Animals can be important in nutrient cycling in particular ecosystems, but few studies have examined how this importance varies along environmental gradients. In this study we quantified the nutrient cycling role of an abundant detritivorous fish species, the gizzard shad (Dorosoma cepedianum), in reservoir ecosystems along a gradient of ecosystem productivity. Gizzard shad feed mostly on sediment detritus and excrete sediment-derived nutrients into the water column, thereby mediating a cross-habitat translocation of nutrients to phytoplankton. We quantified nitrogen and phosphorus cycling (excretion) rates of gizzard shad, as well as nutrient demand by phytoplankton, in seven lakes over a four-year period (16 lake-years). The lakes span a gradient of watershed land use (the relative amounts of land used for agriculture vs. forest) and productivity. As the watersheds of these lakes became increasingly dominated by agricultural land, primary production rates, lake trophic state indicators (total phosphorus and chlorophyll concentrations), and nutrient flux through gizzard shad populations all increased. Nutrient cycling by gizzard shad supported a substantial proportion of primary production in these ecosystems, and this proportion increased as watershed agriculture (and ecosystem productivity) increased. In the four productive lakes with agricultural watersheds (>78% agricultural land), gizzard shad supported on average 51% of phytoplankton primary production (range 27-67%). In contrast, in the three relatively unproductive lakes in forested or mixed-land-use watersheds (>47% forest, <52% agricultural land), gizzard shad supported 18% of primary production (range 14-23%). Thus, along a gradient of forested to agricultural landscapes, both watershed nutrient inputs and nutrient translocation by gizzard shad increase, but our data indicate that the importance of nutrient translocation by gizzard shad increases more rapidly. Our results therefore support the hypothesis that watersheds and gizzard shad jointly regulate primary production in reservoir ecosystems.

No abstract

1. Catchments export nutrients to aquatic ecosystems at rates and ratios that are strongly influenced by land use practices, and within aquatic ecosystems nutrients can be processed, retained, lost to the atmosphere, or exported downstream. The stoichiometry of carbon and nutrients can influence ecosystem services such as water quality, nutrient limitation, biodiversity, eutrophication and the sequestration of nutrients and carbon in sediments. However, we know little about how nutrient stoichiometry varies along the pathway from terrestrial landscapes through aquatic systems. 2. We studied the stoichiometry of nitrogen and phosphorus exported by three catchments of contrasting land use (forest versus agriculture) and in the water column and sediments of downstream reservoirs. We also related stoichiometry to phytoplankton nutrient limitation and the abundance of heterocystous cyanobacteria. 3. The total N : P of stream exports varied greatly among catchments and was 18, 54 and 140 (molar) in the forested, mixed-use and agricultural catchment, respectively. Total N : P in the mixed layers of the lakes was less variable but ordered similarly: 35, 52 132 in the forested, mixed-use and agricultural lake, respectively. In contrast, there was little variation among systems in the C : N and C : P ratios of catchment exports or in reservoir seston. 4. Phytoplankton in the forested lake were consistently N limited, those in the agricultural lake were consistently P limited, and those in the mixed-use lake shifted seasonally from P-to N limitation, reflecting N : P supply ratios. Total phytoplankton and cyanobacteria biomass were highest in the agricultural lake, but heterocystous (potentially N fixing) cyanobacteria were most abundant in the forested lake, corresponding to low N : P ratios. 5. Despite large differences in catchment export and water column N : P ratios, the N : P of sediment burial (integrated over several decades) was very low and remarkably similar (4.3-7.3) across reservoirs. N and P budgets constructed for the agricultural reservoir suggested that denitrification could be a major loss of N, and may help explain the relatively low N : P of buried sediment. 6. Our results show congruence between the catchment export N : P, reservoir N : P, phytoplankton N versus P limitation and the dominance of heterocystous cyanobacteria. However, the N : P stoichiometry of sediments retained in the lakes was relatively insensitive to catchment stoichiometry, suggesting that a common set of biogeochemical processes constrains sediment N : P across lakes of contrasting catchment land use.

The stoichiometric composition of autotrophs can vary greatly in response to variation in light and nutrient availability, and can mediate ecological processes such as C sequestration, growth of herbivores, and nutrient cycling. We investigated light and nutrient effects on phytoplankton stoichiometry, employing five experiments on intact phytoplankton assemblages from three lakes varying in productivity and species composition. Each experiment employed two nutrient and eight irradiance levels in a fully factorial design. Light and nutrients interactively affected phytoplankton stoichiometry. Thus, phytoplankton C:N, C:P, and N:P ratios increased with irradiance, and slopes of the stoichiometric ratio versus irradiance relationships were steeper with ambient nutrients than with nutrients added. Our results support the light-nutrient hypothesis, which predicts that phytoplankton C:nutrient ratios are functions of the ratio of available light and nutrients; however, we observed considerable variation among lakes in the expression of this relationship. Phytoplankton species diversity was positively correlated with the slopes of the C:N and C:P versus irradiance relationships, suggesting that diverse assemblages may exhibit greater flexibility in the response of phytoplankton nutrient stoichiometry to light and nutrients. The interactive nature of light and nutrient effects may render it difficult to generate predictive models of stoichiometric responses to these two factors. Our results point to the need for future studies that examine stoichiometric responses across a wide range of phytoplankton communities.

Climate-change models predict more frequent and intense summer droughts for many areas, including the midwestern United States. Precipitation quantity and intensity in turn drive the rates and ratios at which nitrogen (N) and phosphorus (P) are exported from watersheds into lakes, but these rates and ratios are also modulated by watershed land use. This led us to ask the question, is the effect of precipitation on phytoplankton nutrient limitation dependent on watershed land use? Across 42 lakes, we found that phytoplankton in lakes in agricultural landscapes were usually P limited but shifted to strong N limitation under increased drought intensity, and that droughts promoted N-fixing cyanobacteria. In contrast, phytoplankton in lakes with forested watersheds were consistently N limited, regardless of drought status. This climate-land use interaction suggests that droughts may increase the incidence of N limitation in agriculturally impacted lakes. N limitation would likely impair valuable ecosystem services such as drinking water, fisheries, and recreation by promoting the occurrence and severity of cyanobacterial blooms.