Temporal and Spatial Changes of Chlorophyll a in the KEEP Study Waters off Northern Taiwan

The distribution data of chlorophyll a collected from 11 cruises between 1990-1993 in the KEEP study area was analyzed to demonstrate its temporal and spatial changes. Chlorophyll a concentration in the euphotic layer was generally low in March (0.02-0.67 mg.m-3) and high in November (0.02-3.79 mg.m -3 ). Sampling stations showed the highest chlorophyll a concentration during each cruise was mostly in the water along the brim of the East China Sea continental shelf, thus, indicating an association between the upwelling and bottom topography. Surface chlorophyll a concentration was negatively related to surface water temperature on all cruises. The distribution of the low surface water temperature was mostly caused by the upwelling. In win­ ter, however, the shelf was flooded with cold water from the north, and the chlorophyll a concentration was low. . The strength and the position of the center of the upwelling affected the dynamics of the chlorophyll a concen­ tration. When the upwelling strength was not strong enough to reach water surface, the surface chlorophyll a concentration at the upwelling stations re­ mained low. A significant negative relationship between the surface water temperature and the chlorophyll a concentration in all stations still existed because the chlorophyll a concentration in the warm Kuroshio water was even lower than in other regions. Dynamic temporal and spatial variations in chlorophyll a distribution were observed at the upwelling center surveyed repeatedly at one-week intervals. The vertical distribution of chlorophyll a showed a shallow subsurface maximum between 0 and 25 m in the upwelling region in contrast to between 50 and 75 m in the Kuroshio water. Surface chlorophyll a and the 0-100 m integrated chlorophyll a concentrations showed significantly positive correlations during nine out of the eleven cruises. (


IN TRODUCTION
The water on the continental shelf off northeastern Taiwan has long been a major fishing ground for both demersal and pelagic fi shes.A permanent upwelling caused by the intrusion of the meandering Kuroshio Current into the continental shelf of the East China Sea has been reported (Liu et al., 1992) in the area.Accordingly, cold and nutrient-laden water has been observed at the shelf break (Liu et al., 1992).High concentrations of chlorophyll a have occurred in summer (Chen, 1992) as well as in winter (Chen, 1994) in the upwelling region.
Chlorophyll a distribution has been regarded as an important index for the estimation of phytoplankton biomass and the capacity of its primary production.Inorganic nutrients, especially nitrate, are often a limiting factor constraining the growth of phytoplankton in the euphotic zone in subtropical and tropical oceans (Lalli and Parsons, 1993).Upwelling is, thus, one of the principal causes of nutrients coming up to the euphotic zone from deeper the water in these areas.
A 5-year (1989-1993) integrated research project, the Kuroshio Edge Exchange Pro cesses (KEEP), was initiated to study the dynamics of the physical, chemical and biological aspects of oceanography in describing the exchange processes of the mixing waters between the Kuroshio Current and the East China Sea continental shelf.The present report summarizes the spatial and temporal changes of chlorophyll a concentration in the aforementioned waters based on the results of 11 cruises conducted from 1990-1993.The effects of •upwelling and the related parameters on the chlorophyll distribution were emphasized to provide a better understanding of the mechanisms of the distribution and, thus, for better numeric prediction.

MATERIALS AND METHODS
During the 1990-93 period, oceanographic studies including chlorophyll a measure ments were conducted on 20 cruises in the KEEP study area.Complete data sets were obtained from 11 of the cruises.These observations covered every month of the year except for January, February, June and August.On all cruises, samples for the determination of chlorophyll a concentrations were collected from the R/V Ocean Researcher I using a rosette multi-sampler attached to a CTD probe.Two replicated I-liter water samples were taken from 0, 3, 10, 25, 50, 75 and 100 m.Filtration was conducted on board using Whatman GFIF glass-fiber or GF/C filters under low vacuum ( < 100 mm Hg).The fi ltered membranes were kept frozen in darkness at -20° C until analyses.Chlorophyll a concentration was mea sured according to the method of Strickland and Parsons (1972) for extracted samples using a fluorescence spectrophotometer.Integrated water column chlorophyll a concentration (Chi-a, mg.m-2) to 100-m depth was calculated from the following equation: where Ci and Ci+i are the chlorophyll a concentrations (mg.m-3 ), and Di and Di+ 1 are the depths of the upper and lower limit of sample i.
The data of water temperatures and nitrate concentrations from the Chemical Oceanog raphy Data Bank of the National Science Council R. V. Ocean Researcher I, Regional Instrument Center, were used in the statistical analyses.Because the nomenclature of the same sampling station varied on different cruises, new names were assigned to all stations studied.A total of 55 stations were identifi ed, and their localities together with their original names on each cruise are depicted in Table I.These stations covered the main area that was thoroughly studied in detail during different phases of the KEEP project.

RESULTS AND DISCUSSION
Chlorophyll a concentration in the study waters dynamically changed spatially and temporally.In the euphotic layer, the minimum value (0.02-0.67 mg.m-3 ) was observed in March (OCI-271) and the maximum (0.02-3.79 mg.m-3 ) in November (OCI-260).Surface (0 m) chlorophyll a concentration ranged from 0.03 to 3.79 mg.m-3 with higher values occurring in November and December and lower values (0.04-0.14 mg.m-3 ) in July.The Kuroshio water had the lowest chlorophyll a concentrations among all stations on each cruise ranging between 0.03-0.16mg.m-3 on the surface and 5.93-23.34mg.m-2 on the upper 100 m when stations 53, 54 and 55 were used as representative.Often the highest chlorophyll a c oncentrations of all stations on each cruise occurred at the shelf break, while an intermediate chlorophyll a concentration occurred in the shelf water.
When surface chlorophyll a concentrations of all stations on the 11 cruises were corre lated with surface water temperatures, significant negative relationships were observed (Table 2), indicating that the region with low surface water temperature had a high chlorophyll a concentration.Surface water temperature in the study area ranged from 16.7 to 28.9°C.Tem perature alone could explain 22-80% of the variations in chlorophyll a concentration (Table 2).The low surface water temperatures probably represent the upwelling phenomenon in most seasons and the coastal continental cold water in winter (cruise 271, Chen, -1994), whereas, .thenegative relationship of surface water temperature and the chlorophyll a concentration may also be reinforced or simply due to the low chlorophyll a concentration in the warm Kuroshio water.The latter was observed in Cruise 323 (Table 2) when the upwelled water was not brought up to the surface (Figure 1).The surface chlorophyll a concentrations, there fore, remained at low levels ranging from 0.04-0.14mg.m-3 in the study stations (Figure 2).The upper 100 m chlorophyll a concentrations were also very low (8.56-36.48mg.m-2).Among these low chlorophyll a concentrations, a significant negative relationship of surface chlorophyll a and surface water temperature still existed on the cruise due to the very low chlorophyll a concentration in the warm Kuroshio water (Figure 2).The result of this cruise also elucidated the importance of the upwelled strength.And the abundance of inorganic nutrients• to the phytoplankton biomass is especially critical in the euphotic surface water.
The relationship between surface chlorophyll a concentration and surface N0 3 -N level (Table 3) was, however, not as consistent as that between surface chlorophyll a concen tration and temperature.In these events, when significant correlations could be established (March, April, May and September), high N0 3 -N concentrations were accompanied by high chlorophyll a concentrations.No such correlations were established otherwise.It seems that temperature was generally a more indicative factor than was the N0 3 -N concentration in estimating the distribution of phytoplankton biomass.N0 3 -N is regarded as the most limit ing nutrient to regulate phytoplankton growth in most oceanic regions (Ryther and Dunstan, 1971;Goldman, 1975).:The negative relationship between chlorophyll a concentration and temperature was most likely to have been caused by the effects of the upwelling (Chen, 1992).The cold but nutrient-laden upwelled water supported the growth of phytoplank ton.The status of blooming, on the other hand, was decided by the maturation stages of the upwelling (Takahashi et al., 1986).In winter, nutrient-laden cold water from the north flooded the shelf, and the chlorophyll a concentration on the shelf was lower than that on the shelf break (Chen, 1994).Thus, besides nutrients, many other factors, such as grazing activity (Frost, 1991), concentration of trace metals (Martin et al., 1990;Minas and Minas, 1992), water temperature, light extinction rate and turbulence (J acques and Minas, 1981) might have worked individually or together to affect the chlorophyll a concentrations.The variation of the determination of coefficients (R 2) among cruises (Table 3) also reflected these possibilities.Stations with chlorophyll a concentration higher than 80% of the maximum value obtained in all stations in each cruise were located mostly at the brim of the continental shelf (Figure 3) where the bottom depth was• between 100 and 200 m.Several stations had such a high chlorophyll a concentration on more than 2 cruises.They were all located along the brim of the shelves (the Chilung Shelf and the East China Sea Shelf) and canyons (the Mien-Hua Canyon and the North Mien-Hua Canyon, Figure 4).This indicates that the upwelling mechanism is probably closely associated with the bottom topography when the Kuroshio water intrudes the shelf.And the strength of the upwelling and /or the location of the upwelling center seem to vary from time to time around the brim of the shelf where a permanent upwelling system was reported (Chem et al., 1990).The representative high chlorophyll a concentrations in the upwelled water around the shelf brim in November are shown in Figure 5. Similar phenomena for different seasons have been reported elsewhere (August and September in Chen 1992, March in Chen 1994).
Chlorophyll a concentrations in the shelf waters were usually between those of the Kuroshio and the shelf break.Occasionally, they were higher than that of the shelf break.During Cruise 260, the shelf stations (Stations 8 and 13) had higher surface chlorophyll a  concentrations (3.79 and 3.02 mg.m-3 ) and integrated upper 100 m chlorophyll a concen trations (139.9 and 218.8 mg.m-2) than those of Station 19 at the shelf break (surface concentration: 1.60 mg.m-3 and integrated upper 100 m concentration: 84.6 mg.m-2).The upwelling phenomenon was clearly observed at the shelf break (Figure 5) during this cruise, with low surface water temperature (21.59°C) and high nitrate concentration (3.37 µM).In contrast, shelf stations (Stations 8 and 13) had both relatively higher surface water tempera tures (23.14-23.40°C)and lower nitrate concentrations (0. 43-0.8 µM ).The mechanism of this high chlorophyll a concentration with low nitrate concentration at the shelf water warrants further study.
Although chlorophyll a concentration at the shelf break was usually higher than in the Kuroshio and the shelf waters, the chlorophyll a concentration at the shelf break varied with time.For example, in March the surface and the upper 100 m chlorophyll a concentraitons at the shelf break were 0.67 mg.m-3 and 31.79 mg.m-2, respectively (Station 19, Cruise 271), while in November they were up to 2.28 mg.m-3 and 66.5 mg.m-2 (Station 25, Cruise 260) and 0.88 mg.m-3 and 122.6 mg.m-2 (Station 30).These great temporal changes in chlorophyll a concentrations could also be manifested by comparing observations conducted during two consecutive cruises (Cruises 260 and 261) with a one-week interval.The surface chlorophyll a concentration at Station 19 dropped from 1.60 to 0.59 mg.m-3 and at Station 35 from 1.29 to 0.47 mg.m-3 , respectively as the upwelled center moved northwestwardly from Stations 19 and 35 (Figure 5) toward a new center closer to the shelf at Stations 10 and 23 (Figure 6).Thus, not only the dynamics of the upwelling strength but also the meandering of the upwelling center spatially and temporally affected the chlorophyll a distribution.
The vertical distribution of chlorophyll a in the study area showed a shallow subsurface maximum between 0 and 25 m in the upwelling region, while it was between 50-75 m in the Kuroshio water (Figure 7 and Chen, 1992).A similar vertical distribution with shallow subsurface maximum was observed at the shelf when its chlorophyll a concentration was high (Figure 8).Chlorophyll a is apt to develop a maximum near the surface in eutrophic waters and at greater depths in oligotrophic waters (Ichimura, 1956) .Regression analysis between integrated water column chlorophyll a concentration and respective surface chlorophyll a concentration in the present study indicates, in general, a positive relationship between the two (Table 4) . .Surface chlorophyll a concentration, accordingly, can be effectively (21 %-72% of the variation explained) used to estimate the water column chlorophyll a concentration which estimates all phytoplankton biomass in a water column.
The good correlations between surface water temperature and surface chlorophyll a concentration and between surface chlorophyll and integrated chlorophyll imply a possible direct link between surface temperature and integrated water column chlorophy IL The re gression analysis of the two variables indicates good correlations (Table 5).The fact that these regression coefficients are lower than those between surface and integrated water col umn chlorophyll indicates that the surface chlorophyll concentration is still a better parameter than surface temperature in estimating water column phytoplankton biomass.

Fig. 2 .
Fig. 2. Relationship of surface water temperatures and surface chlorophyll a concentrations collected on Cruise 323 (J uly 1992).

Fig. 3 .Fig. 4 .
Fig. 3. Location of the study site off the northeastern coast of Taiwan.Numbers indicate sampling stations.Stations having at least once a surface chloro phyll a concentration higher than 80% of the maximum value obtained at all stations on each cruise are circled within the dashed line.Among these, those having such high chlorophyll a concentration in more than two cruises are stippled.25°40'
Relationships between chlorophyll a concentration integrated over the upper 100 m (E Chl 1 0 o m ) and surface water temperature (Temp) ob tained between September 1990 and May 1993.

Table 1 .
Locations of the sampling stations and their original station names on respective cruises of the R/V Ocean Researcher I

Table 2 .
Relationships of surface chlorophyll a concentration (Chl) and surface water temperature (Temp) observed between September 1990 and May 1993.

Table 3 .
Relationships between surface chlorophyll a concentration (Chi) and nitrate concentration (N03) observed between September 1990 and May 1993.

Table 4 .
Relationships between surface chlorophyll a concentration (Chl) and chlorophyll a concentration integrated over the upper 100 m (EChl1oom) obtained between September 1990 and May 1993.