Influence of zooplankton stoichiometry on nutrient sedimentation in a lake system

We explored rates and stoichiometry (C : N : P ratios) of sinking particles in a temperate reservoir during a 2-yr period. Plankton was sampled weekly, and a sediment trap placed below the metalimnion collected sinking particles. There were no significant relationships between the stoichiometry of entrapped material and seston or zooplankton stoichiometry. However the differences in the entrapped C : P and N : P ratios between consecutive trap samplings were negatively correlated with the time variations of the zooplankton C : P and N : P ratios. Zooplankton C : P and N : P ratios were positively correlated with the percentage of copepod biomass in total zooplankton biomass .250 mm and negatively correlated with the percentage of cladocerans. Zooplankton biomass .250 mm reduced the fraction of N and P primary production lost to sinking (export ratio). The residuals of the N export ratio versus zooplankton biomass relationship were negatively correlated with the zooplankton N : P ratio, whereas there was a positive relationship with the residuals of the P export ratio relationship. These observations support the hypothesis that the regulation of elemental homeostasis in the herbivorous zooplankton consumers occurs at least partly at the assimilation/egestion level. Elements ingested in excess—P for the herbivorous copepods and N for many cladocerans—are concentrated into sinking feces, whereas the deficient elements are captured into biomass. In lakes, the vertical flux of small particulate matter essentially comprises two components: the sinking of ungrazed phytoplankton cells (e.g., Reynolds et al. 1982) and the sinking of the feces or fecal pellets of planktonic primary consumers (Sarnelle 1999). Both types of particles will settle from the upper layer below the thermocline if their sinking rates are higher than their mineralization rates. A simple heuristic model (Elser et al. 1995) explicitly represents the processes involved (Eq. 1). Sx 5 rz,xexgPx 1 rxsx(1 2 g)Px (1) 1 To whom correspondence should be addressed. Present address: Université du Québec à Trois-Rivières, GREA, CP 500, Trois-Rivières, Québec, Canada G9A 5H7 ([email protected]). Acknowledgments We thank Véronique Jacquet and Raphaël Willame for field assistance, and we thank Dag Hessen, Gérard Lacroix and two anonymous reviewers for helpful comments on earlier drafts of this manuscript. This work was supported by the National Fund for Scientific Research (Belgium). Sx is the sedimentation rate of the element x; rz,x is the fraction refractory to mineralization of the egested (ex), grazed (g) fraction of the production rate of element x (Px), and rx is the corresponding refractory fraction for the sinking (sx), ungrazed (1 2 g) fraction of the production rate of element x. For any given time interval, we can express the sedimentation rate as a fraction of elements incorporated by autotrophic activity during that interval. This fraction is referred to as the ‘‘export ratio’’ (ER; Eppley and Peterson 1979) and can be defined for any element x (ERx; Elser et al. 1995). 21 ER 5 (S P ) 5 r e g 1 r s (1 2 g) x x x z,x x x x ⇔ ER 5 r s 1 g(r e 2 r s ) (2) x x x z,x x x x This expression defines a relationship between export ratio and grazing intensity (g) in which ERx is a linear function of g with a y-intercept of rxsx and a slope of rz,xex 2 rxsx. The slope of this formula explicitly formalizes the influence of zooplankton on sedimentation of autotrophic production. Zooplankton, by its grazing activity, enhances the export ratio if the fraction of feces refractory to mineralization is 906 Darchambeau et al. Fig. 1. Location of the Esch-sur-Sûre reservoir with morphometric and ecological summary information. The location of the sampling and sediment trap deployment station is indicated by S. higher than the refractory fraction of sinking phytoplankton cells (rz,xex . rxsx). However, in turn, zooplankton will decrease the export ratio if rz,xex , rxsx. The direction of the relationship between zooplankton (via its grazing) and sedimentation of particulate matter is thus dependent on (1) lake morphometry, which determines the fraction of phytoplankton that directly sediments (sx), (2) phytoplankton community characteristics (via rx and sx), and (3) zooplankton community characteristics (via rz,x and ex; Elser et al. 1995). The egested fraction of an element x (ex) is a function of the digestive ability of animals. Stoichiometric theory indicates that it could also be a function of consumer elemental needs (Sterner and Elser 2002). P-rich consumers, such as daphniids (Andersen and Hessen 1991; Hessen and Lyche 1991), have high demands of P from food; thus, the fraction of P egested from Daphnia must be lower than for genera with a lower P demand. In turn, copepods are rich in body nitrogen (N), and the fraction of N egested should be lower than for zooplankton species with a lower body N content. Moreover, apart from zooplankton stoichiometry, the fraction of feces refractory to mineralization (rz,x) also depends on their structure. Copepods surround their feces with a peritrophic membrane, which largely increases fecal cohesion and therefore the probability of sinking out of the epilimnion. So we can predict that copepods are likely to increase the export ratio of P and decrease the export ratio of N. However, for daphniid species, their effect on nutrient sedimentation will depend on the cohesion of their feces. If the feces sink before remineralization, we can predict a positive effect of daphniids on N sedimentation and a negative effect on P sedimentation. In this study, we monitored the elemental contents of seston, zooplankton, and settling particles in a mesoeutrophic lake during the annual period of stratification for 2 yr. We determined sinking of elements as a fraction of primary production to test the effects of zooplankton stoichiometry on vertical particulate-bound nutrient exports. Our results indicate that zooplankton biomass has a positive effect on N and P retention in the epilimnion and that zooplankton stoichiometry determines stoichiometry of settling particles. Materials and methods Field data acquisition—This field study was conducted in the Esch-sur-Sûre reservoir in Grand-Duchy of Luxembourg. A map and a summary of morphometric and ecological characteristics are provided in Fig. 1. According to the Organization for Economic Cooperation and Development (1982) classification, the reservoir is considered a mesoeutrophic waterbody (Dohet and Hoffmann 1995). The survey was conducted at a station (maximum depth 30 m) located in the middle of the lake, representative of whole lake conditions (Thys et al. 1998). Seston and zooplankton were sampled weekly during the period of stratification (roughly from April to October) in 1999 and 2000. Lake stratification was determined weekly according to the temperature and oxygen vertical profiles obtained with a Hydrolab DS-4 multiprobe. The lower limit of the epilimnion was at 2–3 m in early May and continuously deepened until it was at 11–14 m in late September. The lower limit of the metalimnion started at 7–8 m in May and deepened to 17–20 m in September. Zooplankton was sampled with a 17-cm diameter, 250-mm mesh net towed vertically in the epilimnion. Six zooplankton samples for elemental analysis were collected on each sampling occasion and immediately filtered on preignited (12 h at 5008C), preweighted Whatman GF/C filters and directly frozen in dry ice. Epilimnetic zooplankton was additionally sampled with a 50-cm diameter, 50-mm mesh net to determine community composition and densities. Triplicate samples were collected and pooled to reduce heterogeneity in zooplankton horizontal distribution and sampling variability. The collected zooplankton was immediately narcotized in soda water, rinsed, and preserved within 4% formalin (Haney and Hall 1973). For seston analysis, a pooled sample was constituted on each sampling occasion from discrete samples (3 liters) collected with a Ruttner bottle and spaced every meter in the epilimnion. Another pool was constituted for the metalimnion. From each pool, one subsample of 1–2.5 liters were filtered on a Whatman GF/C filter and directly frozen in liquid nitrogen for pigment analysis by high-performance liquid chromatography (HPLC), and six subsamples of 0.15–0.5 liters were filtered on preignited Whatman GF/C filters and directly frozen in dry ice for elemental analysis. A sediment trap was deployed at the top of the hypolimnion, and its deployment depth (8–21 m) was adjusted every 2 weeks according to the thermal and oxygen stratification of the water column. The trap consisted of one 15.4-cmdiameter, 133-cm-long polyvinyl chloride collection tube suspended from a floating pontoon. The trap was initially filled with GF/C-filtered lake water taken at the immersion depth of the trap. One liter of an inhibitory, high-density solution (180 mmol L21 HgCl2, 10% w/w NaCl) was added with the use of a small tube at the bottom of the trap. This solution inhibits the breakdown of entrapped material without catching swimmers (Lee et al. 1992; data not shown). The trap was recovered every 2 weeks, and the upper 4 liters were discarded. The remaining volume (;20 liters) was carefully poured into a large basin and gently mixed. The 907 Zooplankton and nutrient sedimentation water was filtered through a 250-mm Nitex screen, and six subsamples of particles ,250 mm were collected for elemental analysis on preignited Whatman GF/C filters. In addition, lake water was collected with a Ruttner bottle at the same depth as the trap immersion depth and submitted to the same procedures to correct trap contents for ambient par-