Variability of phytoplankton pigment ratios across aquatic environments

Transcription

1 European Journal of Phycology ISSN: (Print) (Online) Journal homepage: Variability of phytoplankton pigment ratios across aquatic environments Jean-Pierre Descy, Hugo Sarmento & Harry W. Higgins To cite this article: Jean-Pierre Descy, Hugo Sarmento & Harry W. Higgins (2009) Variability of phytoplankton pigment ratios across aquatic environments, European Journal of Phycology, 44:3, , DOI: / To link to this article: Published online: 19 Aug Submit your article to this journal Article views: 671 View related articles Citing articles: 18 View citing articles Full Terms & Conditions of access and use can be found at Download by: [ ] Date: 30 December 2017, At: 15:22

2 Eur. J. Phycol., (2009), 44(3): Variability of phytoplankton pigment ratios across aquatic environments JEAN-PIERRE DESCY 1, HUGO SARMENTO 1 AND HARRY W. HIGGINS 2 1 Laboratory of Freshwater Ecology, URBO, University of Namur, Rue de Bruxelles 61, B-5000 Namur, Belgium 2 CSIRO Marine and Atmospheric Research, GPO Box 1538, Hobart 7001, Australia (Received 5 February 2008; revised 30 March 2008; accepted 28 October 2008) The majority of phytoplankton pigment studies are from marine, estuarine and oceanic waters, and commonly use estimates of the ratio between marker pigments and chlorophyll a (chl a) for calculating the contribution of phytoplankton groups to total chl a. In this study, we examined pigment ratios obtained with CHEMTAX processing of field data from a range of tropical and temperate freshwater bodies with contrasting water transparency, depth of the mixed layer, and trophic state. The pigment ratios obtained from processing data from fresh waters corresponded quite well with existing values from pure cultures, and were compared with the marine marker pigment:chl a ratio calculated with CHEMTAX using identical procedures. In deep, stratifying lakes a large variation of some pigment ratios with depth was observed, as well as seasonal variation relating to changes in water column structure. There was considerable variation of average phytoplankton pigment ratios among the freshwater bodies studied. Pigment ratios were significantly correlated to indicators of nutrient availability, to depth of the euphotic zone, or to a proxy of light availability. The substantial variation in marker pigment:chl a ratio confirms that algorithms which account for natural variation of pigments in phytoplankton groups are required for accurate assessment of phytoplankton groups based on marker pigment concentration alone. Key words: phytoplankton, diagnostic pigments, HPLC, CHEMTAX, lakes, rivers, tropical, temperate Introduction The use of marker pigments to identify and quantify the various algal groups of phytoplankton has been widely and successfully used in marine, estuarine (Wright & Jeffrey, 2006 and references therein), freshwater environments (Woitke et al., 1996; Yacobi et al., 1996; Edelkraut et al., 1997; Descy et al., 2000, 2005; Fietz & Nicklisch, 2004, 2005; Buchaca et al., 2005) and even for the study of photosynthetic bacteria (Steenbergen et al., 1989; Hurley & Watras, 1991; Vila et al., 1996; Goericke, 2002). However, as the biomass estimation of algal groups is derived from pigment concentrations in the water column by calculations based on marker pigment:chlorophyll a (chl a hereafter) ratios, the technique is prone to errors related to changes in the cell pigment content, which may result from acclimation to the light climate and from nutrient deficiency. Variation of pigment ratios in phytoplankton has been addressed in studies on cultures, e.g. by Goericke & Montoya (1998), who conducted culture experiments with several marine planktonic algae. Correspondence to: J.P. Descy. jpdescy@fundp.ac.be They observed that, in most cases, light-harvesting carotenoids (for example, fucoxanthin in diatoms, alloxanthin in cryptophytes) co-varied with chl a in nitrogen- and light-limited cultures. In the species tested, xanthophylls:chl a ratios showed less variability than the pigment content per cell. The observed ratio, however, still varied by a factor up to two as a function of growth rate or irradiance, with the possible result of errors when relying on ratios derived from cultures for estimating the contribution of algal groups to chl a biomass. Similar constant photosynthetic carotenoids:chl a ratios were observed by Schlu ter et al. (2000), who analysed mesocosm field data using CHEMTAX (Mackey et al., 1996), and ratios from laboratory cultures grown under different light conditions. However, in some taxa, the marker:chl a ratio increased by at least 100% from low light to high light. This was the case among the diagnostic pigments, e.g. zeaxanthin, a photoprotective carotenoid, which would be expected to increase under high-light conditions. Similar studies were carried out by Nicklisch & Woitke (1999), who simulated the effects of natural fluctuations of light and photoperiod in several freshwater algae, with varying results according to the species tested. ISSN print/issn online/09/ ß 2009 British Phycological Society DOI: /

3 J.-P. Descy et al. 320 Carotenoid:chl a ratios in most species tested were relatively insensitive to the light conditions, but in the filamentous cyanobacterium Planktothrix agardhii, in which the zeaxanthin content per cell did not vary, the zeaxanthin:chl a ratio changed significantly under fluctuating light as compared to constant light. Fewer studies have demonstrated intra-class variations of pigment composition and cell content (Nicklisch & Woitke, 1999; Zapata et al., 2004; Muylaert et al., 2006 and references therein), which are potentially a problem for the assessment of phytoplankton biomass based on pigments, but can also help increase the discriminative power of the pigment method. The recent study by Schlu ter et al. (2006), however, represents significant progress in defining adequate pigment ratios for the main freshwater phytoplankton classes by taking into account the light conditions. That study also considered intra-class variation of pigment content: for instance, large variation in pigment composition occurred among cyanobacteria species, as well in their in response to environmental factors. In contrast to laboratory studies, large variations in pigment ratios have been seldom reported from field data. Brunet et al. (1993) reported variation of the (diatoxanthin þ diadinoxanthin):chl a ratio by a factor up to five along an inshore offshore transect in the English Channel; this apparent increase of photoprotective chromophyte pigments was well correlated to vertical light attenuation and turbidity. Similar observations were made by Riegman & Kraay (2001) with a decrease in the (diadinoxanthin þ diadinoxanthin):chl a ratio with increasing depth. Some changes were also found when processing pigment data from different depth categories with CHEMTAX (Descy et al., 2000, 2005; Higgins & Mackey, 2000; Higgins et al., 2006). Some studies combined HPLC analysis with other techniques, such as flow cytometry, to follow changes in cell pigment in specific populations of autotrophic picoplankton (Veldhuis & Kraay, 2004). The obvious problem for interpreting variations of marker:chl a ratios from field measurements is that a change may result from changes in the marker cell content, or from chl a cell content, or from changes in both marker pigment and chl a, in different proportions. However, the experimental studies referred to above indicate that, in case of acclimation to low light, the increase of lightharvesting pigments per cell occurs in similar proportions, so that the ratio of a photosynthetic carotenoid to chl a does not show substantial variation. In contrast, as a result of exposure to high light, chl a per cell decreases whereas photoprotective carotenoids increase; therefore, a photoprotective pigment:chl a ratio is expected to increase in response to high light. Following this reasoning, changes in pigment ratios, such as those resulting from calculations by an optimisation algorithm, can to some extent, be used to explore how, in nature, phytoplankton groups respond to changes in light exposure, either in a given water body (with depth or as a result of seasonal variations of light), or among water bodies of different transparency and trophic status. In particular, changes in the ratios of photoprotective carotenoids to chl a should present an unambiguous response, whereas little change would be expected for the ratio of light harvesting carotenoids to chl a. In the present study, we used various HPLC determined phytoplankton pigment datasets, mostly from fresh waters, but also from a set of oceanic samples. The datasets were either from previously published studies dealing with phytoplankton classes assessment using CHEMTAX (e.g. Descy et al., 2000), or from reprocessing of published data (e.g. Descy et al., 2005), or from new, unpublished pigment data. We processed these data sets with CHEMTAX, following identical procedures, and examined the variation of key marker pigment:chl a ratios in different algal groups, and compared our results with literature data from pure phytoplankton cultures and field studies. Material and methods Study sites Several freshwater bodies were studied, sometimes for several years, with sampling intervals sufficiently frequent (usually weekly or occasionally bi-weekly) to account for seasonal variability, and for changes with depth and water column structure in lakes. Two pigment datasets were from tropical, oligotrophic systems, Lake Tanganyika, off Kigoma, Tanzania (Descy et al., 2005) and Lake Kivu, off Bukavu, DR Congo (Sarmento et al., 2006, 2007), both located in East Africa. These two lakes, located in the Western Rift, are meromictic, but have distinct characteristics. Lake Tanganyika has very active hydrodynamics, due to its large size and exposure to winds, and a mixed layer which varies from 25 m (in the rainy season) to >100 m (dry season), depending on location (Coulter, 1991). In comparison, Lake Kivu lies closer to the Equator and is smaller and located at higher elevation; it is oligotrophic with a mean chlorophyll a of 2.3 mgl 1, and has characteristics closer to shallower lakes in the region, e.g. Lake Victoria (Sarmento et al., 2006). As a result, Lakes Kivu and Tanganyika have different phytoplankton assemblages with different seasonal dynamics and distribution in the water column. Further details on the differences between these two lakes can be found in Sarmento et al. (2006).

4 Pigment ratios across aquatic environments 321 Pigments ratios from two temperate, oligotrophic lakes, Crystal Lake and Little Rock Lake (Descy et al., 2000 and located in Northern Wisconsin, USA were also used. Finally, other datasets came from a meso-eutrophic reservoir, located at Esch-sur Suˆre, Grand-Duchy of Luxembourg (Thys et al., 2003), from two eutrophic reservoirs (Lake Ry Jaune and Lake Falemprise, Belgium); and from a eutrophic river, the River Meuse in its Belgian stretch (Gosselain, 1998; Viroux, 2000). These water bodies can be ranked according to their mean euphotic layer thickness, with a gradient from the more transparent waters of Lake Tanganyika during the rainy season through to the more turbid River Meuse. Macronutrient concentrations and trophic status also vary considerably, so that the whole range of trophic status was covered, as shown by total phosphorus (TP) and chl a concentrations (Table 1). The ranking by euphotic depth does not totally match the ranking by TP and chl a, probably because the North Wisconsin lakes, particularly Little Rock Lake, have substantial amounts of dissolved organic carbon (DOC), which influences light penetration. Also, the two tropical lakes have deep mixed layers, and as phytoplankton is distributed evenly throughout this layer, they have lower chl a per unit volume and higher water clarity. Despite these small differences among the trophic gradient, highly significant correlations were found between mean depth of the euphotic zone (Zeu) and mean chl a, chl a and mean total phosphorus (TP), TP and DIN/ TN, and between Zeu and TP (Table 2). The ratio of the mean depth of the mixed layer to the mean depth of the euphotic zone (Zm:Zeu) was used as a proxy of average exposure to light of the phytoplankton in a fully mixed water column. Average values of environmental factors were calculated either on an annual basis (for the tropical lakes, which were permanently stratified, and for the river which was fully mixed at all times), or for the stratified period in temperate lakes. Pigment ratios were determined for the corresponding periods. For a marine reference, we used data from the warm pool region of the western equatorial Pacific Ocean (WEP) collected during Australian JGOFS cruises (Higgins & Mackey, 2000) and TROPICS cruises (Higgins et al., 2006) in the Bismarck Sea, north of New Guinea. This region of the WEP is highly stratified with a stable and deep oligotrophic warm pool overlying cooler, denser nutrient-rich waters. The deep thermocline and nutricline, combined with a barrier layer produced by shallow haloclines, prevents the upwelling of nutrient-rich water into the euphotic zone even when surface winds from the east are favourable. The isothermal waters in the warm pool contain a number of haloclines and the mixed-layer depth is generally shallower than the top of the thermocline by up to 40 m. The phytoplankton community in this region exhibits a substantial degree of photo-acclimation and is vertically well resolved with a Deep Chlorophyll Maximum (DCM) at the base of the warm pool where there is a balance between decreasing light and increasing nutrients. Although in most marine situations, including the Table 1. Characteristics of the studied locations, with location name, latitude, longitude, mean mixed layer nutrients (TP, TN or DIN, SRSi) and chlorophyll a concentrations. The dominant phytoplankton groups are also indicated (see text for group description). Chl a, mgl 1 Phytoplankton composition TP, mm DIN/TN, mm SRSi, mm Mean Zm, m Zeu, m Elevation, Depth m asl Area, km 2 max, m Location W E Pacific, surface 0 00 S; E < Prochlorococcus, haptophytes, Synechococcus, chrysophytes W E Pacific, Chl max 0 00 S; E < Prochlorococcus, haptophytes, Synechococcus, chrysophytes W E Pacific, >Chl max 0 00 S; E < Prochlorococcus, haptophytes, Synechococcus, chrysophytes, diatoms L. Tanganyika 4 51 S; E Cyanobacteria T2, chlorophytes, diatoms T1 Crystal Lake N; W Chrysophytes, chlorophytes L. Kivu, great lake 2 34 S; E Diatoms T1 and T2, cyanobacteria T1 and T2 L. Kivu, Bukavu Bay 2 34 S; E Diatoms T1, cyanobacteria T1 and T2 Little Rock Lake N; W Chrysophytes, chlorophytes L. Ry Jaune N; 4 24 E Cryptophytes, cyanobacteria T2, chrysophytes, diatoms T1 L. Esch-sur-Suˆre N; 5 52 E Cryptophytes, diatoms T1, chlorophytes L. Falemprise N; 4 24 E All groups, with cyanobacteria T1 blooms River Meuse N; 4 50 E Diatoms T1 and chlorophytes Note: Data on Little Rock Lake and Crystal Lake from LTER, by courtesy of T.K. Kratz, Trout Lake station, University of Wisconsin, Madison, USA.

5 J.-P. Descy et al. 322 Table 2. Significant correlation coefficients (threshold: p < 0.05; except when in italics) between environmental variables and pigment ratios in the freshwater bodies studied. WEP warm pool, nitrate is usually considered as the limiting macronutrient; phosphate concentrations are also usually low. For comparison with freshwater data, in Table 1, we have included the mean euphotic zone and mixed layer depths, and the mean nutrient and chl a concentrations in the surface, DCM and below DCM layers for the 4 JGOFS/TROPICS cruises in the WEP warm pool region. It can be seen that the WEP warm pool fits the pattern of a decrease in chl a as TP (trophic status) is reduced. Field sampling n Spearman R p-value Zeu & CHL A < Zeu & TP CHL A & TP < DIN/TN and TP FUCO:Chl a & CHL A FUCO:Chl a &TP DDX:Chl a & CHL A DDX:Chl a &TP LUT:Chl a & CHL A LUT:Chl a & TP ZEACH:Chl a & Zm:Zeu ZEACY:Chl a & Zeu ZEACY:Chl a & CHL A ZEACY:Chl a &TP Chlb:Chl a & Zm:Zeu Note: The non-parametric Spearman rank correlation coefficient was used. Abbreviations: Zeu: mean depth of the euphotic layer; CHL A: mean chlorophyll a concentration; TP: mean phosphorus concentration; FUCO: Chl a: fucoxanthin:chlorophyll a ratio in diatoms; DDX: Chl a:(diadinoxanthin þ diatoxanthin):chl a ratio in diatoms; LUT: Chl a:lutein:chl a in chlorophytes; ZEACH: zeaxanthin:chl a in chlorophytes; ZEACY: zeaxanthin:chl a in cyanobacteria T2; Chl b: Chl a:chl b:chl a in chlorophytes. Freshwater sampling was carried out, usually fortnightly, except in the winter period in temperate sites, by taking discrete samples at various depths with Van Dorn or Niskin sampling bottles, depending on water column structure, determined from temperature and dissolved oxygen profiles (using a HYDROLAB DATASONDE 4). Light penetration was measured with surface and underwater LICOR quantum sensors; the vertical attenuation coefficient (k ) was calculated and depth of the euphotic layer (Zeu) was estimated as 4.6/k. In the River Meuse, a single sample was taken beneath the water surface. For the temperate lakes, only the stratified period was considered, and several samples were taken in the mixed layer, which was considered to extend from the surface to the top of the thermocline. Sampling procedures used in the marine WEP studies are fully described in Higgins & Mackey (2000) and Higgins et al. (2006). Pigment analysis All pigment analyses were made using similar HPLC methods. Analyses of freshwater samples followed a procedure described in Pandolfini et al. (2000). Briefly, water samples (from 0.5 to 4 l, depending on the water body and phytoplankton biomass) were filtered on Whatman GF/C or GF/F filters, or on material of similar porosity. Extractions were carried out in 90% acetone HPLC grade, with two 15 min sonication steps separated by an overnight storage at 4 C. Pigment analysis was carried out by HPLC using the Wright et al. (1991) gradient elution method, with a Waters system comprising a Waters 996 PDA detector and a Waters 470 fluorescence detector. Calibration was made using commercial external standards (DHI, Denmark). Carotenoids not present in the standard were quantified against fucoxanthin, using as relative response the ratio of the specific absorbance coefficients at 440 nm (Jeffrey et al., 1997) in methanol. Identification of pigments was checked against a library of pigment spectra, obtained by diode array acquisition of chromatograms from pure pigment solutions and from acetone extracts of pure cultures of algae. Chromatograms processing was done with the Waters Millennium software. The analytical procedure for marine pigment samples used similar methods, detailed in Higgins et al. (2006). The results presented here are from processing more than 2000 individual samples. Pigment data processing Abundances of algal taxa were determined from HPLC algal pigment measurements using CHEMTAX. CHEMTAX is a matrix factorisation program developed by Mackey et al. (1996), in order to estimate biomass (in chl a units) of algal classes from concentrations of marker pigments determined by HPLC analysis of water column samples. Essentially, CHEMTAX optimises the marker pigment:chl a ratios for a particular data set, from an initial input matrix which contains the initial estimates of the ratios for the algal classes known to be present. Although the calculation procedure allows for adjustment of the ratios within a userdefined interval, these initial values can influence the final biomass estimates and may be critical in providing correct quantitative phytoplankton assessment. Typical ratios have been given in the literature from cultures of marine (Jeffrey et al., 1997) and, to a lesser extent, for freshwater algae (Nicklisch & Woitke, 1999; Schlu ter et al., 2006). A recent paper by Latasa (2007) described a method to reduce processing errors which may be related to wrong estimates of initial ratios: repeating CHEMTAX processing in successive runs, using the output from each run as the input for the next one, the pigment ratios generally adjust automatically towards the true value, improving initial pigment ratio values and, therefore, biomass estimates. In other words, whatever the uncertainties of the initial ratio values, with repeated CHEMTAX runs as described above, the pigment ratios will eventually converge towards the true value, and a good biomass estimation.

6 Pigment ratios across aquatic environments 323 The processing of freshwater samples depended on the dataset, but in all cases the procedure followed the principle of using the output matrix from each run as the input for the next one for each data bin, as described by Latasa (2007). For the two tropical lakes, bi-weekly pigment data were available for two years (2002 and 2003) and were processed by depth bins, taking into account the depth of the mixed layer, which varied seasonally (details in Descy et al., 2005 and Sarmento et al., 2006): each batch of data contained a single depth within the mixed layer (every 20 m for Lake Tanganyika, every 10 m for Lake Kivu) and a single season (dry season or rainy season). Each bin had at least 20 samples. For the reservoirs, 2 or 4 years of weekly to bi-weekly data were available, every 2.5 m; spring and autumn overturn were distinguished from the stratified summer period, and the data from the spring and fall overturn, from the epilimnion and the metalimnion were processed in separate bins. For the River Meuse, the pigment concentrations from March October 2000 were used. Pigment ratios for the two Wisconsin lakes are those from the study by Descy et al. (2000), from which the initial ratio matrix used for CHEMTAX processing of the lake data was derived. The initial ratio values for the river Meuse data were derived from Descy & Me tens (1996) and from subsequent studies on the same river (Viroux, 2000). Binning and processing of marine pigment data is fully described in Higgins et al. (2006). The pigments used for fitting the algal class abundances in the freshwater samples were: peridinin (PERI), fucoxanthin (FUCO), neoxanthin (NEO), violoaxanthin (VIOL), the sum of diadinoxanthin and diatoxanthin (DDX), alloxanthin (ALLO), lutein (LUT), zeaxanthin (ZEA), chlorophyll b (chl b), chlorophyll a (chl a), echinenone (ECHI) be-carotene (BECAR), bccarotene (BPCAR) and bb-carotene (BBCAR). Summing diadinoxanthin and diatoxanthin (DDX) is justified as these two molecules are converted into each other, depending on the light exposure (Demers et al. 1991). The algal classes used were chlorophytes (NEO, VIOL, LUT, ZEA, chl b), chrysophytes (FUCO, VIOL), cryptophytes (ALLO, BECAR), cyanobacteria T2 (originally known as cyanobacteria T1 but now know as cyanobacteria T2, Jeffrey & Wright, 2006; ZEA), cyanobacteria T1 (originally known as cyanobacteria T2 but now know as cyanobacteria T1, Jeffrey & Wright, 2006; ZEA, ECHI), diatoms T1 (FUCO, DDX), dinoflagellates (PERI, DDX) and euglenophytes (DDX, chl b). In Lake Kivu, a particular pigment type was defined for some diatoms (diatoms T2, FUCO, DDX, BPCAR; Sarmento et al., 2006). Samples containing more than 25% chl a degradation products (phaeophorbides a and phaeophytins a) were discarded. For evaluating the relationships between pigment ratios (w:w) and the environmental variables, owing to the low number of observations (n ¼ 8 or 10), we used the Spearman rank correlation coefficient, calculated with the Statistica Õ software, rel. 5.5, from StatSoft, Inc. Results Changes in the water column within the same environment Variations of pigment ratios with depth, as calculated by CHEMTAX, were observed in the two deep tropical lakes, and were particularly clear for photoprotective pigments (Fig. 1), although other changes in pigment ratios, as for FUCO:chl a in diatoms and chrysophytes or for chl b:chl a in chlorophytes were also evident. The ZEA:chl a ratio in cyanobacteria T2 changed with depth during the rainy season in the water column of Lake Tanganyika (Fig. 1A). In these stable stratification conditions, with a persistent thermocline at m, and often a secondary shallow thermocline (on the calm, sunny days), cyanobacteria T2 presented maximal abundance in the surface layers, where they were exposed to high light; these surface layers were also nutrient depleted. In the dry season, ZEA:chl a was lower than in the rainy season in the surface layer, and did not vary substantially in the water column. In the dry season, deep vertical mixing occurs, driven by south-east trade winds, and cells experience exposure to lower average light. In diatoms, FUCO:chl a tended to increase with depth in most cases; the trend is more reliable for the dry season data, as diatoms become more abundant than in rainy season conditions (Figs 1B, D). In contrast, the ratio of photoprotective pigment, DDX:chl a showed a clear tendency to decline with depth when diatoms were abundant. This was clearer in Lake Kivu, where diatoms are more abundant throughout the year (Fig. 1C). In diatoms and cyanobacteria, the decrease of the ratio of photoprotective pigments to chl a was likely a response to lower light in at the bottom of the euphotic layer or to deep vertical mixing. Changes in the water column of different water bodies The variation of pigment ratios in the different water bodies is shown in Figs 2 5, where the water bodies are ranked according to mean TP concentration, which is inversely related to euphotic depth, as shown in Table 1. Given the large differences in sample number (more than 300 samples for Lake Tanganyika, and only 72 for Crystal Lake) and in the period of studies, a statistical analysis for significant differences was not undertaken. Instead, we plotted the average ratio in each water body for those algal groups which reached high levels of abundance, and analysed the trends. We retained only pigments of algal classes which were present in most water bodies examined, including the marine reference.

7 J.-P. Descy et al. 324 Fig. 1. Variations of some pigment ratios with depth observed in Lake Tanganyika (A and B) and Lake Kivu (C and D), Eastern Africa. Diatoms T1: diatoms with the usual pigment type, i.e. with fucoxanthin and diadinoxanthin-diatoxanthin as major carotenoids. Pigment ratios were obtained by processing the pigment concentrations with CHEMTAX on separate depth/season bins. The FUCO:chl a ratio in diatoms (Fig. 2) was higher in the most oligotrophic lakes, where it reached , i.e. approached the range reported for the oligotrophic oceanic waters of the WEP ( , DCM and above). It was, however, lower in Lake Kivu ( ), where the value was close to that obtained in the mesoeutrophic lake; the lowest value was found in the less transparent and the most eutrophic River Meuse (0.49). This FUCO:chl a ratio was best related to the mean TP of the water bodies (r ¼ 0.68; p ¼ ). The ratio of DDX (diadinoxanthin þ diatoxanthin) to chl a in diatoms varied over a much larger range ( ), and was also negatively correlated with chl a and TP (when values from Lake Tanganyika were removed, see discussion). The alloxanthin:chl a in cryptophytes (Fig. 3) varied relatively little among the freshwater bodies in which flagellates reached relatively high levels of abundance (data are missing for Lake Tanganyika, where cryptophytes were present at very low biomass): it was usually between 0.25 and 0.35 and did not show any trend in our set of freshwater bodies. Observations were somewhat similar for the lutein:chl a ratio in green algae (Fig. 3, which varied in a narrow range ( ), with the exception of the

8 Pigment ratios across aquatic environments 325 Fig. 2. Variations of FUCO:Chl a and DDX:Chl a in diatoms T1, in the Western Equatorial Pacific (data from Higgins & Mackey, 2000 and Higgins et al., 2006, used as marine references) and in the set of freshwater bodies investigated, ranked according their trophic status, increasing from left to right. Fig. 3. Variations of ALLO:Chl a in cryptophytes and LUT:Chl a in chlorophytes, in the Western Equatorial Pacific (data from Higgins & Mackey, 2000 and Higgins et al., 2006, used as marine references) and in the set of freshwater bodies investigated, ranked according their trophic status, increasing from left to right.

9 J.-P. Descy et al. 326 Fig. 4. Variations of ZEA:Chl a in chlorophytes and cyanobacteria T2, in the Western Equatorial Pacific (data from Higgins & Mackey, 2000 and Higgins et al., 2006, used as marine references) and in the set of freshwater bodies investigated, ranked according their trophic status, increasing from left to right. River Meuse (0.38). As lutein:chl a tended to be higher in the eutrophic freshwater bodies, it showed a significant correlation with mean chl a (r ¼ 0.75; p ¼ ) and with mean TP (r ¼ 0.83; p ¼ ). Finally, the ratio of the photoprotective pigment zeaxanthin to chl a (Fig. 4) has a similar behaviour in both groups that possess this pigment in significant amounts, i.e. green algae and cyanobacteria. In green algae, zeaxanthin ratios reached particularly high levels in Lake Tanganyika (0.1), but were much lower in the other water bodies ( ). This ratio was positively correlated with the mean Zm:Zeu (r ¼ 0.83; p ¼ ) in the freshwater bodies, suggesting a response of this pigment ratio to light exposure. However, it is in cyanobacteria T2 that this ratio varied the most, being highest in Lake Tanganyika (maximal in the rainy season, at Zm:Zeu 1) with slightly lower values in Lake Kivu, and declined in the temperate lakes. A significant positive correlation between ZEA:chl a and the mean depth of the euphotic zone was found cyanobacteria T2 (r ¼ 0.77; p ¼ ), as well as negative correlation with mean chl a (r ¼ 0.81; p ¼ ) and mean TP (r ¼ 0.66; p ¼ ). It should be noted that the cyanobacteria T2 in the two tropical lakes were mostly picocyanobacteria, as shown by analysis by epifluorescence and flow cytometry (Sarmento et al., 2008). Although the chl b:chl a ratio (Fig. 5) does not seem to follow any obvious trend, it was significantly correlated with the Zm:Zeu ratio (r ¼ 0.83; p ¼ ). The range was , with the highest values at both ends of the trophic spectrum. Discussion Our results illustrate the trends in the variability of pigment ratios in the set of freshwater bodies studied, and their similarity with ranges observed in culture. Woitke et al. (1996) found a range of and Schlu ter et al. (2006) in the FUCO:chl a ratio in diatom cultures, thus corresponding rather well to the ratio we calculated from our freshwater data ( ). Apart from a light-induced fucoxanthin cycle, which seems to operate in some specific haptophyte species (Stolte et al., 2000; Rodrıguez et al., 2006), fucoxanthin is generally considered as a light-harvesting pigment. Therefore, the cell content of fucoxanthin in diatoms should increase with decreasing light, but the ratio FUCO:chl a may not be affected, owing to a parallel chl a increase. During the 2003 rainy season in Lake Tanganyika, the FUCO:chl a ratio increased with depth

10 Pigment ratios across aquatic environments 327 Fig. 5. Variations of Chl b:chl a in chlorophytes, in the Western Equatorial Pacific (data from Higgins & Mackey, 2000 and Higgins et al., 2006, used as marine references) and in the set of freshwater bodies investigated, ranked according their trophic status, increasing from left to right. (Fig. 1B); in contrast, it was highest in the clearest lakes and in the WEP, and lowest in the turbid River Meuse. A small increase of FUCO:chl a with depth in diatoms (from 0.75 to 0.86) was also noted by Furuya et al. (2003) under summer stratified conditions in the East China Sea. In cultures, increases in the FUCO:chl a ratio in diatoms have been observed in certain coastal (Schlu ter et al., 2000) and freshwater algae (Schlu ter et al., 2006) when grown under lowlight compared with high-light conditions. These results contrast with the observation by Higgins et al. (2006) of a significant decrease with depth of the FUCO:chl a ratio in diatoms from 1.10 to 0.60 in the highly stratified waters of the WEP (Fig. 2). In a study in the stratified Southern Ocean waters near the Antarctic ice edge, Wright & van den Enden (2000) also found high values for the FUCO:chl a ratio in diatoms in surface waters with a significant decrease with depth. A similar decrease in the FUCO:chl a ratio in low light (compared with high light) was found by Lewitus et al. (2005) in culture studies of an estuarine diatom. In addition to irradiance, Goericke & Montoya (1998) indicate in their culture studies of diatoms and prymnesiophytes that nitrate limitation had a variable, and sometimes significant, effect on the growth rates and the FUCO:chl a ratios. As a result, interpreting variations of the FUCO:chl a ratio is not straightforward: it may depend on the relative rate of increase of fucoxanthin vs chl a with decreasing light; culture data suggests a significant intra-class variation (Goericke & Montoya, 1998; Schlu ter et al., 2006) and an effect of nitrogen limitation, presumably on chl a synthesis. This may be the reason why higher ratios were observed in oligotrophic waters than in eutrophic waters, which is consistent with the negative correlation observed between FUCO:chl a and TP. TP, an indicator of trophic status, was highly correlated with DIN or TN concentration (Table 2). DDX:chl a ratios in diatoms from our field data ( ) and from cultures are in a similar range (from 0.05 to up to >0.3 for Asterionella at high light; Schlu ter et al., 2006). Nicklisch & Woitke (1999) reported low values from cultures ( ), which were grown, presumably, at low light. We found the lowest values in the River Meuse and, surprisingly, from the oligotrophic Lake Tanganyika, where higher concentrations of photoprotective pigments in diatoms would have been expected. However, in Lake Tanganyika, diatoms tend to concentrate at the boundary between the mixed layer and the thermocline (Descy et al., 2005), where they are exposed to lower light intensity, and may not need effective photoprotection. Moreover, the most abundant diatom found in Lake Tanganyika during our studies mainly belonged to a colonial species (Nitzschia cf. asterionelloides; Cocquyt & Vyverman, 2005), which may have peculiar pigment composition. In the other oligotrophic lakes, the DDX:chl a ratio was much higher ( ) and approached the values reported for the oligotrophic WEP Ocean ( ). The highest correlations were found with mean chl a (r ¼ 0.71; p ¼ ) and with mean TP (r ¼ 0.86; p ¼ ), suggesting that DDX:chl a decreases with increasing productivity, which

11 J.-P. Descy et al. 328 may result from exposure to lower light in eutrophic water bodies than in the oligotrophic ones. This trend was confirmed by the higher ratio values found for diatoms of the WEP Ocean, and by the decrease with depth shown by Higgins et al. (2006) and in this paper for the diatoms from Lake Kivu (Fig. 1). In cryptophytes, alloxanthin (ALLO) is a lightharvesting pigment (Siefermann-Harms, 1987). The ALLO:chl a ratio varied little in the freshwater bodies studied ( ), still in agreement with the fact that light-gathering pigments co-vary with chl a. Again, these ratios are in a range similar to those reported for cultured freshwater algae ( , Schlu ter et al., 2006) for marine algae ( ; Henriksen et al., 2002); although, on average, ratios in marine algae tend to be slightly lower than for freshwater algae as indicated by the WEP data (Fig. 3). The situation is similar for the ratio of lutein (LUT) to chl a in green algae. The LUT:chl a values obtained from the CHEMTAX runs on the freshwater data sets ( , but a higher value in the River Meuse) were close to the values of Nicklisch & Woitke (1999) for two Scenedesmus cultures ( ) and to the ones of Schlu ter et al. (2006) for three different genera ( ). The range reported by Higgins et al. (2006) from the JGOFS/TROPICS cruises was somewhat lower ( ), but still close to freshwater values. The function of lutein is unclear (Siefermann-Harms, 1987), but our data might indicate a light-harvesting function, as it seems to co-vary with chl a. A photosynthetic role for lutein was suggested by Schmid et al. (1978). The most dramatic variation recorded in this study was for the ratio of the photoprotective pigment zeaxanthin to chl a (ZEA:chl a), in particular in cyanobacteria T2:0.11 in the most eutrophic turbid lake to 0.86 in lakes Tanganyika and Kivu. Yet the maximal values found in the CHEMTAX processing of the tropical lakes data (1.2) is still lower than typical oceanic values (1.9 in the WEP warm pool, according to Higgins et al., 2006). For cyanobacteria T2, comparable values to those observed in the tropical lakes (Fig. 5) were found in Moore et al. (1995) for Synechococcus WH8103, at high growth irradiance. Schlu ter et al. (2006) reported variable zeaxanthin content among five species of cyanobacteria, but again with a similar increase of the ZEA:chl a ratios with increasing light: the range was in Synechococcus, and in Aphanizomenon, thus very close to our range from field data in this study. The values of ZEA:chl a in green algae were lower, but varied by a factor of 10 between the minimum (0.015) and the maximum calculated by CHEMTAX (0.105). Culture results for ZEA:chl a in green algae (Nicklisch & Woitke, 1999; Schlu ter et al., 2006) were again in a similar range ( ) with some variation among species, but the highest ratios were always observed at high light and the lowest at low light. As for DDX:chl a in diatoms, the results were consistent with the expectation that the ratio of photoprotective carotenoids to chl a will show a substantial increase at high light. Finally, as noted above, there was no observed trend in the chl b:chl a ratio. There was, however, a significant correlation with the Zm:Zeu ratio, and a realistic range obtained with the CHEMTAX processing ( ), was shown by the similarity with the range reported in cultures of green algae at different light regimes ( ; Schlu ter et al., 2006). Conclusion Our results show that pigment ratios adjusted after appropriate data processing with CHEMTAX correspond quite satisfactorily with the ratios observed from pure cultures of freshwater algae, particularly those studies where the focus was on the effects of light and nutrients on cell pigment content (Woitke et al., 1996; Nicklisch & Woitke 1999; Schlu ter et al., 2006). This good agreement supports the use of the pigment approach and indicates that ratios provided by CHEMTAX are not merely a numerical result, but have ecophysiological significance. Unfortunately, to date relatively few freshwater ecologists use pigment analysis with HPLC, so that studies on pigment ratios in pure cultures of freshwater algae have been relatively rare, hence few references are available for setting pigment ratio values. The intra-class variation observed in some marine planktonic algae (Zapata et al., 2004) is not as well constrained for freshwater algae (Nicklisch & Woitke, 1999). There is a clear need for additional studies on pigment composition in strains of freshwater algae from different groups, in order to develop a pigment ratio database, similar to that available for marine phytoplankton (Jeffrey et al., 1997), implemented by further analyses in particular groups presenting large chemotaxonomic diversity (Zapata et al., 2004). The discussion of the variation of pigment:chl a ratios with depth in a water column (Fig. 1) indicates some general patterns in the ecophysiological response of photoprotective and light harvesting to changes in irradiance (and nutrients). However, it also highlights the need to identify as accurately as possible (usually by microscopy) the algae actually present within the community as the direction and magnitude of these ecophysiological responses are often species specific (cf. the differences between

12 Pigment ratios across aquatic environments 329 freshwater and coastal marine algae reported by Schlu ter et al., 2006). The large variation of many of these pigment:chl a ratios with depth, particularly in stratified waters, also stresses the need to sub-divide vertically the data sets into more homogenous sub-groups (at least mixed layer and below) to increase the accuracy of the subsequent CHEMTAX calculations. Finally, the observations on substantial pigment ratio variations in the natural environment, in freshwater bodies of different transparency and trophic status, are still inadequate to permit the establishment of reliable pigment ratio matrices, which are necessary for developing standards for the quantitative estimation of algal classes using marker pigments in fresh waters. We recommend, therefore, that more studies be conducted on diagnostic pigment variations in freshwater phytoplankton, including both experiments on cultures and field observations, together with adequate determination of environmental factors. Acknowledgements We are thankful to two anonymous referees that helped improve the manuscript, and to all scientists and technicians who were involved in sample collection and processing, HPLC analysis, and acquisition of various data. Dr Isabelle Thys and Dr Tim Kratz kindly provided environmental data for Lake Eschsur-Suˆre and for the two LTER lakes, respectively. References BRUNET, C., BRYLINSKI, J.M. & LEMOINE, Y. (1993). In situ variations of the xanthophylls diatoxanthin and diadinoxanthin: photoadaptation and relationships with a hydrodynamical system in the eastern English Channel. Mar. Ecol. Progr. Ser., 102: BUCHACA, T., FELIP, M.& CATALAN, J. (2005). A comparison of HPLC analyses and biovolume estimates of phytoplankton groups in an oligotrophic lake. J. Plankton Res., 27: COCQUYT, C. & VYVERMAN, W. (2005). Phytoplankton in Lake Tanganyika: a comparison of community composition and biomass off Kigoma with previous studies 27 years ago. J. Great Lakes Res., 31: COULTER, G.W. (1991). Lake Tanganyika and its Life. Oxford University Press, London, UK. DEMERS, S., ROY, S., GAGNON, R.& VIGNAULT, C. (1991). Rapid light-induced changes in cell fluorescence and in xanthophyllcycle pigments of Alexandrium excavatum (Dinophyceae) and Thalassiosira pseudonana (Bacillariophyceae) a photoprotection mechanism. Mar. Ecol. Prog. Ser., 76: DESCY, J.-P., HARDY, M.-A., STENUITE, S., PIRLOT, S., LEPORCQ, B., KIMIREI, I., SEKADENDE, B., MWAITEGA, S.R. & SINYENZA, D. (2005). Phytoplankton pigments and community composition in Lake Tanganyika. Freshw. Biol., 50: DESCY, J.-P., HIGGINS, H.W., MACKEY, D.J., HURLEY, J.P. & FROST, T.M. (2000). Pigment ratios and phytoplankton assessment in northern Wisconsin lakes. J. Phycol., 36: DESCY, J.-P. & ME TENS, A. (1996). Biomass-pigment relationships in potamoplankton. J. Plankton Res., 18: EDELKRAUT, F., WILTSHIRE, K.H., BLÖCKER, G.W.P. & SCHROEDER, F. (1997). Algal pigments as tracers for biological variability in the Elbe River and estuary. Vom Wasser, 88: FIETZ, S., KOBANOVA, G., IZMEST EVA, L.& NICKLISCH, A. (2005). Regional, vertical and seasonal distribution of phytoplankton and photosynthetic pigments in Lake Baikal. J. Plankton Res., 27: FIETZ, S. & NICKLISCH, A. (2004). An HPLC analysis of the summer phytoplankton assemblage in Lake Baikal. Freshw. Biol., 49: FURUYA, K., HAYASHI, M., YABUSHITA, Y.& ISHIKAWA, A. (2003). Phytoplankton dynamics in the East China Sea in spring and summer as revealed by HPLC-derived pigment signatures. Deep-Sea Res. II, 50: GOERICKE, R. (2002). Bacteriochlorophyll a in the ocean: is anoxygenic bacterial photosynthesis important? Limnol. Oceanogr., 47: GOERICKE,R.&MONTOYA, J.P. (1998). Estimating the contribution of microalgal taxa to chlorophyll a in the field variations of pigments ratios under nutrient and light-limited growth. Mar. Ecol. Prog. Ser., 169: GOSSELAIN, V. (1998). Phytoplancton de la Meuse et de la Moselle et impact du broutage par le zooplancton. PhD thesis, University of Namur, Namur, Belgium. HENRIKSEN, P., RIEMANN, B., KAAS, H., MUNK SORENSEN, H.& LANG SORENSEN, H. (2002). Effects of nutrient-limitation and irradiance on marine phytoplankton pigments. J. Plankton Res., 24: HIGGINS, H.W. & MACKEY, D.J. (2000). Algal class abundances, estimated from chlorophyll and carotenoid pigments, in the western Equatorial Pacific under El Nino and non-el Nino conditions. Deep-Sea Res. I, 47: HIGGINS, H.W., MACKEY, D.J. & CLEMENTSON, L. (2006). Phytoplankton distribution in the Bismarck Sea north of Papua New Guinea: the effect of the Sepik River outflow. Deep Sea Res. I, 53: HURLEY, J.P. & WATRAS, C.J. (1991). Identification of bacteriochlorophylls in lakes via reverse-phase HPLC. Limnol. Oceanogr., 36: JEFFREY, S.W., MANTOURA, R.F.C. & BJØRNLAND, T. (1997). Data for the identification of 47 key phytoplankton pigments. In Phytoplankton Pigments in Oceanography: Guidelines to Modern Methods (Jeffrey, S.W., Mantoura, R.F.C., and Wright, S.W., editors). SCOR-UNESCO, Paris, France. JEFFREY, S.W. & WRIGHT, S.W. (2006). Photosynthetic pigments in Marine Microalgae: insights from cultures and the sea. In Algal Cultures, Analogues of Blooms and Applications (Subba Rao, D.V., editor). Science Publishers, Enfield, USA. LATASA, M. (2007). Improving estimations of phytoplankton class abundances using CHEMTAX. Mar. Ecol.-Prog. Ser., 329: LEWITUS, A.J., WHITE, D.L., TYMOWSKI, R.G., GEESEY, M.E., HYMEL, S.N. & NOBLE, P.A. (2005). Adapting the CHEMTAX method for assessing phytoplankton taxonomic composition in southeastern U.S. estuaries. Estuaries, 28: MACKEY, M.D., MACKEY, D.J., HIGGINS, H.W. & WRIGHT, S.W. (1996). CHEMTAX a program for estimating class abundances from chemical markers: application to HPLC measurements of phytoplankton. Mar. Ecol. Prog. Ser., 144: MOORE, L.R., GOERICKE, R.& CHISHOLM, S. (1995). Comparative physiology of Synechococcus and Prochlorococcus: influence of light and temperature on growth, pigments, fluorescence and absorptive properties. Mar. Ecol. Prog. Ser., 116: MUYLAERT, K., GONZALES, R., FRANCK, M., LIONARD, M., VAN DER ZEE, C., CATTRIJSSE, A., SABBE, K., CHOU, L. & VYVERMAN, W. (2006). Spatial variation in phytoplankton dynamics in the Belgian coastal zone of the North Sea studied by microscopy, HPLC-CHEMTAX and underway fluorescence recordings. J. Sea Res., 55:

13 J.-P. Descy et al. 330 NICKLISCH, A.& WOITKE, P. (1999). Pigment content of selected planktonic algae in response to simulated natural light fluctuations and a short photoperiod. Int. Rev. Hydrobiol., 84: PANDOLFINI, E., THYS, I., LEPORCQ, B. & DESCY, J.-P. (2000). Grazing experiments with two freshwater zooplankters: fate of chlorophyll and carotenoid pigments. J. Plankton Res., 22: RIEGMAN, R.& KRAAY, G.W. (2001). Phytoplankton community structure derived from HPLC analysis of pigments in the Faroe- Shetland Channel during summer 1999: the distribution of taxonomic groups in relation to physical/chemical conditions in the photic zone. J. Plankton Res., 23: RODR_IGUEZ, F., CHAUTON, M., JOHNSEN, G., ANDRESEN, K., OLSEN, L.M. & ZAPATA, M. (2006). Photoacclimation in phytoplankton: implications for biomass estimates, pigment functionality and chemotaxonomy. Mar. Biol., 148: SARMENTO, H., ISUMBISHO, M. & DESCY, J.-P. (2006). Phytoplankton ecology of Lake Kivu (Eastern Africa). J. Plankton Res., 28: SARMENTO, H., LEITAO, M., STOYNEVA, M., COUTE, A., COMPE` RE, P., ISUMBISHO, M.& DESCY, J.P. (2007). Diversity of pelagic algae of Lake Kivu (East Africa). Cryptogamie Algol., 28: 245:269. SARMENTO, H., UNREIN, F., ISUMBISHO, M., STENUITE, S., GASOL, J.M. & DESCY, J.P. (2008). Abundance and distribution of picoplankton in tropical, oligotrophic Lake Kivu, eastern Africa. Freshw. Biol., 53: SCHLU TER, L., LAURIDSEN, T.L., KROGH, G. & JORGENSEN, T. (2006). Identification and quantification of phytoplankton groups in lakes using new pigment ratios a comparison between pigment analysis by HPLC and microscopy. Freshw. Biol., 51: SCHLU TER, L.F., MOHLENBERG, F., HAVSKUM, H. & LARSEN, S. (2000). The use of phytoplankton pigments for identifying and quantifying phytoplankton groups in coastal areas: testing the influence of light and nutrients on pigment/chlorophyll a ratios. Mar. Ecol. Prog. Ser., 192: SCHMID, G.H., RADUNZ, A., KOENIG, F. & MENKE, W. (1978). Characterization of the oxygen-evolving side of photosystem II by antisera to polypeptides and carotenoids. In Photosynthetic Oxygen Evolution (Metzner, H., editor). Academic Press, London, UK. SIEFERMANN-HARMS, D. (1987). The light-harvesting and photoprotective functions of carotenoids in photosynthetic membranes. Physiol. Plantarum, 69: STEENBERGEN, C.L.M., KORTHALS, H.J., BAKER, A.L. & WATRAS, C.J. (1989). Microscale vertical distribution of algal and bacterial plankton in Lake Vechten (The Netherlands). FEMS Microbiol. Ecol., 62: STOLTE, W., KRAAY, G.W., NOORDELOOS, A.A.M. & RIEGMAN, R. (2000). Genetic and physiological variation in pigment composition of Emiliania huxleyi (Prymnesiophyceae) and the potential use of its pigment ratios as a quantitative physiological marker. J. Phycol., 36: THYS, I., LEPORCQ, B.& DESCY, J.-P. (2003). Seasonal shifts in phytoplankton ingestion by Daphnia galeata, assessed by analysis of marker pigments. J. Plankton Res., 25: VELDHUIS, M.J.W. & KRAAY, G.W. (2004). Phytoplankton in the subtropical Atlantic Ocean: towards a better assessment of biomass and composition. Deep Sea Res. I, 51: VILA, X., DOKULIL, M., GARCIA-GIL, J., ABELLA, C.A., BORREGO, C.M. & BANERAS, L. (1996). Composition and distribution of photrophic bacterioplankton in the deep communities of several central European lakes: The role of light quality. Arch. Hydrobiol. Ergebn. Limnol., 48: VIROUX, L. (2000). Dynamique du Me tazooplancton en Milieu Fluvial. PhD Thesis, University of Namur, Namur, Belgium. WOITKE, P., SCHIWITZ, T., TEUBNER, K. & KOHL, J.-G. (1996). Annual profiles of photosynthetic lipophilic pigments in four freshwater lakes in relation to phytoplankton counts as well as to nutrient data. Arch. Hydrobiol., 137: WRIGHT, S.W. & JEFFREY, S.W. (2006). Pigment markers for phytoplankton production. In Marine Organic Mattter: Biomarkers, Isotopes and DNA (Volkman, J.K., editor). Handbook of Environmental Chemistry 2, Part N, Springer, Berlin, Germany. WRIGHT, S.W., JEFFREY, S.W., MANTOURA, R.F.C., LLEWELLYN, C.A., BJORNLAND, T., REPETA, D. & WELSCHMEYER, N. (1991). Improved HPLC method for the analysis of chlorophylls and carotenoids from marine phytoplankton. Mar. Ecol. Prog. Ser., 77: WRIGHT, S.W. & VAN DEN ENDEN, R.L. (2000). Phytoplankton populations off east Antarctica in relation to stratification/ mixing regimes: CHEMTAX analysis of HPLC pigment profiles (BROKE survey, Jan Mar 1996). Deep-Sea Res. II, 47: YACOBI, Y.Z., POLLINGHER, U., GONEN, Y., GERHARDT, V. & SUKENIK, A. (1996). HPLC analysis of phytoplankton pigments from Lake Kinneret with special reference to the bloom-forming dinoflagellate Peridinium gatunense (Dinophyceae) and chlorophyll degradation products. J. Plankton Res., 18: ZAPATA, M., JEFFREY, S.W., WRIGHT, S.W., RODRIGUEZ, F., GARRIDO, J.L. & CLEMENTSON, L. (2004). Photosynthetic pigments in 37 species (65 strains) of Haptophyta: implications for oceanography and chemotaxonomy. Mar. Ecol. Prog. Ser., 270: