Spatial variation of nitrogen fixation in lakes of the northern Great Plains

Cyanobacterial ; blooms are a regular feature of lakes in central North America, but little is known of their importance to the nitrogen (N) cycle and nutrition of aquatic food webs. We hypothesized that N2-fixing cyanobacteria constitute a significant source of N to prairie lakes, that fixed N is effectively transferred to primary and secondary consumers, and that the importance of fixed N is structured on a landscape scale due to spatial gradients of lake chemistry and catchment characteristics. These hypotheses were tested using stable isotope analyses and mass budgets in six chained lakes of the northern Great Plains that were sampled biweekly during summers of 1994–2004. Mean annual N isotope signatures of particulate organic matter (d15N-POM) were highly correlated to the abundance of N2-fixing cyanobacteria (r2 5 0.75, p , 0.001), but exhibited no marked spatial organization. Instead, the temporal variability of d15N-POM was greatest in downstream lakes where N2-fixing cyanobacteria were abundant. Furthermore, seasonal declines in d15N-POM were correlated with reductions in d15N of individual zooplankton taxa (Daphnia spp., Diacyclops thomasi, Leptodiaptomus siciloides, Leptodora kindtii), especially in downstream eastern lakes. N mass balances revealed that the importance of biological N2 fixation relative to total N inputs (up to 77%) and initial N standing stock (up to 201%) were significantly linearly correlated with landscape position (0.59 # r2 # 0.83, p # 0.07) and effective drainage area (0.61 # r2 # 0.98, p # 0.07), as were volumetric and areal estimates of N2 fixation during summer (r2 $ 0.87, p # 0.007). These patterns demonstrate that fixed N is spatially structured, highly predictable based on landscape position, and readily assimilated into aquatic food webs. About 90 to 130 million metric tons of nitrogen (N) are biologically fixed on land each year, while an equivalent amount may be fixed in oceans (Vitousek et al. 1997). In lakes, the importance of biological fixation is extremely variable, with annual rates varying between 200 and 9,200 mg N m22 in eutrophic lakes (Howarth et al. 1988b). In general, N2 fixation tends to be greater in the littoral or shallow regions of lakes than in pelagic zones where fixation rates usually increase from spring to late summer (Wetzel 2001). Recent studies of the importance of N2 fixation in lakes are based on analysis of temporal variations of N2-fixing cyanobacterial abundance (Havens et al. 2003), isotopic composition of primary producers (Rolff 2000; MacGregor et al. 2001), or acetylene to ethylene reduction (Tonno and Noges 2003; Ferber et al. 2004). However, most studies have been limited to a single site, and little is known of how N2 fixation varies on a spatial scale. Because lake position within the landscape is known to strongly influence characteristics such as catchment area, nutrient status, production, and biological diversity (Magnuson and Kratz 2000), the relative importance of atmospheric N2 fixation might be expected to vary in a spatially explicit fashion. Specifically, because the concentrations of conservative elements and nutrients often increase from upstream lakes to downstream lakes (Soranno et al. 1999), because lakes differ in the relative retention rates of N and phosphorus (P) (e.g., Kenney 1990), and because relative supply rates of N : P influence cyanobacterial abundance (Howarth et al. 1988a), the importance of fixed N to lake metabolism is hypothesized to vary along hydrologic gradients. Heterocystous, potentially toxic cyanobacteria are primarily responsible for pelagic N2 fixation, especially in eutrophic lakes with elevated concentrations of dissolved P, low N : P ratios, and high ambient irradiance (Millie et al. 1999) such as are common on the Great Plains of North America. For example, lakes of the prairie ecozone of Canada typically harbor species of the genera Aphanizomenon, Anabaena, and Gloeotrichia, in part because of Prich soils and high irradiance (Barica 1987). However, despite the prevalence of cyanobacteria in the phytoplankton of some lakes, there is little agreement on whether atmospheric N2 is quantitatively incorporated into the aquatic food webs. On one hand, the colonial nature and potential toxicity of heterocystous cyanobacteria make them poor food for zooplankton (de Bernardi and Giussani 1990). On the other hand, strong empirical correlations between cyanobacterial abundance and stable isotope signals in plankton and lake sediments (Talbot 2001) suggest that fixed N2 may be quantitatively important to some lake ecosystems. In particular, seasonal development of N2-fixing cyanobacteria often results in decreased ratios of 15N to 14N (lower d15N values) of algae (e.g., Rolff 2000; MacGregor et al. 2001; Lehmann et al. 2004) because prokaryotes assimilate atmospheric N2 with little fractionLimnology limn-51-04-18.3d 29/3/06 23:25:04 1 Cust # 05-261 1 Corresponding author ([email protected]). 2 Present address: Department of Biological Sciences, University of Alberta, Edmonton, Alberta, T6G 3E9, Canada. Acknowledgments We thank Lorelei Benoit, Martin Callaghan, Brianna Dopson, Twyla Jesse, and Laura Pfeifer for preparing samples; Curtis Brock and Richard Hughes for performing isotopic analyses; and Terry Chamulak (Saskatchewan Watershed Authority) for lake hydrology data. This work was supported by an NSERC postdoctoral fellowship to A.P., a Discovery Grant to P.R.L., and an NSERC PGDD to M.D.G. Manuscript preparation was supported by the Canada Research Chair program. Limnol. Oceanogr., 51(4), 2006, 000–000 E 2006, by the American Society of Limnology and Oceanography, Inc.