Light and productivity of phytoplankton in polar marine ecosystems: a physiological view EGlL SAKSHAUG and DAG SLAGSTAD Sakshaug. E. & Slagstad. D . 1991: Light and productivity of phytoplankton in polar marine ecosystems: a physiological view. Pp. 6S-85 i n 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. folur Research 10(1). This study deals with the modelling of photosynthesis and growth of polar phytoplankton and variations in relevant parameters. Polar rcgions are chardctcriscd by low sun elevations (< 4k50"). extreme seasonal variations in irradiance and day length. and low sea temperatures ( - I . R to 6°C). Due to the latter. maximum phytoplankton growth rates are low (< 0 . 6 d - ' ) . Light absorption by phytoplankton is strongly dependent o n spectral composition (blue oceanic versus green coastal waters), but absorption characteristics (and thus chlorophyll a-normalised photosynthetic efficiency $) d o not differ appreciably between polar and othcr phytoplankton. The maximum chlorophyll- normalised photosynthetic rate P t is, however. lower and dependent on the irradiance to which the cells are adapted. Chla:C ratios vary widely. but within ranges known for other phytoplankton. The carbon- normalised coefficient P', varies little with irradiance. but is clearly dependent on day length and nutrient supply. The corresponding coefficient a' is somcwhat higher in shade-adaptcd than in light-adapted cells. Polar species exhibit a high tolerance for strong light and long days in combination with low temperature relative to other species. The interpretation of P-l functions is discussed, and an empirical formulation is suggested that does not need the Chla:C ratio for predicting the light-limited gross growth rate of polar phytoplankton. Math- ematical simulations of the spring bloom indicate that the depth of the mixed layer and the attenuation of light are the most important variables for determining the photosynthetic rate. The spectral composition of light is of particular importance in low light, e.g. in deeply mixed layers. Generally, the deeper the mixing, the more sensitive thc development of a spring bloom becomes to any algal or environmental variable. Egil Sakshaug, Trondhjem Biological Station. The Museum, Unioersily of Trondheim, Bynesoeieri 46. N - 7018 Trondheim. Norway; Dag Slagstad. SINTEF. Auromaric Conrrol. N-7034 Trondheim-NTH. Norway. 'J*P,~,4~\Ns" Introduction Polar marine environments are characterised by extreme variations in irradiance and day length, and temperature is generally low (- 1.8 to 6°C). The Arctic seas, including the Bering, Greenland and Norwegian Seas, reach as far south as 60"N, and the Antarctic Ocean, defined as extending north to the oceanic Polar Front, reaches latitudes as low as 45-50"s. Photosynthetic and growth rates of phyto- plankton in polar seas may vary significantly because of variations in the light regime, but they are generally low d u e t o low temperature. Phytoplankton is continuously exposed t o a light gradient because of variations in atmospheric light and mixing in the water column. Because light is attenuated in water, the depth of the mixed layer and the rate of mixing are essential in controlling the radiation available t o algae: the deeper the mixing, the less light on the average is available to the phytoplankton community. In polar areas, blooms may be triggered by the melting of sea ice which leaves phytoplankton suddenly exposed t o strong light; a brackish layer which is formed a t the same time restricts the depth of mixing. T h e melting of sea ice thus gives rise t o a bloom which trails the receding ice edge and thus occurs later at higher latitudes. This scenario was suggested by Gran (1931). Although such blooms have been observed in all polar seas covered by sea ice [see Sakshaug (1989) for review], they are not easily identified when the ice edge is poorly defined and when ice drifts across nutrient-depleted waters late in the growth season. The relationship between vertical mixing and the growth of phytoplankton was first proposed by Alexander Nathansohn in 1906 (quoted in Braarud 1935). The relationship was demon- strated by Braarud & Klem (1931) through com- parison of the early blooms in the shallowly mixed Romsdalsfjord in Western Norway with the late blooms in the Norwegian Sea (May-June), and by Riley (1942; 1946) at St. Georges Bank. O n 70 Egil Sakshaug & Dug Slagstad the basis of such observations, Sverdrup (1953) developed a mathematical model in which he introduced the concept of critical depth (zcr). which is defined as the depth above which depth-integrated photosynthesis equals depth integrated respiration: a bloom develops when vertical mixing is less deep then the critical depth (i.e. the integrated photosynthesis exceeds the integrated respiration). The critical depth may approximately be defined as z,, = Ps/kr (1) where Ps is biomass-specific photosynthetic rate at the surface, k is the vertical attenuation coef- ficient of light in seawater and r is the biomass- specific respiration rate of algae. It follows that z,, becomes large when Ps is high or k and r are low. Sverdrup calibrated his model against data from the weather ship M in the Norwegian Sea and derived a critical depth that ranged from 30- 40 m in March to 230-280 m in May. Sverdrup's model is the backbone of many mathematical models for primary production and may be fairly realistic for predicting the onset of spring blooms. Improvements can be made by including a non-linear relationship between pho- tosynthetic rate and irradiance. the effect of the variable spectral composition of light, and self- shading by phytoplankton. Moreover, although Sverdrup was aware of the importance of losses through, for example, grazing and sedimentation, the model included only the algal dark respiration rate. Unless terms for grazing and sedimentation are included, predicted critical depths may be hundreds of metres i n clear waters - far deeper than depths of mixing above which phytoplankton blooms actually occur (4&80m, Sakshaug & Holm-Hansen 1984; Bodungen et al. 1986; Sme- tacek & Passow 1990). Finally, algae can adapt to changes in the environment, usually in a fashion that minimises the impact of environmental vari- ation on photosynthetic and growth rates (Sak- shaug & Holm-Hansen 1986). Adaptation changes the values of algal parameters, some of which are essential i n dynamical phytoplankton models, e.g. P-I coefficients and the Chlorophyll a : Carbon ratio. The present paper deals with light absorption by polar phytoplankton, mathematical models of photosynthetic and growth rates as well as photo- adaptation of polar phytoplankton. The model is examined by analysing recent data from the Barents Sea and comparing observations of spring blooms with predictions. Polar primary pro- duction has been reviewed elsewhere (Platt & Subba Rao 1975; Horner 1976; Alexander 1980; Nemoto & Harrison 1981; Subba Rao & Platt 1984) as has the physiological ecology of polar phytoplankton (Jacques 1983; Sakshaug & Holm- Hansen 1984; Sakshaug 1989; Smith & Sakshaug 1990). The photosynthetically relevant aspects of marine optics have been reviewed recently (see Morel 1988; Sathyendranath & Platt 1989 and references therein; see also Jerlov & Steemann Nielsen 1974; Jerlov 1976; Tyler 1977; Kirk 1985 for general introduction). Atmospheric light in polar areas Sun elevations are generally low in polar areas. The highest elevation (noon, summer solstice) at 80" latitude is 33.5" and at 60" latitude 53.5". On the other hand, at latitudes >66.5", the sun does not set for part of the summer. Because of this, daily incident irradiance on a cloudless day in midsummer may in fact be higher at high latitudes than in equatorial areas (Campbell & Aarup 1989). The diurnal trajectory of the sun is given by I 1 I I A B C Q 4 8 12 16 20 24h Fig. I . Diurnal solar trajectory as a function of latitude. The position of the horizon depends on season: A = at summer solstice. B = at equinox, and C = at winter solstice. A and C are 47" apart. Light and productivity of phytoplankton 71 latitude only and becomes “flatter” the higher the latitude (see Kirk (1985) for astronomical equations), whereas the horizon may be thought of as moving up or down relative to this trajectory according to the seasons (Fig. 1). The math- ematical model by Bird (1984), which predicts spectral scalar irradiance for cloudless days at the sea surface as a function of sun elevation, implies that relative spectral distribution of atmospheric visible light varies little between solar elevations >lo” (Fig. 2). Predictions from this model also correspond well with data from the Barents Sea in that the highest measured scalar irradiances (PAR, i.e. Photosynthetically Active Radiation, integrated over 400-700nm) at sea surface approximate but d o not exceed model predictions (Fig. 3). Generally, cloud conditions and the wind rep- resent major problems in the modelling of ocean and atmosphere dynamics and thus plankton dynamics. The subarctic region and the mid-to- northern part of the Antarctic Ocean are, however, in the atmospheric “low pressure belt” in which atmospheric low pressures move incess- antly eastwards. This usually yields a weather cycle characterised by 4-6 cloudy and windy days followed by 1-3 more or less sunny days and less windy days. This induces rhythm in incident light and vertical mixing and is thus of relevance for primary production (Sakshaug et al. 1991b). Because atmospheric low pressures tend to follow the same trajectory for several weeks, cycles of cloudiness and wind may to some extent be repro- ducible properties of this belt. At high latitudes, in the polar high pressure zone, sunny days may be more frequent but, particularly in the high Arctic, where weather may be calm for extended periods, a thin layer of fog may frequently be formed. *a 0 I I I 0 4 8 12 16 2‘0 2 I h Fig. 3. Scalar irradiance (PAR) at the sea surface on 100% cloudy days (solid circles) and partly cloudy days (open circles) in the period 25 May-6 June, 1987, at 74”N in the Barents Sea. Data provided by G . B . Mitchell. The curve represents predictions for a cloudless day (Bird 1984). Cloud cover and fog may reduce incident light by up to 70-80% (Kirk 1985). In the southern half of the Barents Sea, the average scalar irradiance (PAR) on fully cloudy summer days is on average only 18 ? 6% of the irradiance predicted for a clear day, whereas sunny/partly cloudy days aver- age 55 * 20% (Fig. 2). According to the same set of data, 28 of 46 summer days were fully cloudy, and scalar irradiance (PAR) at 2 m depth was nm Fig. 2. Spectral distribution of incident light (scalar irradiance, relative scale) at the sea surface on a clear day as a function of sun elevation (after Bird 1984). 72 Egil Sakshaug & Dug Slagstad 41 2 5% of scalar irradiance ( P A R ) at the surface. O J 50- Attenuation of light by clear water and sea ice In clear oceanic waters. blue rays penetrate the deepest. For "clearest" ocean water (Smith & Baker 1981). the minimum vertical attenuation coefficient is 0.0168 m - ' at 450 nm wavelength. while U V ( 2 0 0 n m ) and infrared radiation (760nm) have coefficients as high as 3.14 and 2.55 m-I. respectively. Next t o blue, ocean water is most transparent t o violet and green (Fig. 4). A B ,.. E A 2 B 5 10 50 100 300 o . 0 1 4 , ~ ~ . , I , , . , J 2 0 0 300 400 5 0 0 6 0 0 7 0 0 nm hg. 4 . Vcrtical diffusc attenuation cocfficicnt of light and 1 5 light depth in ( A ) "clcarcst" ( b l u c ) ocean watcr (Smith & Baker 1981) and (B) "clcarcst" grccn watcr in Tmndhcimsfjordcn. Norway. early March ( G c i r Johnscn. unpuhl d a t a ) . Fig. i. Vcrticdl diffusc attenuation coefficient ( P A R ) in ( A ) "clcarcat" ocean u a t c r and (6) "clcarcs~" Trondhcimsflordcn u a t c r Samc data a s in Fig. 4. E c a 0 Whereas deep-sea polar waters in winter. par- ticularly in the Antarctic O c e a n , may be fairly similar t o this "clearest" water (Gieskes e t al. 1987). waters o n shelves and in bays and fjords in northern areas may have minimum attenuation coefficients in the green part of the spectrum (500-600 nm). Norwegian fjord and Coastal Cur- rent waters a r e , for instance. characterised by input of dissolved yellow matter, mainly of Baltic origin, and therefore strongly absorb blue and violet radiation (Fig. 4). These waters generally have a much higher attenuation coefficient in terms of P A R than the "clearest" blue ocean water (Fig. 4). Because phytoplankton a r e highly discriminatory with respect t o absorption of blue and green light, shifting the spectral distribution of light in water towards green as a bloom devel- ops, the spectral characteristics of seawater are ecologically important (Fig. 5 ) . Light is greatly attenuated by ice a n d , in par- ticular. snow. Palmisano e t al. (1986) have reported diffuse vertical attenuation coefficients ( P A R ) of 16-45 m - ' for snow, i.e. a 5 0 c m thick layer will only transmit 0.01-3% of incident light. Sea ice has coefficients of 1.5-1.611-l ( P A R ) (Maykut & Grenfell 1975). i.e. a 1 m thick layer of sea ice will let through about 20% of incident light. Consequently, ice and its snow cover reduce incident light so much that there is hardly enough for more than algal subsistence below the under- side of sea ice. Absorption of light by particles The absorption of light by phytoplankton depends on species size/morphology as well as pigment Light and productivity of phytoplankton 73 between shade- and light-adapted cells may also be caused by changes in pigment composition (Fig. 4). The "package" effect is pronounced in a Barents Sea clone of the large-celled centric dia- tom Thalassiosira nordenskioeldii (Fig. 6); e.g. "ac for red light (676 nm) is 0.016 m2 (mg Chla)-] for cells growing at 400 pmol m-z s - ] and 0.0073 for cells growing at 25 pmol m-z s-I. Natural populations in the Barents Sea exhibit the same trend, e.g. "a,(676) averages 0.011 in the well-lit upper 10 metres and 0.0057 mz (mg Chla)-' below 60 m depth (Table 1). 'ac may overestimate absorption of photo- synthetically usable light by algae because some of the absorbed light is unavailable for photo- synthesis. This pertains, for instance, to light absorbed by photoprotective pigments such as diadino- and diatoxanthin which do not transfer energy to Photosystem I1 (Vernet et al. 1989). Chla-normalised absorption of photosynthetically usable light (PUR) may rather be approximated by properly scaled fluorescence excitation spectra (OF), which mimic well the shape of action spectra for oxygen evolution in Photosystem I1 (Neori et al. 1988). Scaling of relative spectra has until recently been a problem, but can now be easily carried out by use of the dye Basic Blue 3 for quantum correction and matching of the peak at 676nm of the corrected spectrum to the cor- composition, but it is presumably relatively inde- pendent of temperature and thus not in principle different for polar and other species (Mitchell & Kiefer 1988). The predominant algal species in polar regions have in vivo light absorption spectra which pri- marily reveal the absorption characteristics of Chlorophyll a (Fig. 6): one major absorption peak at 440nm, i.e. not far from the wavelength of minimum attenuation in blue water, and one at 675 nm (red), which is of less ecological relevance since seawater attenuates red light efficiently in the upper few metres. Spectra of the more impor- tant groupsof polar algae. e.g. diatomsand Phaeo- cystis pouchetii, exhibit in addition a shoulder caused by fucoxanthin and closely related com- pounds at about 500 nm, and minor peaks due to Chlorophyll c at 590 and 630 nm (Fig. 6). Chlorophyll a-specific absorption of light ("a,) by phytoplankton cultures ranges from 0.0043- 0.034 at 676nm and 0.004-0.09m2 (mg Chla)-' at 440 nm (Maske & Haardt 1987). Self-shading in and between chloroplasts may lead to low Chla- normalised absorption by Chla-rich (shade- adapted) cells. This is particularly pronounced in large cells with many chloroplasts (the "package" effect; Kirk 1975; Bricaud et al. 1983; Mitchell & Kiefer 1988; Berner et al. 1989). Large differ- ences, however, in the absorption of blue light Fig. 6. In vivo Chlo-specific light absorption spectra for the large centric diatom Tho/afsiosira nordenskioeldii (Earents Sea clone) grown at 0.5"C and at 400 (Tn-HL) and 25 pmol m - 2 s - ' (Tn-LL). "ac: total absorption. O F : chlorophyll o-fluorescence excitation spectrum (emission 730 nm) scaled to "ac at 676 nm (Sakshaug et al. 1991a). E" - N E 0.04 0.03 0.02 0.0 1 500 600 500 600 nm 74 Egil Sakshaug & Dug Slagstad responding peak of 'ac (Kopf & Heinze 1984; Maske & Haardt 1987; Sakshaug et al. 1991a). For Thalassiosira nordenskioeldii, "F is 42 and 86% of 'ac of light- and shade-adapted cells, respectively, at 440 nm. Hence, OF varies less than 'ac with photoadaptational status in this species, although it is nevertheless lower for shade- adapted than for light-adapted cells (Fig. 6). The large difference in terms of OaC between light- and shade-adapted cells in the blue region is pre- sumably i n part due to the much higher content Om I ; of diadino- and diatoxanthin in the former (Sak- shaug et al. 1991a) and therefore does not rep- resent a simple "package" effect. The absorption of light by all particles in the 500 700 500 700 nm seawater (*a,) can be much larger than 'ac. par- ticularly when the supply of terrigenous material is large (e.g. feeding by glacial rivers), and fre- quently considerably larger also in offshore areas (Morrow et al. 1989). In the Barents Sea. the difference between *ac and 'ac is small in the early phase of ice-edge blooms, but becomes larger i n the late phase, particularly near the surface (Fig. 7). The difference between *a, and 'ac in waters with little terrigenous material appears to be caused primarily by pigments which are creamish yellow and insoluble in methanol. Kiefer et al. (1990) have suggested that these pigments are in the main cytochromes, i.e. respiratory enzymes Fig. 7. Examples of Chlo-specific absorption of light by particles in the Barents Sea (0 and 10 m depth, 2 June 1987). 'a, = total absorption by particles. = in vivo absorption by chloroplast (methanol-soluble) pigments and their derivatives ( E . Sakshaug. unpubl. data). present in procaryotes as well as eucaryotes. This raises the possibility that the difference between *ac and 'ac may be used to estimate respiratory activity in seawater. When phytoplankton stocks are low (e.g. oligotrophic situations), absorption of light by heterotrophs and other non-algal par- ticles may be larger than absorption by chloroplast pigments (Smith et al. 1989). Table 1 . Averaged photosynthetic parameters (24 and 12 h day length) for Thalassiosira nordenskioeldii and Chaetoceros furcellarur grown at 0.5"C and natural populations in the Barents Sea (May-June 1987. 5-25 data sets for each depth range). "F is the integrated value ( W 7 0 0 n m ) of "F taking the spectral distribution of the illumination into account (calculated in analogy with equation 6 ) . Culture data from Sakshaug et al. (1991a); field data provided by F . R e y . E . Sakshaug. M. Vernet and B . G. Mitchell. For explanation of symbols, see text. Cultures ~ Barents Sea pmol m - j s - ' rn 0 025 1 7 0 024 0 011 0 058 0 045 68 405 Depth. m 25 0-20 3 M >70 0.023 0.026 0.025 0.020 0.8 1.6 1.3 0.9 0.014 0 011 0.0094 0 0057 0.0078 0.070 0.035 35 63 52 46 1080 ' rng C (mg Chla).' h - ' ( y m o l m - 2 s - ' ) - ' ' umol m - ? s-I ' rn' (mg Chlo).' rng Chla (mol P S U ) - ' g-at C (mol).' mg C (rng Chla).' h - ' ' growth-normalised values Light and productivity of phytoplankton 75 In equation 4, the system of units conveniently differs from that of equation 3: PB is in g-at C (mg Chla)-' h-l and E, in mol m-2 h - I . Ouis the Chla- normalised specific absorption of photosynthet- ically usable light [m2 (mg Chla)-']; the scaled fluorescence excitation spectrum O F may be used as an adequate approximation (Sakshaug et al. 1991a). t is the minimum turnover time for elec- trons in the photosystems (h), and q is the amount of Chla per photosynthetic unit [mg Chla (mol PSU)-'1. The PSU is here functionally defined, e.g. the existence of different photosystems is disregarded - q t probably mainly represents properties of PSII. $,,,, the maximum quantum yield, converts photons to carbon [g-at C (mol photons)-'] or oxygen [mol O2 (mol photons)-']. Equation 4 is the product of a linear function for photosynthesis versus irradiance multiplied by the Poisson Probability (PP) for a photon to hit a photosynthetic unit with an open reaction centre (i.e. the quantum efficiency). The latter term has the value 1 for zero light and the value zero for infinite irradiance: P B = ($max ODE, ) (linear) x [l - e ~ p ( - ~ u q t E , ) ] / ~ u q t E , . (5) (PP) When multiplied by $,,,, the Poisson Probability term predicts the quantum yield. Equations 3 and 4 are mathematically equiv- alent, thus 9 = 43.2 $,,,"a. P: = 12000 P:/qt, and Ik = 278/"uqt. It should be noted that P: and P-I functions P-I functions describe the photosynthetic response of phytoplankton to irradiance. The sim- plest formulation is a straight line: P = LYE, (2) where P is biomass-normalised carbon uptake (or oxygen evolution), E, is irradiance, and LY is the slope of the line. This linear relationship was used in Sverdrup's (1953) model. Because, however, the quantum yield of photosynthesis decreases with increasing irradiance, the photosynthetic rate in reality forms a saturation function with irradiance. Equation 2 is therefore adequate only at very low irradiances. An empirical and simple saturation function (Fig. 8) was suggested by Webb et al. (1974): PB = P i [ l - exp(-mBE,/P:)] (3) where PB is chlorophyll-normalised carbon uptake [mg C (mg Chla)-l h-'1, P i is maximum carbon uptake, E, is in pmol m-2 S K I , and 9 is the slope of the curve at the origin [mg C (mg Chla)-' h-' E i l ] . The ratio P:/@, the light satu- ration index, is termed Ik. An alternative formulation for equation 3 based on target theory (Arnold 1932; Myers & Graham 1971; Ley & Mauzerall 1982; Dubinsky et al. 1986) has been suggested by Sakshaug et al. (1991a). This formulation makes the physio- logically relevant coefficients explicit (Fig. 8): pB = ($max/qm - e x P ( - " ~ q ~ E , ) l . (4) Fig. 8. P-I curve according to Webb et al. (1974) and target theory notation (Sakshaug et al. 1991a). Because of "u, values for aB and Ik depend on the spectral distribution of the light. 76 Egil Sakskaicg & D a g Slagstad E 1 aS include the conversion factor whereas Ik does not and that P z is spectrally independent in accordance with experimental data (Rochet et al. 1986). Because C@ and I k include the spectrum ' a , they depend on the spectral distribution of ambient light. If Ou (e.g. OF) and c@ for the P-I incubator light source is known, the integrated value aB (400-700 nm) for a particular submarine light regime can be calculated if the spectral dis- tribution of that light is known (Morel 1978; Lewis et al. 1985: SooHoo e t al. 1987): mB = 43.2 $,,, " U = 43.2 - - 7110nm [i,,,,, " d n ) E , . d ; l l / E , , ( P A R ) l . (6) EJA) is the spectral irradiance of the particular light regime and E J P A R ) the integrated (40C 700nm) irradiance for the P-I incubator. Equation 6 implies that mB derived on the basis of a "white" i n g b a t o r light source may differ from the value aB relevant in the field (Fig. 8). A s s u m b an infinitesimally small phytoplankton stock. a" may increase with depth by a factor > 2 in blue oceanic waters because the water becomes "bluer" (Fig. 9 ) . In green waters the converse takes place. When phytoplankton bloom. blue radiation is more rapidly absorbed by algae. so that blue waterwith 10-15 mg Chla m--'resembles green water optically. This has the additional implication that. unless special correction algor- 20 Fig. 9. Variations in the integrated (4W700 nm) value d as a function of changes in spectral distribution of light wlth depth (infinitesimal chlorophyll concentration assumed for A and B ) : A = "clearest" ocean water (Fig. 4): B = "clearest" Trondhcimsfjordcn water (Fig. 4 ) : C = "white" halogen lamp (Osram H 0 1 250); D = "clearest" blue water with 10 mg Chla m-': E = "clearest" blue water with 15 mg Chlo m-'. tthms are employed, satellite images of chloro- phyll in the sea may considerably overestimate phytoplankton stocks in humus-laden waters, such as along the Norwegian coast. Variations in P-I coefficients P-I coefficients a r e adaptive parameters. i.e. their values may change as a response to the environ- ment in which the algae have been growing. T h e rate of adaptation depends o n the coefficient in question, but data are sparse and d o not form a general pattern (Gallegos e t al. 1983; Post et al. 1984: Sakshaug et al. 1987; Cullen & Lewis 1988; Hegseth 1989). Changes in the Chla:C ratio may be completed in a matter of a few hours to some days. P-I coefficients may change in the course of <30 minutes (Lewis & Smith 1983; Legendre et al. 1988); therefore. functions o t h e r than 3 and 4 may correspond better with t h e data if incubation time is longer than this (Jassby & Platt 1976; Platt et al. 1980; Leverenz 1988). Day length and nutrient supply apparently affect Chla-normalised P-I coefficients little compared t o irradiance, although this may b e species-dependent (Sak- shaug et al. 1989: 1991a). aB. P i and Ik of polar phytoplankton exhibit a bewildering variability (see Harrison & Platt 1980: 1986; Smith & Sakshaug 1990 and ref- erences therein). Reported values for $ in white light range from 0.001-1 mg C (mg C h l a ) - ' h - ' Light and productivity of phytoplanktorl 77 sumably related to a low value for "F i n shade- adapted algae. Moreover, data sets from northern areas exhibit a clear positive covariation between $ and P i (Harrison & Platt 1980; 1986). A physiological reason for this covariation may be the concomitant lowering and raising of OF and the product qt, respectively, when light-adapted cells become shade-adapted (Table 1). Variability in @,,, will also cause P i and a" to covary. and "F put constraints on the variation i n 9. For white light, OF is hardly higher than 0.004- 0.015 m2 (Chla)-I (Sakshaug et al. 1991a). This value and the theoretical maximum of 0.125 for $,,, imply an upper limit of 0.08mg C (mg Chla)-' h-' (pmol m-z s-')-' for aB. For total carbon uptake, L@ is likely to be lower than 0.05, whereas $ for filtered samples (the majority of published data) may be as low as 0.0254.040. Assuming a minimum value for "F of 0.004 m2 (mg Chla)-l. the lower limit for aB (filtered samples) becomes 0.007. The large majority of reported values for $ of polar phytoplankton are actually between 0.007 and 0.040. The Chla:C ratio varies inversely with irradiance and day length and proportionally with nutrient supply (Sakshaug & Andresen 1986). Therefore the patterns of variation for the carbon- normalised P-I coefficients P: and & (i.e. P: and aB, respectively, multiplied by the Chla:C ratio) differ from those of P i and aB. In fact, P:, the maximum hourly carbon turnover rate, appears to be relatively independent of irradiance (Sakshaug & Holm-Hansen 1986) but strongly dependent on day length (Table 2) and nutrient status (Sakshaug et al. 1989). & appears to increase somewhat when cells become adapted to low irradiance and short days, largely because "F decreases less than the Chla:C ratio increases (Sakshaug & Holm-Hansen 1986; Sakshaug 1989). Thus shade-adapted cells are photo- synthetically more efficient than light-adapted cells, in spite of the opposite impression given by variations in $. (pmol m-l s-l)-] and for P i from 0.1-15 mg C (mg Chla)-'. The majority of values for L@, how- ever, range from 0.010-0.030 and for P i from 0.3-2.0. Also Ik exhibits a pronounced variability: 0.5-700 pmol m-2 s-I has been reported, although the majority of values lie between 30 and 200 pmol m-2 s-I. Some values, particularly the extreme ones, may be erroneous due to prob- lems in measuring carbon uptake and Chla and extensive manipulation of the samples before incubation (including shocks caused by strong light, and in the special case of ice algae: caused by melting). Data sets may be difficult to compare because of differences i n methodology: incubation time and time of day may be significant (Smith & Lewis 1983; Legendre et al. 1988), the spectral composition of incubator lamps may differ, and whereas some data sets are based on filtration of samples after incubation, others are not. Finally, species composition of communities may play a role. Adaptive variation in natural communities may therefore be difficult to distinguish from vari- ations caused by other sources. @,,,, which is included i n both P! and aB, cannot be measured directly and is therefore a residual in regressions. It thus "absorbs" sys- tematical errors, and, more important, it depends on how much of the flow of fixed energy through the cells is intercepted by the method for meas- uring photosynthesis. Measurement of oxygen release - which takes place in Photosystem I1 and therefore before any fixed energy is spent by dark reactions - may yield values close to the theoretical maximum [0.125 mol O 2 (mol photons)-']; carbon uptake does not, because a fraction of the supplied energy is spent on uptake of nutrients instead of the uptake of carbon. Typi- cal values for carbon uptake (unfiltered samples) are 0.07-0.08 g-at C (mol photons)-' (Langdon 1988). Moreover, estimates of carbon uptake based on filtered samples ( e . g . growth-relevant estimates) yield values as low as 0.04-0.06 g-at C (mol photons)-', i n part because they do not include production of extracellular carbon (Zie- mann et al. 1987; Sakshaug et al. 1991a). Some low values for @,,,ax have been reported because they have been calculated as $pac instead of L@/"F. For light-adapted cells the difference can be considerable (e.g. = 50%). $ of Arctic phytoplankton may be equal or somewhat lower for shade-adapted than for light- adapted algae (Platt et al. 1982); this is pre- Light-limited growth rates Reported growth rates of polar phytoplankton are up to 1.7d-' [see Smith & Sakshaug (1990) for review]; long-term studies (several days) of shipboard cultures, however, indicate maximum rates of 0.35-0.90d-' at <5"C (Sakshaug & Holm-Hansen 1986; Wilson et al. 1986; Spies 78 Egil Sakshaug & Dag Slagstad 1987). The maximum growth rate among cultures of 10 diatoms from the Barents Sea grown at -0.5"C was 0.64 d - ' ? and the average maximum for the ten species was 0.52 d - ' (Gilstad & Sak- shaug 1990). The function by Eppley (1972), which predicts the maximum ternperature-depen- dent growth rate among numerous temperate species grown in laboratory cultures at >2"C, predicts by extrapolation a maximum growth rate of 0.55d-' at 0°C. It thus seems that maximum growth rates of polar phytoplankton on the aver- age are not higher than those that would be exhibited by temperate species at low tempera- tures. Experiments with shipboard cultures of natural communities (Sakshaug & Holm-Hansen 1986: Sakshaug 1989) indicate an optimum range of 35- 105 pmol m-2 s - I for Antarctic ice-edge com- munities, yielding a growth rate of 0.25 d - ' , and 100-250 pmol m-? s - ] for a Barents Sea plankton community. yielding a growth rate of 0.46d-' (Syvertsen, Holm-Hansen and Sakshaug quoted in Sakshaug 1989). Gilstad & Sakshaug (1990) reported a wider optimum range (70 to 330- 500 pmol m-ls-') for cultures of Barents Sea dia- toms and a strong dependence of growth rate on day length. The growth rate is a function of the difference between gross particulate production rate and cellular loss rates (respiration, extracellular pro- duction). It is thus a function of rates which themselves are composite functions of environ- mental variables. The Q l o value of polar phyto- plankton for dark respiration (i.e. the factorial increase due to a temperature increase of 10°C) is apparently higher ( 2 . 3 1 2 ) than that for gross photosynthesis (1.4-2.2; Neori & Holm-Hansen 1982; Tilzer & Dubinsky 1987). Thus a small temperature change may affect the growth rate considerably by altering the respiration rate. Res- piration rates also tend to increase adaptively with increasing irradiance (Falkowski & Owens 1980; Langdon 1988; Sakshaug et al. 1991a) and may be higher during light hours than dark hours (Weger et al. 1989). The steady-state gross growth rate p + r (d-') of phytoplankton, where r is the daily carbon- specific respiration rate (and in principle includes extracellular production), may be modelled (Ban- nister & Laws 1980; Sakshaug et al. 1989; Cullen 1990) : p + r = (Chla:C). D . gPB = g P c . D (7) In this empirical formulation, g P B may represent equations 3 or 4 with growth-normalised coef- ficient values, and the Ch1a:C ratio, accordingly, is in either mg mg-I or mg ( g a t ) - ' , and D is day length (h). Because growth-relevant carbon represents a fraction of the total fixed carbon, whereas the growth rate is an integrated feature over 24 h and P-I coefficients may vary somewhat diurnally, gPB may differ from PB of P-I curves. Sakshaug et al. (1991a) have suggested that @,,, in equation 4 may be adjusted to a growth-rel- evant value the other coefficients being retained. For equation 3 this implies a pro rata adjustment of 2 and PE to growth-relevant val- ues BCY and g P i , respectively, whereas Ik is retained: p + r = (Chla:C) . D . (g@max/qT) x [l - exp(-"uqtE,)] Toble2. Ch1o:C ratio (mg m g ' ) and the product P:(Chlo:C). h - ' ( = P $ ) for cultures and natural populations in the Barents Sea. The upper mixed layer. when thin. may have Chlo:C ratios 0.025. Same data set as in Table 1 . Culrures: Day length 24 12 prnol m - ' SK' 400 25 400 25 Chla:C Pc, 0.013 0.025 0 043 0.031 0 033 0 047 0.049 0.037 Barenrs Sea: Depth. m 0-20 P', 0.021 Chla:C 0.013 30-60 >70 0 021 0.031 0.027 0.028 Light and productivity of phytoplankton 79 For some Barents Sea diatoms (Table l ) , g&,,ax is on the average 0.04g-at C (mol photons)-]. Equation 8 and its counterpart based on equation 3 may also be written in terms of carbon-nor- malised coefficients; then the changed coefficients (q and 'u or $ and P z , respectively) will necess- arily include the variations of the Ch1a:C ratio. Such equations would be mathematically equiv- alent to equation 7 of Kiefer & Cullen (1991 this volume). At very high latitudes, where a light on/ light off cycle may be hard to define, precision may be improved by replacing the product D . gPc with summation of hourly estimates for gPc. In conjunction with an earlier version of this model (Sakshaug et al. 1989), it was suggested that irradiance-dependent variation in photo- synthetic coefficients might be neglected with little loss of precision. Data for Arctic diatoms (Table l ) , however, suggest that irradiance- dependent variations in "F and the product q t should not be entirely neglected. Variability in OF, and thus @', strongly affects predictions of carbon uptake for low light, whereas variability in q t , and thus gP:, affects predictions for high light strongly. The use of properly chosen con- stants, however, is certainly more convenient than the use of variables and should be adequate for most practical purposes. Because of practical difficulties in measuring the Ch1a:C ratio in the field, other equations than equation 8 may be more convenient. Substitution of the Chla concentration for the Ch1a:C ratio may yield gross daily particulate production of carbon instead of gross growth rate (Cullen 1990). Empirical functions without the Chla:C ratio may also be developed, albeit at a price of less gener- ality. Because P$ varies less with irradiance than P: and the Chla:C ratio individually (Table 2) and the growth rate of diatoms appears to be related to day length by half saturation kinetics (Gilstad & Sakshaug 1990), the gross growth rate may be modelled: F + r = (P + r ) m a x ID/(KD + D)I x [ I - exP(-Eo/Ik.)I (9) where (p + r)max is the asymptotic maximum gross growth rate (d-l), and KD is the half saturation constant for day length (D, h). For the ten Barents Sea diatom species studied by Gilstad & Sakshaug (1990), (p + r)max was 0.75 d-' (r assumed to be 0.05d-l), K,, 12.4h and Ik 2 0 p m 0 l r n - ~ s - ' ("white light"; Fig. 10). This set of coefficients is - 0.4 73 1 0 J 0.2 0.4 predicted d - ' Fig. 10. Predicted (equation 8) versus observed average gross growth rates for 10 Barents Sea diatoms grown at different irradiances (3-5-500~mol m-* s - I ) and day lengths (4-24 h) and -0.5"C. Standard deviation: 0.035 d - ' . Based o n data from Gilstad & Sakshaug (1990). valid only for nutrient-saturated cells at -0.5"C (no photoinhibition assumed), and may serve as a useful guideline for light-limited gross growth rates of phytoplankton in polar waters. Photoinhibition Photoinhibition may be defined as a lowered pho- tosynthetic or growth rate when light becomes strong. On the basis of equation 8, it is evident that photoinhibition of growth can be mediated through an increase in the respiration rate, a decrease in the Chla:C ratio, a change in the photosynthetic coefficients of a combination of these. Strong light does inhibit growth of polar phyto- plankton. In a culture of Antarctic pennate ice- edge diatoms, the growth rate dropped from 0.25 to 0.19 d-l for an increase in irradiance from 104 to 990 pmol m-2 s-I (Sakshaug & Holm-Hansen 1986). The growth rate of a culture of Thal- assiosira bioculata from the Barents Sea dropped from 0.47 to 0.36 d-' for an increase from 230 to 430 pmol m-2 s-l (Sakshaug 1989). The Chla:C ratio, however, dropped relatively more: from 80 Egil Sakshaiig & Dug Slagstand 0.010 down to 0.0040 and from 0.021 down to 0.013 mg mg-I. respectively. This implies that inhibition of growth has been mediated by a decrease in the Chla:C ratio and that ePB (and thus gP$) may. in contrast. have increased. Photoinhibition in terms of Chla-normalised photosynthetic coefficients appears to be likely only during short-term exposure to strong light. e.g. before the cellular concentration of photo- protective pigments has had time to build u p (Lewis & Smith 1983) or if the ability to build u p high concentrations is lacking, which is highly species-dependent (Sakshaug et al. 1987). Inhi- bition may occur when populations in vertically mixed waters are briefly exposed to the illumi. nation near the surface, and shade-adapted popu- lations appear to be more susceptible than light- adapted populations (Platt et al. 1982). Short-term photoinhibition is usually described by the empirical P-I formulation of Platt et al. ( 1980). PB = P![l - exP(-aBEo/P:)l x exp(-pBE,/P:). (10) Pp represents the potential maximum carbon uptake (same units as P:) and BB is a photo- inhibition coefficient (same units as g ) . For pB = 0. this function is equal to equation 3. is highly variable in polar phytoplankton: reported values range from about zero t o 0.021 [see Smith & Sakshaug (1990). for review]. T h e majority of values however. are <0.0003. This implies a reduction in Pa of < 2 4 % relative to P," at 1000 pnol m - ? s-l. Because irradiance ( P A R ) at 0.5 m depth is not higher than 1000-1500 pmol m-'s-I at 63- 74" latitude and the weather is frequently cloudy (Hegseth & Sakshaug 1983; B . G. Mitchell, unpubl. data), photoinhibition of the growth rate as well as of the short-term photosynthetic response should be of little significance in polar regions and certainly in terms of depth-integrated production. The major effect of photoinhibition may be related to the distribution of species. Whereas polar species apparently grow well in strong light and long days in combination with low tempera- ture. a majority of temperate species d o not (E. Sakshaug 8: K . Tangen unpubl. data). In essence. photoinhibition of growth in temperate species becomes more pronounced with decreasing tem- perature (Talling 1957). Thus, many temperate species may have a n over-summering problem in cold waters d u e to strong light, whereas polar phytoplankton may have an oversummering prob- lem in temperate waters because of high tem- perature. Relative importance of environmental variables To evaluate the relative importance of optically relevant parameters for the development of spring blooms, we have run mathematical simulations of the early phase of the bloom. W e have assumed an infinite vertical mixing coefficient in the wind- mixed (homogeneous) layer and have included the effect of latitude. date and spectral dis- tribution of light in the atmosphere (Bird 1984) and water, including self-shading by algae (see Appendix). Two latitudes, 63 and 74"N, and two types of water, "blue" and "green" (Fig. 4). have been chosen. Algal coefficients represent shade- adapted large-celled centric diatoms. The model demonstrates the importance of the depth of the mixed layer for the progress of the spring bloom (Figs. 11. 12). In fact, depths of mixingof 150-200 m , which may not be infrequent in deep polar seas in winter, early spring or late autumn (Holm-Hansen et al. 1977; Blindheim 1989: Loeng 1989), would preclude blooms large enough t o exhaust t h e mixed layer of winter nutri- ents in any polar sea. To exhaust t h e large nutrient concentrations in Antarctic waters, depths of mix- ing would have to be < I 0 m (Kocmur e t al. 1991; Sakshaug et al. 1991b). Another important factor is the optical quality of the water. Because of t h e strong attenuation of blue light in green coastal waters, the Chla concentration would surpass 1 mg m-3 about 60 days later in green than in blue water for a mixing depth of 40 m (compare Fig. 11B with Fig. 12A). A cloud cover which attenuates light by 70%, which may be a realistic average for the atmospheric "low-pressure'' belt, may delay a bloom by at least 20 days and progressively more the deeper the mixing (Fig. 1lB-D). Moreover, a bloom develops later at a high rather than a low latitude by at least 20 days, depending on the depth of mixing. O n the other h a n d , the delay at high latitudes is in part compensated for by the faster development of the bloom d u e ta the more rapid increase in incident light at high latitudes (Fig. 12B-D). Because measurements of light and light absorption by algae at the spectral level (PUR) require expensive instrumentation and a large amount of data processing relative to measure- ments in terms of P A R , we have evaluated the extent to which the two approaches yield different results. The PUR approach takes into con- sideration the spectral variation i n light with depth (variable %I?), whereas the PAR approach is based on total irradiance (40C700nm) and a constant g$ derived on the basis of “white” hal- ogen lamp incubator illumination (Fig. 9). In green water (Fig. 12), the difference between the PUR and PAR predictions is small, because a$ Light and productivity of phytoplankton 8 1 in the upper layers is not very different from the laboratory estimate. For blue waters. the dif- ference is considerable because of the generally higher value of a$ relative to the incubator value (Fig. 11B-C). In both cases, precise knowledge of and ambient irradiance at the spectral level becomes increasingly important the deeper the mixing because the average photosynthetic rate in the column will increasingly be determined by the spectrally dependent part of the P-I curve near the origin. Consequently, spectral information is particularly important in the case of deeply mixed oceanic waters. The model clearly shows that the progress of Fig. 11. Progress of spring bloom in “clearest” ocean water as a function of the depth (m) of the mixed layer. A: 74”N (“Barents Sea”): B- D: 63”N (“Norwegian Sea”). Incident light according to Bird (1984) except for D (70% reduction). For the Southern Hemisphere the time scale should read August- December. For further explanation, see text and Appendix. 1 I I I Feb. March April May June 4 B C 3 82 Egil Sakshaug & Dag Slagstad Feb. March April the spring bloom is most sensitive t o small changes in the light regime when mixing is d e e p , 1.e. the average irradiance in the mixed column is small. I n photosynthetically inefficient regimes such as preen coastal waters. the sensitivity of the model is large even at moderate mixing depths. Although phytoplankton stocks may be fairly pre- cisely modelled on the basis of crude average values for environmental variables when mixing is shallow, every factor plays an important role when mixing is deep. By ”deep” is meant more than 2 0 m and 8 0 m in green and blue waters, respectively. Thus the development of phyto- plankton biomass in marginal light regimes may be most difficult t o model realistically. The model indicates that blue ocean waters have a potential for allowing blooms in early March in the Northern Hemisphere (or early Sep- tember in the Southern Hemisphere), provided there is shallow vertical mixing. Vertical mixing, however. is simply not shallow enough so early in the season because thermal stratification is out of the question. making meltwater from sea ice the o n l y stabilizing agent. Sea ice may start t o melt in April (October in the Southern Hemi- sphere), but usually does not melt appreciably until May in the Northern Hemisphere and Nov- ember in the Southern Hemisphere (Zwally e t al. 1983; Parkinson et al. 1987). If problems related t o estimating mixing rate and depth can b e adequately solved, it should b e possible to predict gross photosynthesis in the May June Fig. 12. Progress of spring bloom in “clearest” green water at 63”N (“Trondheimsfjorden”) as a function of the depth of the mixed layer (m). For the Southern Hemisphere the time scale should read August- December. For further explanation. see text and Appendix. field in crude terms by means of optical data and relatively simple mathematical models. However, the increase in the phytoplankton stock is also very sensitive to variations in loss rates (sedi- mentation, grazing). Moreover, exponential increase in phytoplankton blooms makes them sensitive to the size of t h e initial stock (Sakshaug et al. 1991b). Initial stocks in permanently open waters may, for example, b e much smaller than the relatively large seeding stocks which may be derived from melting ice. Therefore blooms may develop considerably later in ice-free d e e p waters than in the zone of melting ice. Acknowledgemenfs. -This work, which is part of the Norwegian Research Program for Marinc Arctic Ecology (Pro Mare), was financed by the Norwegian Fisheries Research Council and the Nonvegian Research Council for Science and the Humanities. We would like to thank G. B. Mitchell and M. Vcrnet (Scripps Institution of Oceanography. La Jolla). F. Rey (Institute of Marine Research. Bergcn). and G . Johnsen (Trondhjem Bio- logical Station) for providing unpublishcd data. We also thank K . Andrcscn for producing the figures and t w o referees for most helpful suggcstions. Contribution 248. Trondhjem Biological Station References Alexander. V. 1980: Interrelationship between seasonal sea ice and biological regimes. Cold Regiom Sci. Techno/. 2 . 157- 178. Arnold. W. 1032: Kinetics of photosynthesis in Chlorella. Cold Spring Harbor S y m p . Quannrarioe B i d . 3. 124-127. Light and productivity of phytoplankton 83 Homer. R. 1976: Sea ice organisms. Oceanogr. Mar. Biol. Ann. Reu. 14, 167-182. Holm-Hansen. 0.. El Sayed. S. Z., Franceschini, G . A . & Cuhel. R. L. 1977: Primary production and the factors con- trolling growth in the Southern Ocean. Pp. 11-50 in Llano, G. A . (ed.): Adaptation within Antarctic Ecosystem. Gulf Publ. CO., Houston, Texas. Jacques, G. 1983: Some ecophysiological aspects of the Ant- arctic phytoplankton. Polar B i d . 2, 27-33. Jassby. A. D . & Platt, T. 1976: Mathematical formulation of the relationship between photosynthesis and light for phyto- plankton. Limnol. Oceanogr. 21, 540-547. Jerlov, N. G. 1976: Marine Optics. Elsevier, Amsterdam. 231 PP. Jerlov, N. G. & Steemann Nielsen. E. (eds.) 1974: Optical Aspects of Oceanography. Academic Press. New York. 494 PP . Kiefer, D. A . & Cullen, J. J. 1991: Phytoplankton growth and light absorption as regulated by light, temperature, and nutrients. Pp. 163-172 in Sakshaug. E . , Hopkins, C. C. E. & Britsland. N. A. (eds.): Proceedings of the Pro Marc Symposium on Polar Marine Ecology, Trondheim, 12-16 May 1990. Polar Research lO(1). Kiefer, D. A,, Morrow, J . H . , Stramski, D . & Chamberlin, W. S. 1990: An electron transport hypothesis for seasonal changes in the spectral absorption coefficient of particles in the western Sargasso Sea. Eos 7 1 , 97. Kito, J. T . 0. 1975: A theoretical analysis of the contribution of algal cells to the attenuation of light within natural waters. I . General treatment of pigmented cells. New Phytol. 75,l I- 20. Kirk, J . T. 0. 1985: Light and Photosynthesis in Aquatic Eco- systems. Cambridge University Press. Cambridge. 401 pp. Kocmur, S. F., Vernet, M. & Holm-Hansen, 0.1991: RACER: Nutrient depletion by phytoplankton during the 1989 austral spring bloom. Ant. J. U.S. In press. Kopf, U. Heinze, J. 1984: 2,7-bis(diethylamino)phcnazoxon- ium chloride as a quantum counter for emission measure- ments between 240 and 700 nm. Anal. Chem. 5 6 , 1931-1935. Langdon, C. 1988: On the causes of interspecific differences in the growth-irradiance relationship for phytoplankton. 11. A general review. J. Plankton Res. 10, 1291-1312. Legendre, L. Demers, S . . Garside, C.. Haugen, E. M., Phinney, D. A., Shapiro, L. P., Therriault, J.-C. & Yentsch. C. M. 1988: Circadian photosynthetic activity of natural mar- ine phytoplankton isolated in a tank. 1. Plankron Res. 10. Leverenz, J. W. 1988: The effects of illumination sequence, C 0 2 concentration, temperature and acclimation on the con- vexity of the photosynthetic light response curve. Physiol. Plant. 74, 332-341. Lewis, M. R . & Smith, J.-C. 1983: A small volume. short- incubation time method for measurement of photosynthesis as a function of incident irradiance. Mar. Ecol. f r o g . Ser. 13, 99-102. Lewis, M. R . , Warnock, R. E., Irwin, B. & Platt. T. 1985: Measuring photosynthetic action spectra of natural phyto- plankton populations. J. Phycol. 2 1 , 310-315. Ley, A. C. & Mauzerall. D. 1982: Absolute absorption cross sections for photosystem I1 and the minimum quantum requirement for photosynthesis in Chlorella uulgaris. Biochim. Biophys. Acra 680, 95-106. Loeng, H. 1989: Ecological features of the Barents Sea. Pp. 327-365 in Rey, L. & Alexander. V. (eds.): Proc. 6th Conf. Com. Arcr. l n r . . 13-15 May 1985. E. J. Brill, Leiden. 1-6. Bannister, T. T . & Laws, E. A. 1980: Modeling phytoplankton carbon metabolism. Pp. 243-258 in Falkowski. P. G. (ed.): Primary Productivity in the Sea. Plenum, New York. Berner, T . , Dubinsky. Z., Wyman, K. & Falkowski, P. G. 1989: Photoadaptation and the package effect in Dunaliella tertiolecra (Chlorophyceae). J . Phycol. 25, 70-78. Bird, R . E. 1984: A simple, solar spectral model for direct- normal and diffuse horizontal irradiancc. Solar Energy 32. Blindheim, J . 1989: Ecological features of the Norwegian Sea. Pp. 366-401 in Rey, L. & Alexander, V. (eds.): Proc. 6th Conf. Com. Arc. Int., 13-15 May 1985. E. J. Brill, Leiden. Bodungen, B. von. Smetacek. V., Tilzer, M. M. & Zeitzschel. B. 1986: Primary production and sedimentation during spring in the Antarctic Peninsula region. Deep-sea Res. 33, 177- 194. Braarud, T. 1935: The “ 0 s t ” expedition to the Denmark Strait 1929. 11. Phytoplankton and its conditions of growth. Hualrdd. Skr. 10. 1-173. Braarud. T . & Klem, A. 1931: Hydrographical and chemical investigations in the coastal waters off More and in the Romsdalsfjord. Hualrdd. Skr. 1, 1-88. Bricaud, A , , Morel, A. & Prieur. L. 1983: Optimal efficiency factors of some phytoplankters. Limnol. Oceanogr. 28, 816- 832. Campbell, J. W. & Aarup. T. 1989: Photosynthetically available radiation at high latitudes. Limnol. Oceanogr. 34.1490-1499. Cullen J . J. 1990: Models of growth and photosynthesis. Deep- Sea Res. 37. 667-683. Cullen. J . J . & Lewis, M. R. 1988: The kinetics of algal photo- adaptation in the context of vertical mixing. 1. Plankron Res. 10. 1039-1063. Dubinsky, Z . , Falkowski. P. G . & Wyman. K. 1986: Light harvesting and utilization by phytoplankton. Planr Cell Phys- iol. Tokyo 27, 1335-1349. Eppley, R. W. 1972: Temperature and phytoplankton growth in the sea. Fish. Bull. N O A A 70. 1063-1085. Falkowski, P. G. &Owens. T . G . 1980: Light-shade adaptation: two strategies in marine phytoplankton. PIanr Physiology 66. Gallegos. C. L., Platt, T.. Harrison, W. G. & Irwin, B. 1983: Photosynthetic parameters of arctic marine phytoplankton: Vertical variations and time scales of adaptation. Limnol. Oceanogr. 28, 698-708. Gieskes. W. W. C., Veth, C., Woehrmann. A . & Graefe, M., 1987: Secchi disc visibility world record shattered. Eos 68, 123. Gilstad, M & Sakshaug. E. 1990: Growth rates of ten diatom species from the Barents Sea at different irradiances and day lengths. Mar. Ecol. f r o g . Ser. 6 4 , 169-173. Gran, H . H. (1931): On the conditions for the production of plankton in the sea. Rapp. P.-u. Rdun. Cons. l n f . Explor. Mer 75, 37-46. Harrison. W. G . & Platt, T. 1980 Variations in assimilation number of coastal marine phytoplankton: effects of environ- mental co-variates. J. Plankron Res. 2 , 24%260. Harrison, W. G . & Platt, T. 1986: Photosynthesis-irradiance relationships in polar and temperate phytoplankton popu- lations. Polar Biol. 5 , 15S164. Hegseth, E. N. 1989; Photoadaptation in marine arctic diatoms. Polar Biol. 9. 479-486. Hegseth. E. & Sakshaug. E. 1983: Seasonal variation in light- and temperature-dependent growth of marine planktonic dia- toms in in situ dialysis cultures in the Trondhcimsfjord. Nor- way (63”N). J. Exp. Mar. Biol. Ecol. 67. 199-220. 461-471. 592-595. 84 Egil S a k s h u g Ce Dag Slnprtnd Maake. H . C Haardt. 11 t i 19x7: O u a n t i t a t i \ c i r i i , i i , o absorp- tion Lpcctra ot phytoplankton. Dctrital ahsorption dnd com- parison n i t h fluorcwcnce c\citation spectra. Lirnriol. Oceanogr. 32. 6 3 M 3 3 Maykut. G . A L Grcnfell. T c'. I975 T h e spectral distribution o f light beneath first-year sea ice in thc Arctic Ocean. Limnol Oceanogr. 20, 5 5 l 6 3 . Mitchell, G . B & Kietcr. D. A . IYXX. Chloroph!ll 11-specific ahsorption and fluorescence excitation spectra for Iight-lim- ited phytoplankton Deep-Seu Rer 3 5 . 6 3 M 6 3 . Morel. A . 1978: Available. usable and radiant energ! in relation to marine photos)nthesis. D e e p - s e a Res. -7.5. 673-688. Morel. A . 19RX. Optical modeling of the upper ocean in relation to its hiogenous matter c n n t e n t ( c a w I ~ a t e r s ) . J . G e o p h v s . Re$. 9.3 (CY), 10741)-11V6H Morrow. J H . . Chamhcrlin. W S & Kicfcr. D A . 1Y8Y: A t w o - c o m p n c n t dc\cription of spectral absorption h ) marine particles. Limiiol Oceunogr. 34. lFiML150Y Myers. J . & G r a h a m . J - R . 1971. T h e photos)nthctic u n i t of C'/ilorella m e a w r c d b! r c p c t i t n c short flashcs. P l u m Plii>rol. 48. 282-2x6 h'emoto. T Kr Harrison. H C; I Y X I . fiigh l a t r t u d c c c o s ~ s t c m s Pp 95-126 in Longhur\t. A R ( c d . ) : Annlysir of M a r i n e I . ( u , u / e m ~ Academic Press. Kcu York S e o r i . A . C Holm-Hanscn. 0. 19x2: Effcct of temperature on rate of photos!nthcsis i n Antarctic phytoplankton f o l l i r B i d . I . 3 3 - 3 8 . Ncori. A . Vernct. M . . Holm-Hansen. 0. 8: Haxo, F . T . I Y X X : Comparison of chlorophyll far-red a n d red fluorescence exci- tation spectra with photosynthetic oxvgen action spectra for photosystem I I i n algae. M a r . Erol f r o g . Ser. 4 4 . 297-302. Palmisano. A . C . SooHoo. J . B . . Moe. R . L. & S u l h a n . C . W. 1986: Sea ice microbial communities V l l . Changes i n under-ice spectral irradiance during thc development of ant- arctic sea ice microalgal communities. Mar. Ecol. Prog. Ser. 35. 16S173 Parkinson. C L.. Comiso. J C . Z w a l l y . H . J . . Cavalieri. D J . . Gloersen. P & Campbell. W J 1987: Arcrir Sea Ice. 1973-1976- Sarellire P a s s i o r - M i c r o n ~ a i ~ e O b s e r t m o n s . NASA SP-489, Scientific and Technical Information Branch. Wash- ington. D . C ?Y6 pp. Platt. T . & Suhba R a n . D . V . 1975. Primary production of marine microphytcs P p 24Y-28U in C o o p e r . J P ( c d . ) : Phoros~nrhesii utid froducrii~iii in Differeiir Eriivromnenrs Cambridge L!nirersity Press. Cambridge. Platt. T . . Gallcgos. C . L . C Harrison. W . G . IYRO: Photo- inhibition of photos)nthesis in natural assemblages of marine phytoplankton J M u r . H e , . 3X. 687-7(11. Platt. T . . Harrison. W G . , Irwin. B . . H o r n c . E . P. & Gallcgos. c' L 19x2. Photos!nthcsis a n d photoadaptation of marine phytoplankton in the Arctic. Deep-Seu Res. 2Y. 1151)-1170 Post. A . F.. Duhinsky. 2 . W l m a n . li 8: Falkow'ski. P. G IYKI: Kinetics of light-intensit! adaptation in a marine plankton diatom. M a r . Biol. 8.3. 231-238 Rile). G A . 1042: T h e relationship of \ertical turbulence a n d rpring diatom Rouering\ J .Mar. Res 5 . 67-87, Rilc!. G A 1Y46 Factors controlling phytoplankton popu- lations o n Georgcs B a n k . J . .Mar. Res. 6. 54-73. Rochct. M . . Lcgcndre. L . & Dcmers. S. 1986: Photos!nthetic and pigment responses of sea-ice microalgae to changes i n light intensit! and qualit) J. Crp. .Mar. Biol. Ecol. 101. 2 1 1 - 226 Sakshaug. E IYXY T h e phvvological ecolog\ of polar p h > t o - pldnkton P p . 61-89 in R c ? . L. L Alcxandcr. V . ( c d s . ) . Proc. 6rli Cotif, Cum Arcr. I t i r , 1 3 - / S ,Mav IYX5. E . J . Brill. Lcidcn Sakshaug. E b. Andresen. K. IY86. Effect of light regime upon thc g r o u t h ratc and chcmical composition of a clone of .SA121eronema cosruriim from the Trondheimsfjord. J. flunkroti R r , . 8. 619-637. Sakshaug. E. L Holm-Hansen. 0. 1984: Factors governing pelagicproduction in polaroccanc. P p . 1-1Xin Holm-Hansen, 0 . . Boliz. L. 8; Gilles. R . ( c d s . ) . Marine fhyroplankroti and froducriivri Springer. Berlin. Sakshaug. E L Holm-Hanscn. 0. IYX6: Photoadaptation i n Antdrctic phytoplankton: Variations in growth rate. chemical composition. and P v s I curves. J . Pluiikrori Hes. 8. 459- 473 Sakshaug. E . Demers. S . & Yentsch, C. M . 1087: Tlialassiosira oceunicu and T . p s e u d o f i u t i a : T w o different photodd- aptational rcspomes. M a r . Ecol. frog. Ser. 4 1 . 27s-282. Sakshaug. E.. Andresen. K. & Kiefer. D . A . 1YRY. A steady state description of growth and light absorption in the marine planktonic diatom Skelerotiemu cusrarum. Limfiol. Oceanogr. 3 4 . lYb205 Sahshaug. E . Johnsen. G . . Andresen. K . Kr Vcrnct. M . 1991a: Modeling of light-dcpcndcnt algal photosynthesis and growth: Experiments with thc Barcntq Sca diatoms Thal- ussrosiru norrle~iikioeldii a n d Chuerocero& firrcellurur. D e e p Sea Rei 38. 415-430. Sakshaug. E . Slagstad. D. & Holm-Hanscn. 0. 1YYIh: Factors controlling the development of phytoplankton blooms in the Antarctic Ocean - a mathematical model. M a r . Chem. In press Sath!cndranath. S 8: Platt. T 1989: Computation of aquatic priniar) production. Extended formalism to include effect of angular and spectral distrihurion of light. L i m w l . O c e a n o g r . 3 4 . 188-39K. Smetacck. V. & Passow. U . 1990: Spring bloom initiation and Sverdrup's critical-depth model. Limnol. Oceanogr. 35.228- 233 Smith. R . C . & Baker. K. S . 19x1: Optical properties of the clearest natural waters (?(Mb81)0 n m ) . A p p l . Oprics 2 0 . 177- 1 u4 Smith. R . C . Marra. J . . Perry. M . J . . Baker. K. S . . Swift, E.. Buskey. E & Kiefer. D . A . 1989: Estimation of a photon budget for the upper ocean in the Sargasso Sea. Limnol. Oceatiogr. 3 4 . 1673-1693. Smith. W . 0. & Sakshaug. E . 19Y0: Polar phytoplankton. Pp. 477-525 In Smith. W. 0.: Polar O c e a n o g r a p h y . Parr B: Cliemisrri . Biology and G e o l o g y . Academic Prcss. New York SooHoo. J B . . Palmisano. A . C.. Kottmeier. S . T . . Lizotte. M. P . S u o H o o . S . L . & Sullivan. C. W 1987: Spectral light ahsorption and q u a n t u m yield of photosynthcsis in sea ice microalgae and a bloom of Phaeocysrrs poiccherii from McMurdo Sound. Antarctica. M a r . Ecol. P r o g . Ser. 39. 175- 1x4, Spies. A . lYX7: G r o w t h rates of antarctic marine phyroplankton in the Wcddcll Sea. M a r . Ecol f r o g . Ser. 4 1 . 267-274. S u h b a R a o . D V . & Platt. T. 1984: Primary production of arctic walcrs. f d a r &I/. .?. 191-201. Svcrdrup. H. U. 1053: On conditions for thc vernal blooming of phltoplankton. J. Cons. Perm. Inr. Explor. M e r 18. 287- 295 Talling. J F 1Y57: Photosynthetic characteristics of some fresh- water plankton diatoms in relation to underwater radiation. .Yew f h y r o l 56. 2%50. Tilzer. M . M 8: Duhinsky. Z. IYX7: Effects of temperature and day lcngth on thc mass balance of antarctic phytoplankton. Polar t?io/. 7. 35-42, Light a n d productivity of p h y t o p l a n k t o n 85 sumably does not significantly cxceed the dark resprration ratc of polar phytoplankton. Spectral distribution at the sea surface (no clouds) is a func- tion of sun elevation (Bird 1984) and model output is the s u m of direct and diffuse irradiance. Water vapour pressurc in the atmosphere has been set at 1.42 cm and turbidity r,(0.5) cquals 0 . 1 . Reflection losscs at the surface arc set at 5510% dcpcnding on sun elevation (Kirk 1985). Spcctral distribution of downwelling light is calculated by a model similar to that of Sathyendranath & Platt (1989): Total downwelling I(2.A) at depth z and wavelength I is partitioned into a direct component Id and a diffuse component 1,: I(2.A) = l d ( 2 . A ) + I,(z,I). (13) Vertical attenuation coefficients for diffuse and direct light are given by K,(z.A) = [a(z.A) + bh(z.h)]/cosUd K , ( z . I ) = (a(2.I) + b,,(zJ.)]/b ( 1 4 ) and ( 1 5 ) where a ( z , l ) is the volume absorption coefficient at depth z and wavelength A , and bh(z,A) is thc corresponding backscattcring coefficicnt. 0, is thc sun zenith angle in water and p i s the mean cosine of zenith angles of diffuse light 0, after refraction at the surface (set at 0.6: Kirk 1985). Multiple scattcring is neglected. a ( I ) is the sum of the absorption coefficient of pure water aw(A). either "clearest" blue o r "clearcst" green water (Fig. 4). and absorption due to particles * a r . (Chla*:C). Chla' is thc sum of the Chla concentration and half the concentration of Phaeophytin a . *a,is0.0098 m 2 (mgChla*)-' in "white" halogen lamp illumination. The Phaeophytin a concentration In the Barents Sea is related to Chlu according to the function ( F Rey, unpublished data) Phaeoa = 0.45 Chla + 0.02 (mg m - > ) (16) Algal coefficients are representative for shade-adapted cclls of the large-celled centric diatoms Tha/assiosirn anforcficn and T . nodenskioeldii (at 25 ymol m-'s-') using the scaled fluor- escence excitation spectrum "F for 'u in equations 6 and 8. "F and 'a, for white halogen lamp illumination are 0.0058 and 0.0090mZ ( m g Chlo)-', respectively. The Chl ratio is 0.052 mg mg-I, e&,d, is 0 . 0 4 g-at C (mol photons)-'. and the product qr is 1370 mg(mol PSU).' h. Thus the growth-relevant values RPc and ens ("white" light) are 0 . 3 5 m g C (mg ChIa)-l h - ' and 0.010 mg C (mg Chlu).' h - ' (Fmol m - ? s - ' ) - ' . respectively. The maximum (light-saturated) gross growth rate is 0.44d-I. The equations were solved by a finite difference scheme using a MicroVax computer. The vertical grid point was 5 m and time step was 1 h. The initial concentration of Chla was 0.05 mg m - > . Tylcr. J . E. ( c d . ) 1977: Ligl7f in fhe Sea. Dowdcn, Hutchinson & Ross. Stroudshurg. Pennsylvania. 384 pp. Vernet. M . , Ncori, A . & Haxo, F. T. 1989: Spcctral propcrtics and photosynthetic action in rcd-tidc populations of Proro- cenfrutn micafis and G o ~ 7 y a u / a s p o l y e d r u . M a r . Biol. 193, 365-371. Webb, W. L . , Newton, M. & Starr. D. 1974: Carbon dioxide exchange of Abtus rubm: A mathematical model. Oecologia 17, 281-293. Weger, H . G . . Herzig, R . Falkowski, P. G. & Turpin. D. H . 1989: Respiratory losses in the light in a marine diatom: Measurements by short-term mass spectrometry. Limnol. Oceanogr. 34. 1 1 5 ~ l 1 6 1 . Wilson, D . L . . Smith. W. 0. & Nelson, D . M . 1986: Phyto- plankton bloom dynamics of the western Ross Sea edge. 1. Primary productivity and species-specific production. Deep- Seri Res. 33, 1375-1387. Ziemann, D. A , , Conquest. L. D., Bienfang, P. K . & Kanda. J . 1987: Patterns of primary production and sedimentation during the 1987 spring bloom in Aukc Bay. Alaska. Pp. 29- 194 in APPRISE nfifiual r e p o r f . SFOS APP87-100. School of Fisheries and Ocean Sciences, Unrvcrsity of Alaska, Fair- banks. Zwally, H . J . , Comiso. J . C.. Parkinson. C . L . . Campbell, W. J . . Carsey. F. D . & Glocrsen. P. 1983: Anfarcric Sea Ice. 1973-1976: Safellife Passioe-Microwaue Obseroafions. NASA SP-459, Scientific and Technical Information Branch. Wash- ington. D.C. 206 pp. Appendix: Formu/afion of model Assuming negligible lateral coefficients. phytoplankton dis- tribution is affccted hy vertical turbulent mixing and vertical transport as in the 1-D model where P(1.z) is phytoplankton Concentration at depth z and time t , w is vertical (usually sinking) velocity, K,(z) is the vertical turbulent mixing coefficient. and fhllll describes net phyto- plankton growth. The vertical mixing coefficient is infinite above the depth of mixing and zero elsewhere. For the early phase of the spring bloom. nutrient limitation is ignored. Thus f h l c 4 = p(t.Z){['pc (E,,(h),t.z)] - R) (12) where [pPc (E,,(A),t,z)] is the turnover rate of growth-relevant carbon ( h - ' ) as a function of spectral irradiance. time and depth (see equations 6 . 7 and 8 ) . R rcprcscnts in principle the total loss ratc. i.c. the sum of dark respiration. sedimentation and grazing rates. hut is set at a low value. 0.0046 h-'. which pre-