Alti Associazione Italiana Oceanologia Limnologia, 14, pp. 25-35, luglio 2001

LONG-TERM CHANGES lN THE VERTICAL DISTRIBUTION OF PHYTOPLANKTON BlOMASS AND PRIMARY PRODUCTION lN LAKE GENEV A: A RESPONSE TO THE OLIGOTROPHICA TION

Orlane ANNEVILLE, Christophe LEBOULANGER Station d'Hydrobiologie Lacustre, INRA, BP 511, 74203 Thonon les Bains cedex, FRANCE

ABSTRACT ln Lake Geneva, after several years of eutrophication dissolved inorganic phosphorus concentrations was reduced from 76 Jlg.r' to around 30 Jlg,r', This decrease lead to changes in the phytoplankton composition, but no decrease in the phytoplankton crop has been observed. An analysis of the changes in the vertical distribution of chlorophyll a and primary production reveals a long-term tendancy towards the deepening of the maximum and a more homogeneous distribution of the primary production within the photic layer. This evolution is coherent with the one expected for a lake in course of oligotrophication. Even though the transparency controls the maximum depth of phytoplankton growth, it seems that the observed changes in the vertical distribution of phytoplankton are more likely to be induced by the evolution in the phosphorus vertical gradient. We conclude that the general decrease in phosphorus concentrations induced a change in the vertical distribution of phytoplankton, and altered the summer phytoplankton community toward the selection of species adapted to low-light intensities. RlASSUNTO Modificazioni pluriennali della distribuzione verticale della biomassa algale e della produzione primaria nel Lago di Ginevra in risposta ad un processo di oligotrofizzazione Nel Lago di Ginevra, do po diversi anni di peggioramento delle condizioni trofiche, la concentrazione dei fosforo totale si è ridotta da 76 a circa 30 Jlg ri, Questa diminuzione ha indotto dei cambiamenti nella composizionc dei fitoplancton, sebbene non sia stata osservata una riduzione della biomassa algale. Un'analisi dei cambiamenti nelle distribuzione verticale della clorofilla a e della produzione primaria rivela J'esistenza di una tendenza a lungo termine verso 10 sprofondamcnto nella colonna d'acqua dei valori massimi ed una distribuzione verticale più omogenea della produzione al gale nello stato fotico. Questa evoluzione è coerente con quella che ci potrebbe attendere in un lago in corso di oligotrofizzazione. Sebbene la trasparenza controlli la profondità

25

massima della crescita fitoplanctonica, sembra che i cambiamenti osservati nella distribuzione verticale dei fitoplancton siano da attribuire, più probabilmente, alla evoluzione nel gradiente verticale di fosforo. Possiamo concludere che il generale decremento della concentrazione di fosforo abbia indotto una variazione nella distribuzione verticale dei fitoplancton e modificato la comunità algale estiva selezionando specie adatte a basse intensità luminose. 1. INTRODUCTION The long-term monitoring of Lake Geneva provides fruitful information about the oligotrophication process. ln the sixties, bec au se of the increase of human activity, this lake went through eutrophication. ln the early eighties, annual total phosphorus (TP) started to respond to the reduction in P-Ioading, decreasing from a maximum of 90 J.lg p.rl in the 1970's to 40 J.lg p.rl in 1998. An unexpected increase in phytoplankton biomass was observed in the recent years (Anneville & Pelletier, 2000), characterised by the summer development of filamentous algae such as Mougeotia gracillima (Anneville et al., submitted). Changes in the vertical phosphorus distribution has recently been hypothesised to be one of the key factors responsible for such a phytoplankton response observed since 1992 (Anneville et al., submitted). However, analyses were restricted to the first upper 10 meters for which data on the phytoplankton composition were available. ln this paper we propose to consider a third dimension to this long-term approach by analysing inter-annual changes in the vertical distribution of phytoplankton biomass and activity, in response to change in phosphorus concentration and vertical distribution. Such an approach should provide further clues for a better understanding of the oligotrophication process in lake Geneva, by placing its conceptual scheme within the vertical dimension of the water column. Spatial heterogeneity of plankton distribution has been reported on many occasions and plankton patterns in both the vertical and horizontal directions, have been a point of interest and discussion for several decades (George & Edwards 1976; George & Heaney, 1978; Frempong, 1981; Venrick, 1984; Steele & Henderson, 1992; Abraham, 1998). During the summer, deep temperate lakes are vertically stratified and separated into three strata. N utrient levels are generally low in the epilimnion due to phytoplankton consumption, and increase with depth. Thus, thermal stratification creates a niche which does not exit earlier in the year: the metalimnion, which offers high nutrients but low light and temperature. As a consequence, maximum algal biomass occurs in either the epilimnion or metalimnion, depending on water transparency or primary production levels (Moll & Stoermer, 1982). Tt is assumed that when light intensities are sufficient for al gaI growth, metalimnentic populations exploit the high nutrient levels at the thermocline. However, with increasing nutrient concentration in the epilimnion, epilimnetic algal populations may become large enough to de crea se the amount of light reaching the metalimnion, and to shade out metalimnetic algae. George & Heaney (1978) suggested that as a result of self-shading effects, there should be a tendency towards a larger maxima to form nearer the surface with increasing population density. Thus, during the stratified period, eutrophication or

26

oligotrophication process are expected to induce changes in the vertical distribution of phytoplankton via the selt:shading process. This paper presents a field study showing changes in vertical chlorophyll (1 and primary production distribution over a 25 years period marked by changes in the lake trophic status. The objective is to analyse thc long-term changes in vertical phytoplankton distribution in relation with changes in dissolved inorganic phosphorus (DIP) concentrations and to place the se observations within the conceptual scheme issued l'rom the previous investigations on this ecosystem. 2. METHODS 2.1. Study site and data sets Lake Geneva is a deep and monomictic lake (72 km long, 14 km width, maximal depth of 309 m) located between France and Switzerland. Its total volume is about 89 km1 and mean retention time of the water is around II years. At the end of the 50's total phosphorus increased; this phenomenon was accompanied by a decrease in transparency, a decrease in the recruitment of cxploited Whilefishes, and the occurrence of phytoplankton blooms during summer. Given the economical importance of the Lake, protection measures were soon decided in order to reduce its phosphorus loading and a monitoring program was deveIoped. This program, first initiated in the sixties, has been supervised by the CIPEL (International Committee for the Protection of Lake Geneva waters) since 1974. The physico-chemical and phytoplanktonological data-sets were constituted at the Station d'Hydrobiologie Lacustre of Thonon les Bains by respcctively P. Blanc, J.-c. Druart and J.P. Pelletier. 2.2. Sampling and analytical methods Samples have been collected at the SHL2 station situated at the deepest point of the Great Lake (Fig. 1). Sampling were performed once a month l'rom 1974 to 1980, then twice a month. Vertical series of water samples for Chlorophyll a analyses were taken at nine discrcte depths (0, 1,2, 3.5, 5, 7.5, 10, 15,20 and 30 m). Samples for physico-chemical analyses were collected at several depths l'rom the surface to the lake bottom (309 m). Nutrient and chlorophyll a analyses methods are detailed in one of the annual reports (Monod el (1/., 1984). The euphotic layer was estimated as to 2.5 times the Secchi depth. To determine long-term trends (1974-1998) in dissolved inorganic phosphorus (DlP), were used only averaged data, collected l'rom 0 to 10 m during the end of winter, which represents the period between the winter redistribution of nutrients and the onset of spring algal growth. Vertical changes in DIP concentrations were investigated by the long-term analyse of the DIP-depleted layer thickness. The DlP-depleted layer is defined as the water layer where phosphorus is reduced to concentrations likely to become critical for algal growth. We have considered a threshold of 10 I1g.r' (Sas, 1989). To characterise the shape of thc chlorophyll a vertical distribution, we have considered ils depth of maximal concentration (ChI a maximum), and the depth at which 50% of the total chI (1 concentration per lake surface unit were reached. Differences

27

between samples were tested using the sum-rang nonparametric Mann- WhitneyWilcoxon test. Primary production was determined using the 14C method according to SteemanNielsen (1952), and results were extrapolated from incubation time to the whole daylight period according to Pelletier (1983). Measurements were made by sampling water at ni ne different depths from 0 to 30 m, samples which were spiked with HI4C03(approx. 2 /-lCi, specifie activity of 0.067 /-lei /-lmOr1, in 100 ml of lake water) and replaced at the same depth. After the incubation period (about the third of the day length across solar no on) samples were filtered onto polycarbonate filters, subjected to HCI fumes, and the remaining radioactivity was determined by liquid scintillation counting.

ubonn e

\ \ \ \ Uranse ./ \/ \/ /

~ \ ... - - - Upper Basin - - - ..

//

/

---~--~ Lower Basin 1

10km J

Fig. 1 - Location of sampling station in Lake Geneva. The lake is divided into two parts, the upper and lower basin. Main ri vers are figured, as the surrounding cities. Samplig during the survey was performed at station SHL2, which is the deepest point of the lake (309 m).

3. RESUL TS 3.1. Long-term changes in DIP concentrations in the upper 10 m DIP concentrations increased until the end of the 1970's and reached maximum values of76 J..lg P.L-1 in 1981. From 1981 to 1992 the DIP concentrations presented a strong decrease. Since 1992 the concentrations hardly decreased, and have been oscillating between 30 J..lg P.L-1 and 20 J..lg p.rl. At the annual scale, the decrease in phosphorus concentrations is characterised by longer phosphorus-depleted periods (Fig. 2), but the long-term evolution of the DIPdepleted layer is not regular over the series and may allow the distinction of three periods: The first period, lasting between 1974 and 1985, corresponds to the maximal phosphorus concentrations. During this period, the DIP-depleted layer starts to develop in June and extends till 10 to 15 m depth between June and October. After 1983, it

28

reaches sporadically 20 m depth. During the second period, from 1986 to 1990, the DIP-depleted layer developed earlier in May, and during late spring oscillated between 5 and 10 m. ln summer, it used to extend between 15 and 20 m depth and, sometimes, reached more than 25 m. During the last period, from 1991 to 1998, this depleted layer appeared in spring. Since the month of April, it usually extends to 5 m depth. The DIP depleted layer often extended deeper than 10 m in May and more than 20 m in summer, reaching then frequently a maximum of 30 m. During this third period, the summer DIP-depleted layer has usually been thicker than the mixing layer and the euphotic zone.

80

40 60

20 10

74

o

o 5

5

10

month

10

month

Fig. 2 - Evolution of mean DIP in the 10 m below the surface (lef! figure), and thickness of the DIP depleted layer (right figure), depending on the year (vertical axis) and the month (horizontal axis) of sampling.

3.2. Long-term trend in the shape of the vertical chlorophyll a distributions Month by month analyses of inter-annual changes in chlorophyll a vertical distribution underlined different trends. ln May, chi a maxima were located within the upper part of the euphotic layer, between the surface and 5 m depth (Fig. 3a). Since 1990, they have been occasionally

29

Fig. 3 - Box plot of chI a maximum depth (left column) and euphotic layer thickness (right column). Values are divided in two groups, the first one covering the 1974-88 period, the second one covering the years 1989-98.

30

found deeper in the water column, up to 15 m depth, but were stilllocated within the euphotic layer. During the 1990's, the euphotic layer exhibits a deepening trend but no significant differences between the two periods can be detected (Fig. 3b). The separation between the two periods became more pronounced in June (Fig. 3a). The chI a maxima first located at 0 and 5 m depth, occured then between 5 and 10 m since 1989. Moreover, these two periods presented a significant difference in the chI a maxima depths (p=O.OOOI), but no detectable difference for the limits of the euphotic layer (Fig. 3b) which oscillate without any significant trend. ln July, the difference between the two periods is still significant (p