-1 was observed in that station ( Table 2).  S. costatmn also occurred in station 542A but with a lower density (0.89 x 10-3 cells·l-1 ) than that in station 5323. Trichodesmium thiebautii, which was reported to occur in mass quantities in summer in the Kuroshio Current waters (Marumo, 1957a;1957b), was observed in the present study in relatively large quantities in stations 6020 (36.52 x 10-3 filaments·l-1 ) and 5030 (4.94 x 103 filaments·I-1 ), but none in station 5323.
Similar results of dense diatom concentrations were also observed at stations 5422, 542A and 5323 in the B transect during the September cruise ( Table 2). The densities were 36.6 x 103 cells·l-1 , 12.76 x 103 cells·I-1 , and 13.2 x 103 cells·l-1 respectively. In contrast, densities of between 0.04 x 103 and 2.36 x 103 cells·l-1 were observed at their neighboring stations (Table 2) . During August pennate diatom Thalassionema nitzschioids became dominant instead of the centric diatom S. costatum. T. thiebautii was more abundant in station 542A with a density of 3.07 x 103 filaments·l-1 , while its densities in the remaining stations ranged between 0.04 x 103 and 0.15 x 103 filaments·I-1. · Table 3. Values of Hill's diversity numbers of the phytoplankton assemblages collected on surface water of the B transect in August AIU was smaller at the station 542A than its surrounding stations in both cruises (Table 4). Smaller AIU values of 0.11-0.16 were observed in September than in August (0.35-0.56) at stations 5422, 542A and 5323 (Table 4). This indicated a younger upwelled water mass occurred in September than in August. Markedly higher concentrations of nitrate plus nitrite (2.71-4.06µM) and chlorophyll a (0.66-0.72mg·m-3 ) were also observed in the upwelling stations in September than in August (0.09-0.30µM and 0.13-0.23mg·m-3 respectively) (Table 4). Waters along the 200-m isopleth of the study area showed high phytoplankton standing crops in terms of chlorophyll a concentrations. High nitrate concentration, high salinity and low temperature were also observed. The chlorophyll a maximum was shallower in these areas than in the neighboring waters. These findings agreed with the previous upwelling studies in these waters. The high standing crop and shallow subsurface chlorophyll a maxi mum likely were induced by the increased nutrient supply to the photic zone through vertical mixing. Lohrenz et al. (1988) studied the interrelationship among primary production, chlorophyll a and environmental conditions in frontal regions of the western Mediterranean Sea, and reported similar results to the present study. They found that the integrated pri mary production was inversely related to the depth of chlorophyll a maximum and further suggested that the inverse relationship is regulated by the rate of nutrient supply.
In the present study, while a high density of the blue-green algae Trichodesmium tbiebautii (4.94 x 103 filaments·l-1 ) was found in station 5030 on the Kuroshio Current, the highest density (36.52 x 103 filaments·l-1) among all stations was observed in sta tion 6020 which was far away from the Kurqshio Current. Trichodesmium is an oceanic species and is abundant (10 2 -103 fi_ laments·l-1 ) in summer throughout the Kuroshio Current water and almost absent in winter (Marumo and Nagasawa, 1976). Large blooms of Tri cbodesmium are usually found in oligotrophic oceanic waters. Due to its buoyant nature, surface water with dense Trichodesmium may be transported to a nutrient-rich region, such as station 6020, by storms or meander and ring formation of the Kuroshio Current.
Skeletonema costatum, a cosmopolitan centric diatom, was most dominant at the upwelling station 5323 in August. The water temperature was lower and the nutrients were higher at this station than its neighboring stations. The abundance of S. cost at um had been suggested to be influenced by the terrestrial materials brought into the sea by rivers (Huang, 1986). Abundant occurrences have also been observed in many estuaries or coastal waters of Taiwan (Huang and Chiang, 1974;Huang, 1986). On the other hand, Thalassionema nitzschioides, a pennate diatom occurred abundantly in the upwelled surface water during the September cruise, is also a cosmopolitan species. It is extremely eurythermal, euryhaline and circum-global in distribution and exists mainly in neritic areas.
The dominance of diatoms has been reported in many other upwelled waters. Tant (1976) observed continously occurring diatom blooms in large-scale upwelling waters over a long period of time. Even for a short duration of upwelling water, from days to a few weeks, active growth of diatom populations has been recorded (Yoder et al., 1981;Takahashi and Kishi, 1984). Diatoms became dominant when significant amounts of nutrients are supplied to surface or near-surface waters .
Dominance of diatoms in nutrient-rich water has also been demonstrated in laboratory tests. In a controlled experimental ecosystem where macro-nutrients (nitrate, phosphate and ammonia) were frequently introduced, phytoplankton communities subjected to high nutrients and strong solar radiation were dominated by chain-forming centric diatoms . The effects of these macro-nutrients on the growth of different species of algae were also evaluated by a dialysis bag culture study by Takahashi and Fukazawa (1982). They found that chain-forming S. costatum showed a greatly enhanced growth by the additions of nitrate, phosphate or ammonia. In contrast, growth of the algae was extremely poor at low concentrations of the macro-nutrients. Likewise, Turpin and Harrison (1979) described the dominance of centric diatoms under nutrient-rich conditions. S. costatum was one of the five diatom species dominated in the upwelled water off Peru (Guillen et al., 1971). In the present study the dominance of S. costatum in stations 5323 and 542A reflected the eutrophic condition resulting from upwelling. Turpin and Harr ison (1979) reported that in contrast to the dominance of centric di atoms under nutrient-rich conditions, flagellates are most abundant when nutrients are scarce, and pennate diatoms are favored under intermediate-nutrient conditions. However, in the present study, centric diatom S. costatum dominated when the ambient nitrate concen trations were low, while pennate diatom Thalassionema nitzschioides dominated when nutrient concentrations were high. In a simulated culture experiment of local upwelling in Japan pennates were often found to grow faster than centric diatoms .
However, the pennate diatoms settle to the bottom of the culture flasks in a much faster rate than do the centric diatoms. The reason diatoms are favored under high nutrient conditions could also be associated with the fact that strong turbulence slows the rate of sinking of the large cells from the photic zone (Longhurst and Harrison, 1989). Sinking rate of algae and the upwelling speed of the water mass, thus, may also influence the dominance of pennate or centric diatoms in an upwelled water.