Photosynthesis-irradiance relationships in natural phytoplankton populations of the Barents Sea FRANCISCO REY Rey. F . 1991: Photosynthcsis-irradiance relationships in natural phytoplankton populations of the Barents Sea. Pp.105-116 in Sakshaug, E . . Hopkins. C. C . E. & Britsland. N . A . (eds.): Proceedings of the Pro Mare Symposium on Polar Marine Ecology, Trondheim. 12-16 May 1990. P o / a r Research I O ( 1 ) . An analysis is made of the photosynthesis-irradiance relationships in natural phytoplankton populations in the Barents Sea. The data set comprises 232 experiments carried out during a 10-year period, both in open and ice-covered waters. The variability on the P-1 parameters is discussed and examined in relation to the variation in a variety of environmental conditions. The results suggest that in the Barents Sea. as in other Arctic areas, phytoplankton photosynthesis is mainly controlled by physical variables. However, control of the phytoplankton stock. i . e . by zooplankton grazing, seems also to have a considerable indirect influence on P-I parameters, especially after the spring bloom and the depletion of winter nutrients. “ ‘ ‘ . ~ L A R \ N S ~ ’ ’ Francisco R e y . Institute of Marine Research. P . 0. Box 1870 Nordnes. N-5024 Bergen. N o r w a y (revised M a y 1 9 9 1 ) . Introduction As in other polar seas, phytoplankton in the Bar- ents Sea are exposed to extreme seasonal vari- ations of such environmental conditions as day length, available light, water column stability, and water temperature. Especially important is the fact that the Barents Sea is dominated by two different water masses: cold Arctic waters flowing in from the north and warmer Atlantic waters flowing in from the south, giving rise to a marked front. In addition, the northern halfof the Barents Sea is seasonally covered by ice. Since 1979, the Institute of Marine Research has carried out biological oceanographic inves- tigations in the Barents Sea. Studies have been focused mainly on the feeding conditions of the capelin stock in order to elucidate the large annual and geographical variations i n the growth of this fish species (Skjoldal & Rey 1989). These inves- tigations were incorporated into the Norwegian Research Program for Marine Arctic Ecology (Pro Mare) in 1984 and were concluded in 1989. During this period, a series of phytoplankton photosynthesis-irradiance experiments were car- ried out. The aim of these experiments was to characterise the primary production potential of the Barents Sea. The data set obtained from these experiments is presented here and analysed with respect to the relationships between the par- ameters of the photosynthesis-irradiance curve and other environmental parameters. For logis- tical reasons the bulk of the data was collected during the boreal spring and summer. Material and methods The data employed in this analysis of phyto- plankton photosynthesis were derived from 232 photosynthesis-irradiance experiments carried out between 1980 and 1989. Samples represent different seasons, types of water masses, ice cover, and depths. The samples were usually taken from the upper mixed layer, the subsurface chlorophyll maximum (if present), and from below the pycnocline. Fig. 1 shows the location of the sampled stations while Table 1 presents an overview of the collected data. Two types of incubators were used for meas- uring the uptake of radioactive carbon. The first, called the low-light incubator (LL), was equipped with daylight-type fluorescent tubes (OSRAM 191 Daylight 5000 de Luxe) that, combined with neu- tral density plexiglass filters, gave ten different quantum scalar irradiances in the range from 0 to 390 pmol m-2 s-I. The sample volume was 100 ml of seawater, to which 148 kBq (4 pCi) of radio- active carbon were added (The International Agency for I4C Determination, Denmark). The incubation time was about 4 hours. The second incubator, a high-light type (HL), was equipped with an halogen-metal lamp (OSRAM Power 106 Franckco Rev 80 -. 78 - 16 - 74 - A F Y 0 72 i I t 10 20 30 40 50 Imgitude ( 'E) Fig I Station locations in the Barents Sea Star, 400 W ) that, combined with neutral density filters, gave 11 different intensities in the range from 0 t o 1700 pmol m - ' s - ' of quantum scalar irradiance. To a sample volume of about 300 mi. about 1480 kBq (40 pCi) was added. After gentle a thermostat-controlled water bath. In all cases, incubation was terminated by immediate filtration through Millipore membrane filters of 0.45 pm pore size. Thereafter, the filters were either kept frozen at -18°C for further treatment ashore or analysed on board. T h e filters were fumed for 15 minutes with concentrated hydrochloric acid, after which 7 ml of InstaGel ( u p t o 1986) or Opti- Fluor was added. T h e radioactivity o n t h e filters was measured with a Packard Tri Carb scin- tillation spectrometer using t h e channel-ratio method for estimating counting efficiency. T h e photosynthesis-irradiance curve parameters were estimated using the model described by Platt e t al. (1980): p B = P:(I - e-aB'/p:)e-flB'/pfl PB: photosynthetic rate [mgC (mgCh1)-' h-'1 P! : light-saturated photosynthetic rate in the absence of photoinhibition [mgC (mgChl)-' h-'1 aB: initial slope of t h e P-I curve [mgC (mgChl)-' h-' (pmol m-' s - ' ) - ' ] . I : quantum scalar irradiance ( p m o l m-2 s-I) BB: index of photoinhibition (mgC (mgCh1)-' h-' (pmol m-? s - ' ) - ' ] mixing, 20ml aliquots were poured into Pyrex bottles with a n automatic dispenser and incubated for about 2 hours. When working in open sea areas, the temperature in t h e incubators was kept After fitting t h e equation t o t h e data (Nelder & Mead 1965), the following derived parameters were calculated (Platt e t al. 1980): at in situ temperature by a continuous flow o f P k = P?[cuB/(aB + B B ) ] * [/3"/($ + PB)]@B'"B, subsurface seawater (about 4 m depth). In ice- the maximum observed photosynthetic covered waters and in front areas, the tem- perature in the incubators was maintained at I, = p:/(~", an index of photoadaptation 51°C of the in situ temperature by means of rate [mgC (mgChl)-' h-'] (pmol m-' s-') T o b k 1. Overview of the data sets utilised Number of Year observation$ Month Sampling Incubator Main type of depth, ( m ) 'YPC watermasses 1980 1981 1982 1983 1989 3 '9 8 49 37 ZH 27 39 12 July May. July. August May. June May. June Junc. August August April February. May. June May -5-35 0-45 (L30 6 8 0 LL L L L L LL LL LL LL H L H L Open waters Open waters Open waters Ice-covcrcd waters Open waters Open waters Open waters Ice-covered waters Ice-covered waters Open waters Open water5 LL = Low-light incubator H L = High-light incubator Photosynthesis-Irradiance relationships in the Barents Sea 107 I, = PF/aB In[(# + pB)/pB], the irradiance at which photosynthesis is at maximum (ymol m-* s-l) Ib = P:/flB, an index for photoinhibition sus- ceptibility (ymol m-2 s-') Chlorophyll a concentrations were determined by filtering known volumes of seawater through membrane filters of 0.45 ym pore size. The filters were kept frozen for a few days and then dissolved in 90% acetone. The pigments were extracted overnight at 4'C, centrifuged, and then their fluorescence was measured before and after acidi- fication in a Turner Designs fluorometer. The fluorometer was calibrated against commercially purified chlorophyll a (Sigma Chemical Co.). Nutrients were measured onboard immediately after collection with standard methods adapted to an autoanalyser (Foyn et al. 1981). Results and discussion Environmental parameters The average seasonal distribution of PAR (photosynthetically active radiation) at Bj0rncbya (74'15") which lies at the mean latitude of the Barents Sea is shown in Fig. 2. During the Arctic winter, incident light is hardly detectable. The first recordable light appears in mid-February and increases exponentially thereafter and reaches a maximum total daily irradiance in mid-June. The greatest temporal increase in incoming irradiance is usually observed in April. The average hourly irradiance, however, increases exponentially somewhat earlier than the total daily irradiance and levels off at the end of April. Hydrographic conditions were extremely vari- able, ranging from relatively warm Atlantic waters of high salinity to cold Arctic water masses, in many cases near the freezing point. Fig. 3 shows the temperature-salinity relationships at the depths from where the water samples were col- lected. About 60% of the samples were collected either from Atlantic or Arctic Water proper and about 25% from meltwater. The remainder was collected in areas with different degrees of mixing between the main water masses. Variations in P-I parameters The frequency distributions of the parameters of the P-I curve for the whole investigated period are shown in Fig. 4A-E. For most of the par- ameters, a positive skewness of the values was evident, as has been found in previous polar inves- tigations (Harrison & Platt 1986). Pft, ranged from 0.27 to 7.3 mgC (mgChl)-' h-' with a mean of 1.73; a" ranged from 0.0037 to 0.1633 with a mean of 0.0440 mgC (mgCh1)-' h-' 0 60 120 180 240 300 360 Julian day Fig. 2. Average (lY7G-lY89) seasonal distribution of photosynthetically active radiation (PAR) at Bjarnaya. A . Total daily irradiance. B . Average hourly irradiance (lit hours). 108 Francisco Rey h 32.0 33.0 34.0 35.0 Salinity Fig. 3 Temperature-salinity relationships at the sampling depths. A . Atlantic Water: B . Arctic Water: C I . Spring mclt water: C2 Summer melt water: D Mixed Atlantic and mclt- waters in surnmer/autumn: E . Mixed Atlantic and Arctic W A l C r S (pmol m-? s - ' ) - I ; BB ranged from 0 to 1.085 mgC (mgChl)-' h-' (pmol m-? s - ' ) - ' with a mean of 0.0312; I, ranged from 19 to 891 pmol m-'s-I with a mean of 190; IL ranged from 4 t o 3 1 9 p ~ 1 o l r n - ~ s - ~ with a mean of 47; I h ranged from 49 to infinite. For PB and I h , the number of observations was reduced to 194 by excluding a series of experiments for which the light intensity of one of the incubators (LL) was not sufficient for obtaining a proper estimate of these parameters. Harrison & Platt (1986) found that high latitude phytoplankton populations exhibited lower val- ues for P i . I,, and I, as well as higher aB values than phytoplankton of temperate latitudes. Such differences were not always found in o u r data set. The median P: for all samples was 1.60 mgC (mgChl)-' h - I . which is somewhere between t h e median values of 1.21 and 2.39 found for Arctic and mid-latitudes in eastern C a n a d a . T h e Barents Sea is an area typically characterised by cold Arctic waters in the north and relatively tem- perate Atlantic waters in the south. Assuming that the overall median value for t h e Barents Sea represents an "average" value for temperate and cold waters. the data set was divided into two groups representing temperatures above and below 0°C respectively. This temperature is com- monly used as a boundary between Arctic and Atlantic waters (Loeng 1989). A single A N O V A was then applied t o determine if there were significant differences in the mean values of the P-1 parameters between t h e two subgroups. Only P i and I , mean values were found t o be sig- nificantly different at t h e 5% significance level. Mean Pi was 1.49 below 0°C and 2.01 above 0°C. This is as expected when it is taken into con- sideration that the maximum photosynthetic rate of natural populations is usually a function of temperature. However, o t h e r factors such as species composition, cell size, and biomass might also have an influence o n these differences. Mean I, values were 41 and 54 pmol m-' s-' for below and above 0°C samples, respectively, but this is probably mainly d u e t o variations in P i . Means of the other parameters were not significantly different. mB. the o t h e r parameter determining I I . which in the eastern Canadian Arctic was slightly larger at high latitudes (Harrison & Platt 1986). showed similar mean values for both tem- perature ranges in the Barents S e a . Although the range and median values of P: in the Barents Sea were of the same magnitude as those found in the eastern Canadian Arctic, there were large differences in the a* values between the two regions. Assuming a solar irradiance conversion factor of 4.6 pmol m-'s-' ( W m - ? ) - ' , the Canadian Arctic values showed a median of 0 . 0 1 2 m g C ( m g C h l ) - ' h - ' (pmol m - ? s - I ) - I compared t o 0.037 in the Barents Sea, a difference of 3 ~ . This obviously affected also the differences found between the median Ik val- ues from both regions, i.e. 78 p o l m-: s-' for the Canadian Arctic versus 43 for the Barents Sea. Both data sets are large and include samples taken at different seasons, depths, and hydro- graphical and biological conditions, so if the dif- ferences in mB and Ik a r e real, these comparisons suggest that the natural phytoplankton popu- lations of the Barents Sea may b e more efficient in harvesting low light. T h e possibility of the effect of the different light sources utilised in the Canadian Arctic and the present work on aB and I k cannot be ruled o u t , since both parameters are influenced by t h e light spectral Composition. The values of the P-I parameters from the Barents Sea a r e in the same range as those reported from several investigations carried out in the Antarctic Ocean (Table 2 ) . All t h e experi- ments in the Antarctic Ocean were carried o u t under controlled light conditions in laboratory incubators, with the exception of those of Sak- Photosynthesis-Irradiance relationships in the Barents Sea 109 1.73 I , I , I 0.0 A 0.15 0.10 c & i r 2 3 0.05 1.6 3.2 4.8 6.4 8.0 - 15 0.20 0.15 6 f 0.10 r 0.05 0 B 0.15 0.10 P t .k $ 3 3 0.05 C 0.044 D Fig. 4A-E. Frequency distribution of P-I parameters. Numeri- cal values represent mean values. 110 Francisco Rey Table 2. P-l parameters of phytoplankton of selected polar areas ( M e a n \ ~n pdrenthcscs) n p: b B 1, I t Rcfcrence Antarctic Isolated species 3 0 . g 9 . 3 0.0150.029 m 8 n - Rivkin & ( 5 . 5 ) (0.023) ( 6 1 ) Putt (1987) Antarctic Natural population 7 2 8-3.7 0.064-0.075 - 4 2 4 9 Rivkin Early spring ( 3 . 1 ) (0.069) (45) e t at. (1989) Antarctic Natural population 20 0.2-2.4 0 . W . M 6 s 5 3 0 34-177 Brightman Winter ( 1 . 2 ) (0.026) (219) (54) & Smith (1989) Antarctic Natural population 8 0 8-4.4 O . o O w . o J 9 16c-570 3 b l Y O Sakshaug & Summer (2.6) (0.029) (331) (103) H o l m H a n s e n (1986) Arctic Natural population 21 6 1.2' 0.012' 36X' 7X' Harrison & All seasons Platt (1986) Arctic Natural population 232 0 % 7 3 o . o o u 1 163 I s 8 9 1 & 3 1 Y This work All seasons ( 1 . 7 ) (13.044) ClX)) (47) * Median values Units: P t : mgC (mgChl)-' h - l nB: mgC (mgChl) ' h - ' ( p o l m-' s - ' ) - l . I,. Ik: pmol m-: s - ' shaug & Holm-Hansen (1986) which were carried out under natural light on deck. Relationships between the P-I parameters There is a pronounced correlation between P t and eB in both natural populations and laboratory cultures (Harding e t al. 1981, 1985; MacCaull & Platt, 1977; Harrison & Platt 1986). This cor- relation is also evident in the Barents Sea although the scattering of values is high (Fig. 5 ) , especially where aB values were higher than 0.1. These values should be viewed with caution since they lie above the theoretical limit for 9 of phy- toplankton (Sakshaug & Slagstad 1991 this volume). Values of eB lower than 0.1 showed a slightly better correlation with P z at tem- peratures above 0°C than at below 0°C. P:. as expected from theoretical consider- ations. showed a slight positive correlation with both I,,, and I,. although there was a large spread in the data. u B , also as expected, was inversely correlated with the two above-mentioned light parameters (Fig. 6 A and B ) . Neither Pw or eB were correlated with I b . P-I parameters and environmental factors Previous studies have shown that it is possible to assess the effect of environmental variables on 8 7 6 5 O I 9 I I 1 0 0.05 0.10 0.15 0.20 Q [ mgC(mgCNj* h-l (pnol m%-3-'1 Fig 5 Relationshlp bctween P E and 0' for samples taken helo% (circles) and abovc (squarea) 0°C Photosynthesis-lrradiance relationships in the Barents Sea 11 1 lo00 T ly- 600 _1 m looO 1 800 4 0 1 2 3 4 5 6 7 0 A [mgC(mgQlljl i1 1 Fig. 6A-B. Relationships between two P-I parameters. 8 l 7 1 0 0.05 0.10 0.15 0.20 B a B I mgC(mgChl)-l h-' (Ilmol ~ n - ~ s - ' ) - ' ] I 1 t 0 10 20 30 A Ophcal depth Fig. 6A-B. Relationships between two P-I parameters 1 I I I 0 10 20 30 C optical depth 1 I 1 0 10 20 30 B Optical depth m m m I I 0 10 20 30 D optical depth Fig. 7 A - D . Relationships between the optical depth and P-I parameters. 112 Francisco Rev photosynthesis through their effect o n the P-I parameters. P:. for example seems t o be. on an annual basis, predominantly controlled by tem- perature, while mB is controlled by light history (Platt & Jassby 1976; Harrison & Platt 1980; Sakshaug e t al. 1991). T h e same results have been found for phytoplankton growing in the laboratory (Mortain-Bertrand et al. 1988). T h e effect of temperature on the P-I parameters of natural phytoplankton populations of the Barents Sea have already been commented upon above. With regard to light. since irradiance data for the whole set of P-I data was not available. the optical depth (sampling depth mean light attenuation coefficient at sampling station) has been chosen as an index of the light conditions at t h e sampling depths. None of the P-I parameters (P:, wB, I,. IL or I,,) showed a clear relationship t o optical depth. However. for all parameters except Ik, there was a general trend for maximum observed Tahk 3. Simple rank correlation matrlx of photosynthetic and en\ironmental parameters. Barents Sea. 198&-1989. All coefficients based on 232 sample, PBMAX ALFA BETA PBMAX ALFA BETA IMAX IK IB OPTDEPTH TEMP SALT NO3 PO4 s104 N 0 3 / S 1 0 4 CHLOR PHAEOP PHAEOP/CHL IB OPTDEPTH TEMP SALT NO3 PO4 s104 N 0 3 / S 1 0 4 CHLOR PHAEOP PHAEOPiCHL PO4 9 0 4 N 0 3 ' S 1 0 4 CHLOR PHAEOP PHAEOP 'CHL PHAEOP CHL I OMMJ 0 4253 * 1 0 4067 ' I 0 4075 ' I 0 4948 ' I - 0 Oh22 - 0 1758 ' 2 0 2478 * I 0 0947 (1 0495 0 0394 0 1x20 ' 2 -0 0289 -0 0455 00111 -n 0528 IB I O(m I 1 0 3 5 5 -0 0427 - 0 0351 -0 0876 -0 I070 -0 0489 -0 0214 0 "9 0 0342 -0 (03005 PO4 1 O(KX) U 7846 ' I 0 6543 ' I -0 0188 - 0 uo42 0 0 4 1 4 PHAEOP CHL I oooo loo00 0 0397 -0 4264 ' I -0 5041 ' I 0 0489 -0 0616 0.0003 -0 0680 0 2075 '2 0.2052 '2 0 1809 ' 2 0 1355 '3 -0.1939 '2 -0.1183 0.1857 '1 OPTDEPTH 1 .oooo - 0.0577 0.2656 ' 1 0 1666 '3 0.1577 '3 0.0410 0 3338 ' I 0.3246 ' I 0.3434 ' 1 -0.0214 SiO4 1 .inw 0.2940 'I - 0 .O?XlJ 0 0107 0.0130 1 I M X M l -0 1389 '3 0 XXI ' I - 0 XI69 ' I -0 11113 0 1360 ' 3 0 0U)l 0 114x4 ( I 0604 0 0718 -0 0107 -0 0075 -0 0463 - 0 0295 TEMP I W K K ) 0.3715 ' I -0.3182 * I - 0 3413 ' I -I) 1321 * 3 -0.4496 * I -11.2348 * I -0 1973 ' 2 o.ni4o SO3lSiO4 I I X X X ) 0 2281 ' I 0.2131 '2 0.0075 IMAX I K I oooo 0.8178 ' I 0.3594 * 1 0 . I677 *3 0 . I083 -0.lXX9 * 2 -0.2119 '2 -0.0436 -0.1729 '2 0.1195 0.0741 - 0 . I Z O X -n.1426 '3 SALT 1 .O(MI 0.2499 ' I 0.2557 ' I 0.3240 * 1 0 1057 0.1789 '2 0.1595 '3 -0.1904 '2 CHLOR I .0uiWJ 0.8084 * 1 -0.4951 ' I 1 .0000 -0.1426 *3 -0.1277 0.2315 ' I 0.1907 * 2 -0.1639 '3 - 0 . l X l l J '2 -0.01 17 -0.1871 '2 0.1249 0.0560 -0.1746 ' 2 NO3 I .OWXI 0.9525 ' 1 0.7665 *1 0.6965 ' I -0.0184 -0.0209 -0.0018 PHAEOP 1 . WO) 0.0143 * I = 99 Y C ; confidence l e \ c l - - 99 O r < confidence h e 1 ' 7 = 95 UCr confidence h e 1 * 7 - Photosynthesis-Irradiance relationships in the Barents Sea 113 values t o decrease with increasing optical depth (Fig. 7A-D). Although the trend is strongly influenced by eight samples obtained at optical depths larger than 10, these samples were col- lected in two different years (1985 and 1987), under quite different biological conditions (see below). It can then be assumed that the observed trend is real and not influenced by some common factor. The lack of correlation between I k and depth in high-latitude samples has also been observed by other authors (Platt et al. 1982; Har- rison & Platt 1986) and has been ascribed, together with smaller depth-related differences of the other P-I parameters, to the difference in the development of density stratification in high latitudes areas as compared with temperate regions. In order to assess more qualitatively the influ- ence of different environmental factors on the P-I parameters, a Spearman rank correlation analysis was carried out on the whole data set (Table 3). P: was positively correlated with temperature and silicate, but negatively correlated with the optical depth. aB was positively correlated with nutrients and a grazing index (phaeopigments/ chlorophyll) and negatively correlated with bio- mass (chlorophyll). I, and Ik were positively correlated with hydrographic variables as tem- perature and salinity and negatively correlated with nutrients. When the data were analysed by seasons, some important differences appeared in the extent and nature of the environmental variables' effects on P:. For example during win- ter (February-March), P: was negatively cor- related with temperature but started t o show a positive correlation in spring (May-early June), exhibiting the strongest correlation during the summer months (July-August). The oppostie was seen in the correlation of P: with the optical depth, being positively correlated in winter and early spring (April) and negatively correlated dur- ing spring and summer. The relationship between aB and other environmental factors did not show such marked differences from season to season. However, the relative strength of the different correlations changed from one season to another, with nutrients being more important in the spring and the grazing index during summer. With the exception of the influence of temperature on P t that occurred through all seasons, no other environmental variable showed such strong cor- relation with the P-I parameters. The same lack of strong correlation between P-I parameters and environmental variables has been observed in other regions (Harrison & Platt 1986). A more quantitative assessment of the import- ance of the environmental variables on P-I par- ameters can be obtained b y using stepwise multiple regression analysis (Harrison & Platt 1980, 1986). For this purpose the following environmental parameters were selected: tem- perature (as an hydrographic variable), chloro- phyll (to represent phytoplankton biomass), nitrate (for nutrients), optical depth (for light conditions), the ratio between phaeopigments and chlorophyll (to represent grazing impact), and the ratio between nitrate and silicate (to represent main dominating phytoplankton groups; Skjoldal & Rey 1989). Table 4 sum- marises the results of these regression analyses using P k and aB as the dependent variables for the whole data set. Temperature and the phaeo- pigments/chlorophyll ratio were found to be the main variables controlling P:, although together they explained just 17% of the variation. The other four selected parameters explained only an additional 6%. For aB, the phaeopigment/ chlorophyll ratio explained 8.5% of the total vari- ation with nitrate as a second important variable with about 3%. The rest of the variables accounted only for 3% of the total variation. As in the rank correlation analysis there were also large seasonal variations in the contribution of the different selected parameters in explaining variations of P: or aB. During winter (n = 8). Table 4. Stepwise multiple regression analysis of P i and uB on selected biological and environmental variables. Based on 232 samples. p: Variable Temperature Phaeop/Chl Optical depth Nitrate Chlorophyll Nitrate/silicate (YB Phaeop/Chl Nitrate Optical depth Temperature Chlorophyll Nitrate/silicate Multiple R MR' 0.297 0.088 0.413 0.171 0.440 0.194 0.460 0.212 0.491 0.241 0.493 0.243 0.291 0.085 0.334 0.112 0.361 0.130 0.371 0.138 0.372 0.139 0.374 0. I40 Delta R' F-ratio 0.088 22.28 0.083 23.58 0.023 18.24 0.019 15.26 0.029 14.32 0.002 12.06 0.085 21.23 0.027 14.40 0.018 11.3s 0.008 9.07 0.001 7.27 0.001 6.09 114 Francisco Rey nutrients and biomass accounted for about 7S% of the variations in Pf)l while biomass and tem- perature were responsible for about 28% of the variations i n early spring ( n = 27). In spring (n = 133), temperature and optical depth accounted for about 149% of the variations and the ratio phaeopigments/chlorophyll and temperature were responsible for 52% of the variations during summer (n = 64). These results suggest a seasonally increasing influence of temperature on the photosynthetic capacity of the phytoplankton in the Barents Sea. This is probably related t o the increasing difference in temperature between the shallow upper mixed layer and the layers below the pycnocline that occurs throughout the spring and summer, especially in waters that have been covered by ice during winter. I n the case of aB. biomass and the ratio phaeopigments/chlorophyll contributed 61 % and 35% of the variation, respectively, during winter and early spring. In spring the variables respon- sible for most of the variation were the optical depth and nitrate, accounting for just 10%. I n summer, the optical depth, the ratio phaeopigments/chlorophyll, and temperature contributed about 30% to the variation. The importance of the light history (i.e. optical depth) in controlling aB. as it has been found in other investigations (Platt & Jassby 1976). was not evi- dent until spring and was almost negligible on a several year basis. The same lack of relationship has been found earlier (Harrison & Platt 1986). The relatively large importance of the ratio phaeopigments/chlorophyll as one of the vari- ables determining a sustantial part of the vari- ations of P t and aB, both on an annual and a seasonal basis, has not been described before for Arctic phytoplankton, probably because it has not been included in the analyses. The relative amount of phaeopigments has been suggested as an index of grazing pressure by herbivore zooplankton on phytoplankton (Shuman & Lor- enzen 1975; Welschmeyer & Lorenzen 1985), but little is known about how it can affect phy- toplankton photosynthesis. Most probably the effect is indirect, either through the recycling of nutrients or the reduction of biomass. CBte & Platt (1983). working in a small marine inlet, also found a strong positive correlation between the phaeopigment/chlorophyll ratio and P and emphasised the importance of nutrient recycling in that coastal area. Kristiansen & Lund (1989) found that in the Barents Sea ammonium was the most important nitrogen source for phytoplank- ton in late spring and summer. This period is characterised by a well-developed pycnocline that gives rise to an upper oligotrophic layer with phytoplankton growth supported by remin- Tubk 5. Mean values and standard deviation (in parentheses) of photosynthetic parameters and environmental variables from two biological situations in the Barents Sea. Depth strata Temperature Chlorophyll P! P', a n a' 1, I k * ] 0 - 2 'lo-' M a y J u n e 1987 10 metre n = 12 30 metre n = 7 >40 metre n = I ? A u g u t I985 10 metre n = 8 2G30 metre n = R 31-40 metre n = 8 > 4 l metre n = J 0.70 (-1.48-2 12)' 1.10 (-0.61-2.05)' U.98 (-0.7%?.03) 3.81 0 2 0 (-1.7.lh.48)' -0.91 (-1.78-2.83)' 0.Y4 (0 73-1.09)' (2.03-5.39)' 2.638 (2.010) 3.571 (1.138) 3.283 ( I ,362) 0.346 (0.176) 0.886 (0.492) 1.789 ( 1 . 4 5 1 ) 0.823 ( 0 . M ) 1.452 1.455 (0.751) 0.962 (0.549) (n.708) 1.735 (0.365) 1.295 (0.480) 1 283 ( I .2S3) 0.Y37 (0.171) 2.53 ( 1 . 2 4 ) 2.56 (1.31) 1.68 (0.96) 2.34 2.46 (0.87) 3.40 (3.67) 4.05 ( I .01) (0.49) 0.0244 0.0275 (0.0133) 0.0206 (0.0085) (0.0139) 0.0348 (0.0090) 0.0347 (0.01 07) 0.0561 (0.0458) 0.0526 (0.0 186) 0.43 (0.24) 0.48 (0.23) 0.36 (0.15) 0.47 (0.12) 0.66 (0.17) 1.48 (1.37) 2.31 (0.Y7) 214 (37) 1 89 (103) 1 1 1 (39) 93 ( 1 3 ) Range. Photosynthesis-lrradiance relationships in the Barents Sea 115 eralised nutrients (Rey & Loeng 1985). Under these conditions, the effects of grazing on nutrient recycling could be of major importance in con- trolling phytoplankton photosynthesis in the Bar- ents Sea after the spring bloom and the strengthening of the pycnocline. The fact that phytoplankton in this area is not severely nitrogen limited (Kristiansen & Lund 1989) supports this idea. On an overall basis neither biomass, nutrients, nor phytoplankton species composition seemed to have any major influence on controlling the variations in the P-I parameters. The same was observed by Harrison & Platt (1986) while work- ing with phytoplankton populations in the eastern Canadian Arctic. These authors found their results to be consistent with nutrient studies that showed no signs of nutrient-stress. In general, the results of this analysis of the P- I parameters in the Barents Sea confirm previous findings in other Arctic regions (Harrison & Platt 1986), i.e. the general lack of strong correlation between the P-I parameters and covarying environmental factors and the clear correlation between PE and aB. The main difference was that light did not have a clear role in determining variations in the photosynthetic efficiency, a", as it had been previously suggested. However, it is inevitable that when working with such large sets of data some single relationships become masked and do not appear as significant. This is demon- strated in Table 5 , where P-I parameters have been summarised along with some environmental factors from two different biological situations: in the spring of 1987 at the end of the phytoplankton bloom, dominated by the prymnesiophyte Phaeo- cystis pouchetii (Wassmann et al. 1990); and in the summer of 1985 after oligotrophic conditions have developed in the upper layer and a strong subsurface chlorophyll maximum has developed (Rey unpubl. data). During the spring situation, there were no differences in the P-I parameters with depth in the euphotic zone, with the excep- tion of I, that decreased with depth. Below the euphotic zone (>40m depth), only small dif- ferences in the P-I parameters were observed. This indicates that no adaptation to low irradiance had occurred, probably due to a strong vertical mixing. In the summer the situation was com- pletely different. The chlorophyll normalised P- I parameters were similar throughout the upper 30 metres, but they showed a sharp increase in aB at the region of the subsurface chlorophyll maximum and below, while P z , I, and Ik decreased with depth. This supports the findings of Beardall & Morris (1976) that a diatom growing at low intensities has enhanced ability to utilise sub-optimal intensities and reduced ability to util- ise saturating levels. The most dramatic differ- ence, however, was found in the carbon normalised P-I parameters. Both P: and ac increased strongly with depth below the euphotic zone, suggesting that one of the mechanisms by which the phytoplankton adapted to the low light levels was increasing their cellular concentrations of photosynthetic pigments. These findings illus- trate small scale effects on the P-I parameters, such as light adaptation, which may not always be revealed in a global analysis as the one made here. Acknowledgements. -Sincere thanks to F. Mora for his help in collecting and processing the data. My gratitude also to B. Stensholt for advice in statistical matters and to E. Sakshaug for providing valuable comments. Thanks are also due to two anonymous reviewers for their helpful suggestions. This work was partially supported by the Norwegian Fisheries Research Council. References Beardall, J. & Morns. I . 1976: The concept of light intensity adaptation in marine phytoplankton: some experiments with Phaeodacrylum tricornutum. Mar. Biol. 37, 317-387. Brightman, R. I. & Smith. W . 0.. Jr. 1989: Photosynthesis- irradiance relationships of antarctic phytoplankton during austral winter. Mar. Ecol. Prog. Ser. 5 3 , 143-151. CBt6, B. & Platt. T. 1983: Day-to-day variations in the spring- summer photosynthetic parameters of coastal marine phy- toplankton. Limnol. Oceanogr. 28. 32Ck344. Feyn. L.. Magnussen. M. & Seglem. K. 1981: Automatic analy- sis of nutrients with 'on-line' data treatment. A presentation of the construction and performing of the system in use at the Institute of Marine Research vessels and laboratory. Fisken Hau., Serie 8. 1981(4), 1-39. Harding, L. W . . J r . . Meeson. B. W.. Prezelin. B. B. & Sweeney. B. M. 1981: Die1 periodicity of photosynthesis in marine phytoplankton. Mar. Biol. 6 1 , 95-105. Harding. L. W . , Jr., Meeson. B. W. & Fisher, T. R.. J r . 1985: Photosynthesis patterns in Chesapeake Bay phytoplankton: short- and long-term responses of P-I curve parameters to light. Mar. Ecol. Prog. Ser. 26. 99-1 11, Harrison, W. G. & Platt. T. 1980: Variations in assimilation number of coastal marine phytoplankton: Effects of environ- mental co-variates. 1. PIankton Res. 2, 249-260. Harrison, W. G. & Platt. T. 1986: Photosynthesis-Irradiance relationships in polar and temperate phytoplankton popu- lations. Polar Biol. 5 . 153-164. Kristiansen, S. & Lund, B.Aa. 1989: Nitrogen cycling in the Barents Sea-I. Uptake of nitrogen in the water column: Deep- Sea Res. 36, 255-268. 116 Franci.sco Rey Loeng. H. 1989: Ecological featurcs of the Barents Sea. Pp. 3 M 1 in Rey. L. & Alexander. V. (eds.): Proceedings of rhe Surh Conference ofrhe Cofnlre Arcrique l n f e r ~ i a r i o n a l , 13- I5 May 1985. E . J. Brill. Leiden. MacCaull. W. A. & Platt. T. 1977: Die1 variations i n the photosynthetic parameters of coastal marine phytoplankton. L i m n o l . Oceanogr. 22. 73-73 1 Mortain-Bcrtrand. A . . Dcscolas-Gros. C. & Jupin. H. 1988: Growth. photosynthesis and carbon metabolism in the tem- perate marine diatom S k d e r o n e m a cosfafurn adapted to loa temperature and l o w photon-flux density. M a r . Biol. 100. 135-1.11. Neldcr. J . A. & Mead. R . 1965: A simplex method for function minimization. Cornpurer 1. 7. 30b31.7. Platt. T . & Jasshy. A . 1976: The relationship between photo- synthesis and lrght for natural assemblages of coastal marine phytoplankton. J . P h w o l 12. 421-430. Platt. T.. Gallegos. C. L & Harrison. W . G . 1080: Photo- inhibition of photosynthesis i n natural assemblages of marine phvtoplankton J . Mur. Res. 38. W7-701. Platt. T . Harrison. W . G . . Irwin. B.. Horne. E. P. & Gallegos. C. L. 1982: Photosynthe5is and photoadaptation of marine phytoplankton in the Arctic. Deep-Sea Res 29. 115Y-1170. Rey. F & Loeng. H . 19x5: The influence of ice and hydro- graphic conditions on the development o f phytoplankton in the BarentsSea. Pp. 4%63 in Gray. J. S. & Christianscn. M . E. (eds.): Marine Biology of Polar Regions a n d €ffecfs of Srress o n Marine O r g a n u m s . John Wile? & Sons Ltd.. England. Rivkin. R . B. & Putt. M. 1987: Photosynthesis and cell division by antarctic microalgae: comparison of benthic. planktonic and ice algae. J. P h y d 23. 223-229. Rivkin. R. B.. Putt. M . . Alexander. S. P.. Meritt. D. & Gaudet. L. 1989 Biomass and production in polar planktonic and sea ice microbial communities: a comparative study. M a r . Biol. 101. ?7-L283 Sakshaug. E . & Holm-Hansen. 0. 1986: Photoadaptation in Antarctic phytoplankton: variations in growth rate. chemical composition and P versus I curves. 1. PIankron Res. 8 . 459- 473 Sakshaug. E . & Slagstad. D. 1991: Light and productivity of phytoplankton in polar marine ecosystems: a physiological view. Pp. 6%85 i n Sakshaug. E.. Hopkins. C . C. E. & Britsland. N . A . (eds.): Proceedings of the Pro Mare Sym- posium on Polar Marine Ecology. Trondheim. 12-16 May 1990. Polar Research 1 0 ( / ) . Sakshaug. E.. Johnsen. G . . Andresen, K . & Vernet. M. 1991: Modeling of light-dependent algal photosynthesis and growth: experiments with the Barents Sea diatoms Thal- assiosira nordenskioeldii and Chaeroceros fitrcellarus. D e e p - Sea Res. 38. 41-30, Skjoldal. H. R. & Rey. F. 1989: Pelagic production and varia- bility of the Barents Sea ecosystem. Pp. 241-286 in Sherman, K . & Alexander. L. M . (eds.): B i o m a s s y i e l d s a n d g e o g r a p h y o f large marine ecosvsrems. AAAS Selected Symposium 11 1, Westview Press. Inc.. Colorado. USA. Shuman. F. R. & Lorenzen. C . J . 1975: Quantitativedegradaton of chlorophyll by a marine herbivore. Limno1. Oceanogr. 20, 5Ru-585 Wassmann. P . . Verncl, M . . Mitchell. B. G. & Rey. F: 1990. Mass sedimentation of Phaeocysris p o u c h e f i i in the Barents Sea. Mar. Ecol. P r o g . Ser. 66. 183-195. Welschmeyer. N . A . & Lorenzen. C. J . 1985: Chlorophyll budgets: zooplankton grazing and phytoplankton growth in a temperate fjord and the Central Pacific Gyre. L i m n o l . Oceanogr. 30. 1-21