Differences in chemical composition among zooplankton and their food sources can have important consequences for nutrient cycling in lakes. To date, studies of the interaction between plankton stoichiometry and nutrient cycling have focused primarily on temperate lakes. Here we explore the implications of plankton stoichiometry for nutrient cycling in tropical Lake Malawi. Three seston size fractions (picoplankton, nanoplankton and net plankton) displayed seasonal variations in C:P, C:N and N:P ratios. On average, picoplankton displayed lower C:P, C:N, and N:P ratios than nanoplankton and net plankton. Average C:N and C:P ratios for all seston size classes were much higher than Redfield ratios (C:N:P = 106:16:1). The elemental composition of individual species of crustacean zooplankton showed little temporal variation. However, there were large inter-species differences in P and N content, which can be expected to result in differences in N:P recycling ratios. The zooplankton community of Lake Malawi is dominated by copepods (>80% of total zooplankton biomass), which have a relatively low P content and high N:P ratio, resulting in a relatively low N:P recycling ratio. This might be expected to promote N limitation of phytoplankton and dominance of N-fixing cyanobacteria, but in some seasons the effects of zooplankton nutrient cycling appears to be masked by nutrient inputs from rivers, the atmosphere and vertical mixing.
Allochthonous nutrient inputs regulate net ecosystem productivity (Polis et al., 1997), but ecologists have also recognized the vital role that biota play in the cycling of nutrients in marine and freshwater ecosystems (Sirotnak and Huntly, 2000; Vanni, 2002). Most importantly, a significant portion of the nutrients that support primary production in lakes is derived from excretion by zooplankton (Sterner et al., 1992). Considerable attention has focused on nitrogen (N) and phosphorus (P) since they are the nutrients that most commonly limit primary production (Elser and Urabe, 1999). Nutrient release rates are affected by the nutrient composition of a consumer's body relative to that of its food source (Sterner et al., 1992). In general, consumers assimilate a large fraction of the consumed limited nutrient into their tissues and excrete a large fraction of nutrients that are available in excess of their body requirements (Olsen et al., 1996). Preferential assimilation of a specific nutrient by a consumer will enhance limitation of primary production by that nutrient, as consumers will recycle less of the scarce nutrients and more of the non-limiting nutrients (DeMott et al., 1998). An individual with a relatively low body N:P ratio should release nutrients at a N:P ratio higher than that for an individual with a high body N:P ratio when feeding on the same food source (Andersen and Hessen, 1991). Since various zooplankton taxonomic groups have different elemental compositions (Andersen and Hessen, 1991; Sterner et al., 1992), it follows that their recycling potential for N and P will be dissimilar, and so changes in zooplankton species dominance has the potential to alter phytoplankton nutrient status and switch between N and P limitation (Sterner et al., 1992). Methods of estimating nutrient release rates from animals are available (Olsen and Østgaard, 1985). However, quantification of P release rates from natural zooplankton populations is difficult, especially in P-deficient systems, where released P is rapidly taken up by P deficient phytoplankton and bacteria (Elser and Urabe, 1999). An alternative is the application of stoichiometric methods, which use stoichiometric differences among zooplankton and phytoplankton to infer the ecological implications of consumer-driven nutrient cycling (Andersen and Hessen, 1991; Elser and Urabe, 1999; Walve and Larsson, 1999).
Lying at the southern end of the East African Rift valley, Lake Malawi is permanently stratified at ∼230 m (max. depth = 700 m). Dissolved N and P concentrations in the surface mixed layer (SML) are very low, but increase with depth (Hamblin et al., 2003). Phytoplankton production in the lake is seasonal (Patterson et al. 2000) and appears to fluctuate between nitrogen and phosphorus limitation (Hecky et al., 1993; Guildford and Hecky, 2000). The open water zooplankton community of the lake is dominated by two species of calanoids Tropodiaptomus cunningtoni and Thermodiaptomus mixtus, and two species of cyclopoids Mesocyclops aequatorialis aequatorialis and Thermocyclops neglectus, with occasional biomass peaks of two cladoceran species, Diaphanosoma excisum, and Bosmina longirostris, and larvae of the midge Chaoborus edulis (Twombly, 1983). Few direct measurements of nutrient cycling within Lake Malawi and the other African Great Lakes exist. Initial measurements of nutrient cycling by fish were conducted by Emidio et al. (2003) who determined that nearshore fish provide nearly half of the total periphyton N and P demands. The goal of the present study was to assess stoichiometric differences among major zooplankton taxa and seston size classes in Lake Malawi, and to use these measurements to determine how seston stoichiometry may influence zooplankton nutrient recycling.
Materials and methods
Samples for measurement of abundance and elemental composition of picoplankton, nanoplankton, net plankton and zooplankton species were collected at least once per month over an annual cycle from a 90 m deep station in Lake Malawi, at the northern end of the southeast arm (14.059S and 034.952E). Bulk water samples were collected and mixed from depths of 5, 15, 20, 25, 30 and 40 m (CTD casts indicated the chlorophyll fluorescence maximum was usually between 15 and 30 m) using a Niskin sampler. At the laboratory, sequential filtration was used to separate each sample into three size fractions: net plankton (20–100 μm) nanoplankton (2.0–20 μm) and picoplankton (<2.0 μm) (see Ngochera and Bootsma , for details). Chlorophyll a measurements and the analyses of particulate C, N and P were carried out according to the methods described by Stainton et al. (1977). Molar units are used for all C:N:P ratios unless otherwise noted. Zooplankton samples were collected using an 80 µm-mesh Wisconsin net (30 cm diameter) pulled at approximately 1 m s−1 from bottom to the surface. Zooplankton samples were sorted and picked immediately after collection under a dissecting microscope. Approximately 100 individuals of the most common zooplankton species T. cunningtoni, M. a. aequatorialis, T. neglectus, D. excisum and B. longirostris and the lake fly larvae C. edulis were picked, transferred onto pre-combusted GF/F filters and dried in a desiccator.
where L is the maximum accumulation efficiency of the nutrient that is in shortest relative supply (i.e. the fraction of grazed N or P that is allocated to growth). We applied the values of 0.3, 0.5 and 0.75 for L (Walve and Larsson, 1999). A similar range of values was used by Sterner (1990). Comparisons of elemental composition among seston size classes and zooplankton species were made using one way ANOVA.
Total algal biomass estimated as chlorophyll a varied seasonally, from 0.4 to 1.4 µg l−1, with a mean of 0.78 ± 0.30 µg l−1 while total zooplankton biomass fluctuated between 16 and 36 mg C m−3, with a mean of 24 mg C m−3 (Figure 1a). Copepods dominated the total macrozooplankton biomass contributing more than 80%. The calanoid T. cunningtoni dominated the total zooplankton biomass (42%) followed by M. a. aequatorialis (16%), and nauplii (4 to 28% with the lowest contribution occurring in January). Contribution by the two cladocerans, D. excisum and B. longirostris was minimal.
To determine the effect of elemental seston composition on zooplankton nutrient recycling, C:N:P ratios of zooplankton and seston were compared (Figure 1b). Seston appearing to the left of the dashed lines in have higher N:P ratios while those on the right have lower N:P ratios than is required for zooplankton balanced growth. The N:P ratios reported here are means of the different zooplankton species and phytoplankton size classes over the entire sampling period. The recycled N:P ratio is lower for copepods than for cladocerans and the midge C. edulis that have higher and lower body N:P ratio respectively (Table 1). A comparison of the stoichiometry of the different zooplankton groups indicates high C:P ratios for the cyclopoid copepods (M. a. aequatorialis and T. neglectus), while the midge, C. edulis showed the lowest C:P ratio with intermediates observed in D. excisum and T. cunningtoni (Figure 1c). The differences in stoichiometry among the zooplankton species groups are similar to those reported in the literature for cladocerans and copepods (Andersen and Hessen, 1991; Figure 1d).
Seston P concentrations increased from low values before and during hot stratified season to maximal values during the rainy and mixing period (Figure 2a). In contrast, total seston C and N concentrations varied minimally but were lowest during the transition period (Figure 2a). All the three phytoplankton size classes showed C:P, N:P and C:N ratios that were much higher than the Redfield ratio, except for C:P and C:N for picoplankton during the mixing period (Figure 2a and c). One-way ANOVA revealed significant differences among the C:P ratios of phytoplankton size classes. Post hoc t-tests indicated that the C:P ratio of picoplankton was significantly lower than those of nanoplankton and net plankton (p = 0.0004 and p = 0.000093, respectively) while no significant difference was observed between nanoplankton and net plankton (p = 0.09). N:P comparisons between picoplankton and nanoplankton and picoplankton and net plankton also displayed significant differences (p = 0.002 and p = 0.00014, respectively). However, there was no significant difference between the N:P ratios for nanoplankton and net plankton (p = 0.10, Figure 2d). Zooplankton P concentration increased from low values during the transition, hot stratified and at the onset of the rainy season (Figure 2e). In contrast, zooplankton C and N concentration decreased from high values in the transition and hot stratified season to low numbers at the onset of the rainy season, before increasing again in the mixing season (Figure 2e). All zooplankton species had higher C:P and N:P ratios in the hot, stratified season, and lower ratios between February and May (Figure 2f and h). A one-way ANOVA indicated significant variation among species for both the C:P and N:P ratios, (p = 0.003 and 0.04, respectively) but there was no significant inter-species difference in C:N ratios (p = 0.38). C:N ratios were confined to a relatively narrow range (Figure 2g) with an average of 7.2 ± 1.8, in contrast to the wider ranges observed for phytoplankton (Figure 2c).
Our results reveal significant differences in C:N:P ratios between picoplankton and the other two seston size classes. While there were variations in the temporal trends of each size class, all three size classes tended to have lower C:P and C:N ratios following the rainy season through the mixing season, suggesting that they are all regulated by similar forcing mechanisms. The mean C:P, C:N, and N:P ratios for phytoplankton in Lake Malawi are considerably higher than the Redfield ratio of 106:1, 6.6:1 and 16:1 (Redfield, 1934; Sterner, 1995). The low C:P and C:N ratios during the mixing season and the rainy season, and the high ratios during the hot stratified season correspond to earlier studies that observed high and low photosynthetic rates during these seasons respectively (Bootsma, 1993a). As indicators of algal P status, the high C:P and N:P ratios for nanoplankton and net plankton during the stratified season signify low to severe P deficiency during that period (Healey and Hendzel, 1980). That picoplankton displayed lower C:P and C:N may be due to their larger surface area: volume ratio, which increases the efficiency of nutrient uptake (Fogg, 1995).
The high C:P and N:P values for copepods observed between October and December are of interest. Zooplankton are considered stoichiometrically homeostatic (Sterner et al., 1992). While seasonal variation in P and N content has been documented for marine copepods (Båmstedt, 1986), it is not a common occurrence in freshwater systems. Although there was no significant correlation between the C:N:P ratios of seston and any of the zooplankton species, there were apparent similarities in the seasonal stoichiometric patterns of the two. C:P ratios in particular followed similar seasonal trends (Figures 2b and f). In addition to stoichiometry, phytoplankton species composition which varies seasonally in Lake Malawi (Bootsma, 1993b), may also explain some of the temporal variations observed in zooplankton stoichiometry (Sterner and Schwalbach, 2001; Hessen et al., 2003). Zooplankton C:N ratios were less variable than C:P and N:P ratios, in agreement with observations for other lakes (Sterner and Elser, 2002). A comparison of our results with those observed in other lakes (Figure 1c and 1d) suggests that taxonomic differences in zooplankton stoichiometry are similar for temperate and tropical lakes.
Recycling of P and N by zooplankton can influence the biomass, production, and community structure of phytoplankton in marine and freshwater ecosystems (Sterner et al., 1992). James and Salonen (1991) observed that these nutrients can provide up to 100% of phytoplankton's net daily nutrient requirements. The magnitude however, is influenced by environmental and biological factors (Sterner, 1990; Sterner et al., 1992). According to homeostatic theory, zooplankton will selectively recycle N or P, depending on whether their N:P ratio is higher or lower than that of their food source. D. excisum and C. edulis have low N:P ratios compared to nano- and net- plankton and hence will selectively recycle N. Similar results were reported by Andersen and Hessen (1991), when comparing N and P recycling by cladocerans and copepods. However, if these zooplankton species in Lake Malawi are feeding on picoplankton that have a low N:P ratio, they will selectively recycle P. Little is known of size selectivity by D. excisum, but this genus is generally capable of filtering particles within the picoplankton size range (Hessen, 1985). The copepods T. cunningtoni, M. a. aequatorialis and T. neglectus have higher body N:P ratios than picoplankton and net plankton and hence will selectively recycle P when utilizing these two food sources. This is in agreement with the model results of Sterner (1990) which indicated lower N:P recycling ratios for copepods than for cladocerans. However, in the present study the N:P ratio of nanoplankton tended to be greater than that of these copepod species, and if nanoplankton are a major food source, they will tend to selectively recycle N and promote P limitation of phytoplankton. The above analyses do not consider the potential effect of egestion of fecal pellets by copepods. However, Sterner (1990), citing Le Borgne (1982), suggests the N:P of copepod fecal pellets can be similar to that of the zooplankton themselves, and so the effect of fecal pellet formation on selective nutrient recycling may be minimal.
The relative importance of zooplankton nutrient cycling in regulating phytoplankton production and species composition will depend on the magnitude of other nutrient loads, which are greatest during the mixing season (due to flux from the nutrient-rich hypolimnion) and the rainy season (due to river loading). The C:P and C:N ratios for seston suggest that nutrient limitation is greatest in the hot stratified season and early rainy season. Zooplankton nutrient recycling likely has the greatest potential to affect nutrient availability to phytoplankton during this period. At this time of year copepods with a high N:P ratio comprise over 90% of total zooplankton biomass. If picoplankton with a low N:P ratio make up a large fraction of zooplankton food during this period, grazing and nutrient excretion by copepods will promote N limitation. This is in agreement with observed peaks in heterocystous cyanobacteria biomass and N fixation in these months (Bootsma, 1993a; Gondwe et al., 2008). However, seston N:P ratios in Lake Malawi are frequently well above the Redfield ratio of 16:1 (Figure 2d), indicating that phytoplankton are more P limited than N limited for much of the year. Nutrient recycling by copepods may to some extent relieve P limitation, but it can only promote a switch to N limitation when seston N:P ratios are relatively low and when accumulation efficiencies are high (Table 1). While a large range of seston N:P ratios was observed in this study, picoplankton showed a stronger tendency to have low N:P ratios. Hence the ability of zooplankton to promote N limitation may depend on the size class of phytoplankton that they preferentially feed on. While recent studies have improved our understanding of lower food web trophic linkages in Lake Malawi (Ngochera and Bootsma, 2011), more information about size selectivity by zooplankton is needed to discern their role in Lake Malawi's nutrient cycle.
Seston stoichiometry in Lake Malawi varies seasonally. Seston C:N:P ratios suggest that phytoplankton are strongly nutrient limited through much of the year, with some relief of nutrient limitation apparent during the mixing season and the rainy season, when there is likely a greater nutrient supply to the surface mixed layer. The severity of nutrient stress appears to depend on size class, with picoplankton having lower C:P and N:P ratios than nanoplankton and net plankton. As a result, the influence of zooplankton grazing on selective N vs. P recycling will depend on the size class of phytoplankton that zooplankton graze on, highlighting the need for further studies to characterize size selective feeding by zooplankton.
Zooplankton elemental composition varies little throughout the year, although there are seasonal trends in the C:P ratio that roughly correspond to those in phytoplankton. There are significant differences in C:N:P ratios among zooplankton taxa, similar to those observed for temperate lakes. Through most of the year the zooplankton is dominated by copepods, which have a relatively high N:P ratio. While the effect of zooplankton nutrient recycling is likely masked by external nutrient loads to the epilimnion in the mixing season and rainy season, it may be more important in the warm, stratified season. While seston N:P ratios suggest that phytoplankton are generally P limited in Lake Malawi, selective P recycling by copepods may lead to stronger N limitation in the warm, stratified season, as reflected in the greater dominance of N-fixing cyanobacteria usually observed at this time of year.
Maxon J. Ngochera http://orcid.org/0000-0002-7524-6019
Harvey A. Bootsma http://orcid.org/0000-0002-9690-0155