Primary production and phytoplankton in three Galician Rias Altas ( NW Spain ) : seasonal and spatial variability *

The coasts of Galicia (NW Spain) seasonally receive the influence of wind-driven upwelling pulses that allow high yields of exploitable marine biological resources. The upwelling causes the fertilization of coastal and shelf areas with deep water nutrients, in discrete events that can occur between March and October (Fraga, 1981; Blanton et al., 1984). The seasonality of surface winds, along with coastal morphology and the presence of several coastal embayments or rias favours biological production processes (Fraga and Margalef, 1979; Tenore et al., 1982). The resulting catches of fish and shellfish amount to about 1.6 x 1011 g y-1 over SCI. MAR., 62 (4): 319-330 SCIENTIA MARINA 1998

the shelf, and an equivalent figure of raft-cultivated blue mussels inside the rias (Xunta de Galicia, 1992).
The rias are unique systems because of their morphology and orientation with respect to the main coastline and, with their freshwater inputs, greatly influence the fate of the nutrients released by the upwelling (Blanton et al., 1984;Fraga, 1996).The Rias Baixas, located to the south of Cape Fisterra, are oriented in a NE-SW direction, which favours the net influx of upwelled water and high levels of primary production (Fraga and Margalef, 1979;Tenore et al., 1982).River runoff can be seasonally important in most Rias Baixas, contributing significantly to water dynamics and biogeochemistry (eg.Alvarez-Salgado et al., 1996).In contrast, the Rias Altas, located to the north of Cape Fisterra display a variety of orientations and receive smaller inputs of freshwater.The Rias Altas support a significant number of mussel rafts and local fisheries, although their total yield is lower than that of the Rias Baixas (Xunta de Galicia, 1992).Because of the economical importance of the Rias Baixas and their resources, there are several studies dealing with their primary production (see Varela et al., 1984 for a review).However, there are very few published data on this subject for the Rias Altas.(Bode et al., 1994a;Casas, 1995;Bode et al., 1996;Varela et al., 1996).
In this paper we describe the main features of phytoplankton biomass, species composition and areal estimates of primary production in characteristic stages of the seasonal cycle in the Rias of Ferrol, Ares and the Bay of La Coruña.These results are compared with those available for the Rias Baixas and the continental shelf of Galicia.A description of general hydrographic characteristics, dissolved nutrients and plankton observed during this study can be found in Varela et al. (1996).

METHODS
Daily irradiance was measured at our laboratory in La Coruña using a LiCOR 2π sensor.Large scale oceanographic conditions were estimated by the upwelling index values, calculated using the procedure described in Blanton et al. (1984) and wind data provided by the Centro Zonal de La Coruña (Instituto Español de Meteorología).Water samples and observations were collected during COPLA-393 (11-29 March 1993), COPLA-893 (3-10 August 1993) and COPLA-1293(15-20 December 1993) cruises, as part of a large scale study on the continental shelf off La Coruña (Varela et al., 1996).In the present study we include results from the stations located inside the Bay of La Coruña, the Rías of Ferrol and Ares, and the adjacent coastal areas (Fig. 1).Temperature, salinity, and irradiance profiles were obtained at each station with a CTD Seabird SBE-25 equipped with a LiCOR spheric sensor.Water and phytoplankton samples were collected using Niskin bottles at the depths of 100, 50, 25, 10 and 1 % of the incident irradiance (I 0 ).Additional samples were taken at depths below the euphotic zone where bathymetry permitted.Chlorophyll-a concentrations were determined fluorometrically in samples from the euphotic zone in all stations.These samples were filtered through Whatman GF/F filters, and pigments were extracted in 90% acetone and analysed using the procedure described by Yentsch and Menzel (1963).Phytoplankton composition was determined in Lugol preserved samples from the euphotic zone in selected stations (Table 1).
Primary production was measured by two procedures in stations visited before noon (Table 1).The results from a nearby station (Sta.14, 43°20.74'N, 8°34.38'W, 65 m depth) were included to complete the production values for the March cruise.Aliquots from each light intensity level within the euphotic zone were inoculated with 4 µCi (148 kBq) of 14 CO 3 Na and were incubated on board for at least 2 h around noon in simulated in situ (SIS) conditions.The incubator was refrigerated by running surface seawater and illuminated by sunlight.Light levels were simulated using neutral density screens.In addition, photosynthesis -irradiance (P-I) curves were determined in 60 ml aliquots from water collected at the surface (100 % I 0 ) and at the base of the euphotic zone (1 % I 0 ), inoculated with 2 µCi (74 kBq) of 14 CO 3 Na and incubated for up to 2 h in a linear incubator.These aliquots were illuminated with halogen lamps of 2 mmol quanta m -2 s -1 of initial irradiance.The incubator was refrigerated with surface seawater.All incubations were terminated by filtration onto Whatman GF/F filters, that were acidified with two drops of 5 % HCl overnight and counted on a LKB Wallac 1409 liquid scintillation counter using Instagel‚ as the scintillation medium.The P-I curves were adjusted using the functions described by Platt and Jassby (1976), in the case of no photoinhibition, and Platt et al. (1980) in cases of apparent photoinhibition.The chlorophyll-a normalized photosynthetic parameters α B (production eficiency, mg C mg Chl-a -1 h -1 µmol quanta -1 m 2 s), P B S (light-saturated production rate, mg C mg Chl-a -1 h -1 ), P B M (maximum production rate, mg C mg Chl-a -1 h -1 ), β B (photoinhibition, mg C mg Chl-a -1 h -1 ), and saturation irradiance (I k µmol quanta m -2 s -1 ) were used to describe the photosynthetic properties of the phytoplankton in each station and to compute areal estimates of primary production.
Primary production estimations in the whole study area were made using the procedure of Sathyendranath and Platt (1993).First, primary production rates at each sampling depth were computed using the photosynthetic parameters α B and P B M , and the irradiance received at that depth.The resulting chlorophyll-normalized production values were then converted into carbon units using the chlorophyll concentration measured at the same depth.Water column rates were obtained by integration to the depth of the euphotic zone.The integral was com-PRIMARY PRODUCTION IN RIAS ALTAS 321 puted by numerical integration of 1 m intervals, estimating values located between the discrete sampling depths by exponential smoothing.For comparative purposes we used the average daily irradiance value measured during each cruise and the depth attenuation coefficient measured at each station to estimate the irradiance received at each depth.The daily primary production rates obtained from this method were compared to estimations obtained by the SIS method.

Oceanographic conditions
The cruises were representative of significant oceanographic situations related to the annual cycles of upwelling and solar irradiance (Fig. 2).The COPLA-393 cruise was made after the first marked upwelling pulse of the year, with positive values of the upwelling index for several days before the cruise and average irradiance values of ca.20 mol quanta m -2 d -1 .The summer cruise COPLA-893 was also made in favourable upwelling conditions, but in this case the irradiance values were near the annual maximum.In contrast, mostly downwelling conditions and very low irradiances occurred during the COPLA-1293 cruise.
Surface temperature and salinity values indicated that the stations located inside the rias received little influence from freshwater discharges during the March cruise (Fig. 3 A).Most of the area had surface temperatures between 12.4 and 13°C, that were close to the annual minimum in this region (Casas et al., 1997).Chlorophyll concentrations were high near La Coruña (ca.150 mg Chl-a m -2 ), but reached very low values in the Ria de Ares and in the open shelf stations.In contrast, the Ria de Ferrol had relatively high chlorophyll concentrations (up to ca.50 mg m -2 ).
Higher temperature variations were found during the August cruise, when salinity values inside the rias, but also at the shelf, were lower than in March (Fig. 3 B).As a result, strong salinity gradients ocurred from the inner to the outer stations in the rias.The prevailing upwelling conditions during this cruise were not clearly reflected in the surface, but the relatively low surface temperatures and high chlorophyll concentrations in the stations near La Coruña suggested that shelf stations in the western part of the study area were affected by the upwelling.In December surface temperatures showed minima values near the coast and inside the rias (Fig. 3 C).Surface salinity reached maximum values during this period, and chlorophyll concen-trations were in general low, reaching values higher than 20 mg m -2 at open shelf stations.

Vertical distribution
The vertical distribution of these variables in a transect from the Ria de Ares to the open shelf, shows the existence of a body of relatively warm water that extends from the surface to ca. 40 m depth near the mouth of the ria in the March cruise (Fig. 4 A).The freshwater influence was restricted to the ria and caused marked stratification in the upper 20 m of the water column.In contrast, shelf waters were more homogeneous, with salinity values between 35.75 and 35.80 psu.Chlorophyll concentrations in this transect were higher inside the ria, but these values were low when compared with those of Ria de Ferrol or the stations near La Coruña.
During the August cruise, the stratification of the upper layer of the water column was well defined in the whole transect (Fig. 4 B).The influence of the upwelling was not apparent except in the subsurface layer of the deepest station that had water with salinity values higher than 35.80 psu.Chlorophyll concentrations were generally high in the upper 20 m and inside the ria, where values higher than 5 mg m -3 were found near the bottom.The stratification of temperature and salinity values found during the December cruise affected only to the coastal stations, while the outermost station displayed a vertically mixed water column (Fig. 4 C).Chlorophyll concentrations were very low at all stations but a surface maximum of ca.0.8 mg m -3 was measured near the mouth of the ria.
The study area was divided in four zones corresponding to La Coruña Bay (Sta.3b, 3c and 4), Ría de Ares (Sta. 31,32,33 and 34),Ría de Ferrol (Sta. 35,36 and 37) and the open shelf (Sta. 2,3a,20 and 38).The profiles of chlorophyll were quite homogeneous in all zones during March and December, but subsurface maxima were clearly defined in August, particularly in the open shelf (Fig. 5).Most of these subsurface maxima were near 5 m depth in the rias and near 10 m depth in the open shelf.SIS primary production profiles showed the maximum values close to the surface in December, but in a subsurface layer in all other cruises.These subsurface maxima were generally above the chlorophyll maxima, but in the open shelf stations during the August cruise both maxima occurred at the same depth (Fig. 5 B).The largest differences in chlorophyll and production profiles between cruises were in the absolute values reached, where values from the December cruise were ca. 10 times lower than those of March.
The maximum production values measured were those of St. 4 in La Coruña Bay during August (ca.50 mg C m -3 h -1 ).

Phytoplankton species composition
Flagellates dominated phytoplankton abundance distributions in nearly all cruises (Table 2).However, the diatom Chaetoceros socialis Lauder reached high abundances in March, while Leptocylindrus danicus Cleve and Skeletonema costatum (Greville) Cleve were dominant in August.Small dinoflagellates were always present, but species of this group had higher abundances in summer when red-tide forming organisms, like Alexandrium lusitanicum Balech and Prorocentrum micans Ehrenbergh, exceeded on average 90 cells ml -1 .Abundance values of phytoplankton were low during the December cruise (< 30 cells ml -1 , Table 2).There were some differences in the composition of phytoplankton in the three rias, particularly in the March cruise, when the Ria de Ferrol and the stations near La Coruña had the highest abundances of Chaetoceros socialis and the Ria de Ares and the outer shelf zone were dominated by flagellates.

Photosynthetic characteristics and areal production
The photosynthetic parameters α B and P B M were significantly different between cruises (ANOVA Kruskall Wallis test, n = 18, p < 0.001) and there were also statistically significant differences in I k (p < 0.05).However there were no significant differ-PRIMARY PRODUCTION IN RIAS ALTAS 325 ences in these parameters between the different zones (p > 0.05).Therefore, photosynthetic parameters of each cruise were averaged in order to compute primary production in all stations (Table 3).Phytoplankton in March was well adapted to the increasing radiation received during the spring, as evidenced by high P B M values in surface waters.However it was also well adapted to relatively low irradiance values because of the high values of a B and average values of I k lower than 150 µmol quanta m -2 s -1 .The differences between P-I curves from the surface and the bottom of the euphotic zone, particularly in the apparent photoinhibition response of samples from the latter, indicated the presence of a vertical stratification within the euphotic zone.The adaptation to high irradiances was more evident in the samples from the August cruise, with characteristic low α B and high P B M values.Saturation irradiances exceeded 200 µmol quanta m -2 s -1 and the values of the photoinhibition parameter β B reached maximum values measured during this study.In summer there were no clear differences in the values of photosynthetic parameters between the upper and lower limits of the euphotic zone (Table 3).This can be explained by the frequent occurrence of upwelling pulses through the summer, that cause the movement of phytoplankton from deep layers to high irradiances near the surface.In December the phytoplankton was still adapted to relatively high irradiances, indicated by low values of α B and I k values similar to those of March.However, the low P B M values and the low degree of variation between surface and deep samples indicated that mixing events were frequent in the water column and that the production was probably light limited.There was a significant linear correlation between water column estimates of daily primary production using the average values of a B and P B M and the values measured in some stations with the SIS method (r = 0.819, n = 16, p < 0.001).The slope of the regression line between these variables was not significantly different from 1 (1.343 ± 0.251, s.e.) and the intercept was not significantly different from 0 (-22.269 ± 280.851), as determined by Student-t tests (p < 0.05).Consequently, the primary production values estimated from P-I parameters and chlorophyll profiles are statistically equivalent to those measured by the SIS method, more frequent in other studies of primary production made in Galician waters (e.g.Bode et al., 1996).
Primary production rates were higher in the western stations of the study area during March and August, both with equivalent maximum values exceeding 3000 mg C m -2 d -1 (Fig. 6 A, B).The spatial pattern of primary production was also similar in both seasons, excepting the lower production rates of the Ria de Ares in March.In contrast, primary production values estimated for the December cruise were ca. 1 % of those measured during previous cruises.In this case, only one station located near the mouth of the Ria de Ares reached more than 20 mg C m -2 d -1 , while maximum values in the open shelf stations were ca.14 mg C m -2 d -1 (Fig. 6 C).The mean values measured during the cruises (Fig. 7) were statistically significant (ANOVA, p < 0.001), but the values of the March and August cruises were equivalent and significantly higher than the mean production measured in December ('a posteriori' Student-Newman-Keuls test, p < 0.05).However, due to the large variations found in each zone described, there were no significant differences in daily production estimates between zones (ANOVA, p > 0.05).

DISCUSSION
The described variations in biomass and species composition are consistent with the annual cycle and the oceanographic stages related to the phytoplankton described for the coast of La Coruña (Valdés et al., 1991;Bode et al., 1994a;Casas, 1995;Bode et al., 1996;Casas et al., 1997).However, maximum abundance and chlorophyll concentration values found during this study were lower than those reported previously (Casas, 1995).This can be attributed to the occurrence of the main upwelling peaks before the cruise (as in the case of March) or at the end of the cruise (as in August) and a delay in the response of phytoplankton to the upwelling forcing of up to 13 days (Casas, 1995).There is also a correspondence between our findings and the seasonal succession of phytoplankton in other Galician rias (eg.Margalef et al., 1955;Mariño et al., 1985;Figueiras and Niell, 1987).Therefore, our study provides results characteristic of phytoplankton blooms (spring and summer cruises) and of low phytoplankton abundance, biomass and production (winter cruise).
Phytoplankton blooms are caused in the study area by two main phenomena.On one hand, blooms produced during the spring can be originated by small density gradients near the surface of the water column, as those found in the Ria de Ferrol in this study.The first occurrence of these blooms is generally in March (Casas, 1995;Casas et al., 1997), when the mean daily irradiance at the sea surface is greater than 20 mol quanta m -2 s -1 .The density gradients can be caused by freshwater runoff, as in the rias, or by the upwelling.In the latter case, the upwelling cause the movement of water from the ocean to the shelf, where water was generally well mixed at this time of the year (Fraga, 1981;Casas et al., 1997).In some cases, water with a salinity higher than the shelf water can be found near the shelf break, as part of the poleward current that flows seasonally in the Eastern North Atlantic (Frouin et al., 1990).This water may reach inner shelf and coastal areas causing density gradients in the water column (Varela et al., 1996;Casas et al., 1997), and can be detected in our March cruise as the shelf water with a salinity higher than 35.75 psu (Fig. 4 A).A similar phenomenon was described by Castro et al. (1994) in the same area.The role of upwelling pulses during these blooms would be more related to the creation and maintenance of the stratification by forcing the penetration of a dense watermass into the shelf than to the fertilization with new nutrients, because nutrient concentrations in shelf water are already high at this time of the year (Valdés et al., 1991;Varela et al., 1996;Casas et al., 1997).
In contrast, blooms produced in late spring or in summer, as in our August cruise, would be more dependent on the nutrient enrichment caused by the upwelling.Even when surface nutrients are rarely depleted in this coast, their concentrations in the surface during summer generally do not exceed 0.5 mmol m -3 in the case of nitrate (Valdés et al., 1991;Casas et al., 1997).Upwelling events can introduce a significant amount of new nutrients into the euphotic zone (Bode and Varela, 1994), allowing for discrete phytoplankton blooms that can be traced through most of the spring and summer (Casas et al., 1997).During the August cruise surface nutrient concentrations were low, but shelf stations located to the north of the Ria de Ferrol had more than 2 mmol nitrate m -3 and temperatures lower than 14.5°C at 10 m depth, indicating a recent upwelling (Varela et al., 1996).There was also a difference in the composition of phytoplankton populations between spring and summer blooms.Spring blooms were typically dominated by Chaetoceros socialis (as in our March cruise) and other species of the same genus, while the composition of phytoplankton during summer blooms was often more diverse (Casas, 1995).In our study, the phytoplankton of the summer cruise was dominated by two species of diatoms, but several flagellates and dinoflagellates reached significant abundances.The high abundance of flagellates in all cruises may indicate a high potential for the oxidation of organic matter in the water column.Varela et al. (1996), based on counts of flagellates with and without autofluorescence, reported that only during the March cruise was there an equivalent number of autotrophic and heterotrophic cells, while during the August and December cruises more than 80% of cells were heterotrophic.This result supports the hypothesis that a large fraction of the organic matter produced by phytoplankton after an upwelling pulse in summer is consumed in the water column by the existing microbial populations (Bode and Varela, 1994).
Our measurements indicate that in all cruises phytoplankton cells were well adapted to the existing light regimes.Maximum production values (P B M ) occurred near the surface and during summer, and maximum saturation irradiances were always close to incident irradiances at noon in each season.The largest differences in photosynthetic parameters occurred between the seasons studied.This feature, common in coastal and temperate areas (eg. Schofield et al., 1993), contrasts with the constancy of these parameters in oceanic environments (Platt et al., 1992).The pattern described by our values of α B differs with that of average seasonal values for coastal areas in the north Atlantic reported in Sathyendranath et al. (1995), because in our case the values of α B and P B M observed in spring were comparable or even higher than those observed in summer.The later authors recognize that the coastal environments are heterogeneous and that the effect of local conditions in photosynthetic parameters may be of greater importance than seasonal changes.This emphasizes the value of experimental estimations of these parameters in local areas where major environmental changes occur, like the upwelling forcing in our case.
The fact that our results do not show a statistically significant depth dependence of either α B and P B M support the hypothesis that mixing phenomena, involving most of the euphotic zone, are very frequent in the studied area even during summer, the season where thermal stratification is likely to occur.Several authors (eg.Lewis et al., 1984 andLizon et al., 1995, among others) have shown that phytoplankton populations can be physiologically heterogeneous only in moderately mixed water columns.Upwelling pulses, that occur in this area with a periodicity of 4 to 22 days (Casas, 1995), would allow phytoplankton cells to be exposed to high irradiance levels as the picnocline is moved upwards.In addition, there are other physical processes that can produce the same effect, like tides and internal waves.As a consequence, these cells are adapted to frequent oscillations in the light field, as in other upwelling areas (eg.González-Rodriguez, 1994).It is not surprising that diatoms are a significant fraction of phytoplankton in the study area because they can tolerate large variations in irradiance (Sakshaug et al., 1987).However, other studies have shown that photosynthetic parameters at various depths within the mixed layer can be very different, following also circadian variations (Lizon et al., 1995).Therefore, the values provided in the present study must be taken as preliminary, and more experimental data are required in future studies for a better understanding of light adaptations of phytoplankton in this area.
Average water-column integrated primary production values during either spring and summer cruises were comparable to those measured in the Galician shelf (Bode et al., 1994a;1996).In contrast, values for the winter cruise were among the lowest recorded in this region.The highest values of chlorophyll and primary production were found in shelf stations near La Coruña.Primary productivity data collected over several years in the mid-shelf zone often exceeded values measured near the coast (Bode et al., 1996).Also several studies have pointed out the existence of mid-shelf and shelf-break upwelling areas in spring and summer (Varela et al., 1991;Tenore et al., 1995;Bode et al., 1994b;Prego and Bao, 1997).In our case study, the thermohaline front that divided the shelf in a north-south direction during the March cruise (Varela et al., 1996) probably was caused by an intrusion of a saline wedge of the poleward current over the shelf, having associated high chlorophyll concentrations and primary production values.
Published data of daily primary production in the Rias Baixas are similar to our measurements (Table 4).Fraga (1976) report for the Ria de Vigo maximum and mean primary production rates of 2.82 and 0.71 g C m -2 d -1 respectively.González et al. (1982), in their study of the Ria de Pontevedra in winter, cite values between 0.05 and 0.46 g C m -2 d -1 .Tenore et al. (1982) report values for the Ria de Arousa that range between 0.03 and 0.10 g C m -2 d -1 in winter and are higher than 1 g C m -2 d -1 during phytoplankton blooms in spring and autumn.Even when the resulting value of annual primary production for La Coruña (608 g C m -2 y -1 ) is somewhat higher than those measured by the SIS method at the same stations by Casas (1995), the small surface and volume of the Bay of La Coruña compared with the two Rias Baixas gives a total annual value that represents only from 10 to 25 % of primary production in the Ria de Arousa.At the same time, this value is very similar to that of Ria de Ares, three times larger.The lower absolute value of total annual production corresponds to the Ria de Ferrol, because of its small size.
The results obtained in this study are consistent with the hypothesis of a differential effect of the upwelling in the Rias Altas when compared to the Rías Baixas.In the latter, the upwelling dynamics cause an export of the organic matter produced inside the rias to nearby shelf areas, that display sediments enriched in organic matter (López-Jamar et al., 1992).In the Rías Altas, the irregular distribution of upwelling pulses over the shelf could be responsible for the low organic content (López-Jamar et al., 1992) and the distribution of biogenic silica (Prego and Bao, 1997) of shelf sediments in this area.There is evidence of low sedimentation of particulate matter near the coast of La Coruña during active upwelling pulses (Bode et al., 1998).Most of the produced organic matter could either be exported to offshore areas by the dominant surface circulation or remineralized in the water column (Bode and Varela, 1994).

Fig. 1 .
Fig. 1. -Map of the study area with the position of the sampling stations.
FIG. 4. -Spatial distribution of temperature (°C), salinity (psu) and chlorophyll concentration (mg m -3 ) in a transect from the inner part of the Ria de Ares to the open shelf during the cruises COPLA-393 (A), COPLA-893 (B) and COPLA-1293 (C).
FIG. 5. -Vertical profiles of chlorophyll concentration (closed symbols and continuous lines, mg m -3 ) and SIS primary production rates (open symbols and dashed lines, mg C m -2 h -1 ) in four zones during the cruises COPLA-393 (A), COPLA-893 (B) and COPLA-1293 (C).The station number is indicated by arrows.
326 A. BODE and M. VARELA

TABLE 1 .
-Stations where simulated in situ production incubations (SIS), phytoplankton species composition (PSC) or P-I curves were measured during the different cruises.

TABLE 2 .
-Abundance (cells ml -1 ) of the dominant phytoplankton species in each sampling period.sd: standard deviation, n: number of samples.

TABLE 3 .
-Photosynthetic parameters obtained from P-I curves averaged by sampling periods (mean and standard deviation, sd) of P-I curves obtained for each cruise.Parameter units and symbols as described in the methods section.