SUMMARY1. Nutrients released from lake sediments can influence water column nutrient concentrations and planktonic productivity. We examined sediment nutrient release [soluble reactive phosphorus (SRP) and ammonia (NH þ 4 )] at two sites in a eutrophic reservoir (Acton Lake, OH, U.S.A.) that differed in physical mixing conditions (a thermally stratified and an unstratified site). 2. Sediment nutrient release rates were estimated with three methods: sediment core incubations, seasonal in situ hypolimnetic accumulation and a published regression model that predicted sediment phosphorous (P) release rate from sediment P concentration. All three methods were applied to the deeper stratified site in the reservoir; however, we used only sediment core incubations to estimate SRP and NH þ 4 release rates at the shallow unstratified site because of the lack of thermal stratification. We also compared the total P concentration (TP S ) of sediments and the concentration of P in various sediment fractions at both sites. 3. Anoxic sediments at the stratified site released SRP at rates more than an order of magnitude greater than oxic sediments at the shallow unstratified site. However, P accumulated in the hypolimnion at much lower rates than predicted by sediment core incubations. In contrast, NH þ 4 was released at similar rates at both sites and accumulated in the hypolimnion at close to the expected rate, indicating that P was 'lost' from the hypolimnion through biogeochemical pathways for P, such as precipitation with inorganic material or biological uptake and sedimentation. 4. TP S was significantly greater at the deeper stratified site and organically bound P accounted for >50% of TP S at both sites. 5. We examined the magnitude of SRP fluxes into the study reservoir in 1996 by comparing the mean summer daily SRP fluxes from anaerobic sediments, aerobic sediments, stream inflows and gizzard shad excretion. While the SRP release from anaerobic sediments was high, we hypothesise that little of this SRP gained access to the epilimnion in mid-summer. SRP flux to the reservoir from aerobic sediments was less than from gizzard shad excretion and streams. Large interannual variability in thermocline stability, gizzard shad biomass and stream discharge volumes, will affect SRP loading rates from different sources in different years. Therefore, construction of P budgets for different years should account for interannual variation in these parameters.

Resource pulses affect productivity and dynamics in a diversity of ecosystems, including islands, forests, streams, and lakes. Terrestrial and aquatic systems differ in food web structure and biogeochemistry; thus they may also differ in their responses to resource pulses. However, there has been a limited attempt to compare responses across ecosystem types. Here, we identify similarities and differences in the causes and consequences of resource pulses in terrestrial and aquatic systems. We propose that different patterns of food web and ecosystem structure in terrestrial and aquatic systems lead to different responses to resource pulses. Two predictions emerge from a comparison of resource pulses in the literature: (1) the bottom-up effects of resource pulses should transmit through aquatic food webs faster because of differences in the growth rates, life history, and stoichiometry of organisms in aquatic vs. terrestrial systems, and (2) the impacts of resource pulses should also persist longer in terrestrial systems because of longer generation times, the long-lived nature of many terrestrial resource pulses, and reduced top-down effects of consumers in terrestrial systems compared to aquatic systems. To examine these predictions, we use a case study of a resource pulse that affects both terrestrial and aquatic systems: the synchronous emergence of periodical cicadas (Magicicada spp.) in eastern North American forests. In general, studies that have examined the effects of periodical cicadas on terrestrial and aquatic systems support the prediction that resource pulses transmit more rapidly in aquatic systems; however, support for the prediction that resource pulse effects persist longer in terrestrial systems is equivocal. We conclude that there is a need to elucidate the indirect effects and long-term implications of resource pulses in both terrestrial and aquatic ecosystems.

Periodical cicadas emerge from below ground every 13 or 17 years in North American forests, with individual broods representing the synchronous movement of trillions of individuals across geographic regions. Due to predator satiation, most individuals escape predation, die, and become deposited as detritus. Some of this emergent biomass falls into woodland aquatic habitats (small streams and woodland ponds) and serves as a high-quality allochthonous detritus pulse in early summer. We present results of a two-part study in which we (1) quantified deposition of Brood X periodical cicada detritus into woodland ponds and low-order streams in southwestern Ohio, and (2) conducted an outdoor mesocosm experiment in which we examined the effects of deposition of different amounts of cicada detritus on food webs characteristic of forest ponds. In the mesocosm experiment, we manipulated the amount of cicada detritus input to examine if food web dynamics and stability varied with the magnitude of this allochthonous resource subsidy, as predicted by numerous theoretical models. Deposition data indicate that, during years of periodical cicada emergence, cicada carcasses can represent a sizable pulse of allochthonous detritus to forest aquatic ecosystems. In the mesocosm experiment, cicada carcass deposition rapidly affected food webs, leading to substantial increases in nutrients and organism biomass, with the magnitude of increase dependent upon the amount of cicada detritus. Deposition of cicada detritus impacted the stability of organism functional groups and populations by affecting the temporal variability and biomass minima. However, contrary to theory, stability measures were not consistently related to the size of the allochthonous pulse (i.e., the amount of cicada detritus). Our study underscores the need for theory to further explore consequences of pulsed allochthonous subsidies for food web stability.

We examined the biomass-dependent effects of common carp (Cyprinus carpio) on water quality in 10 ponds at the Eagle Mountain Fish Hatchery, Fort Worth, Texas, USA. Ponds contained 0-465 kg ha )1 of common carp. We measured limnological variables at weekly intervals for four weeks in early summer, after which ponds were drained and the biomass of fish and macrophytes was determined. Common carp biomass was significantly positively correlated with chlorophyll a, total phosphorus, total nitrogen, and Keratella spp. density and negatively correlated to bushy pondweed (Najas guadalupensis) biomass. In addition, we combined our data with data from comparable studies to develop more robust regression models that predict the biomass-dependent effects of common carp on water quality variables across a wide range of systems.

The prevailing paradigm in subterranean ecology is that below-ground food webs are simple, limited to one or two trophic levels, and composed of generalist species because of spatio-temporally patchy food resources and pervasive energy limitation. This paradigm is based on relatively few studies of easily accessible, air-filled caves. However, in some subterranean ecosystems, chemolithoautotrophy can subsidize or replace surface-based allochthonous inputs of photosynthetically derived organic matter (OM) as a basal food resource and promote niche specialization and evolution of higher trophic levels. Consequently, the current subterranean trophic paradigm fails to account for variation in resources, trophic specialization, and food chain length in some subterranean ecosystems. We reevaluated the subterranean food web paradigm by examining spatial variation in the isotopic composition of basal food resources and consumers, food web structure, stygobiont species diversity, and chromophoric organic matter (CDOM), across a geochemical gradient in a large and complex groundwater system, the Edwards Aquifer in Central Texas (USA). Mean δ13C values of stygobiont communities become increasingly more negative along the gradient of photosynthetic OM sources near the aquifer recharge zone to chemolithoautotrophic OM sources closer to the freshwater-saline water interface (FWSWI) between oxygenated freshwater and anoxic, sulfide-rich saline water. Stygobiont community species richness declined with increasing distance from the FWSWI. Bayesian mixing models were used to estimate the relative importance of photosynthetic OM and chemolithoautorophic OM for stygobiont communities at three biogeochemically distinct sites. The contribution of chemolithoautotrophic OM to consumers at these sites ranged between 25% and 69% of total OM utilized and comprised as much as 88% of the diet for one species. In addition, the food web adjacent to the FWSWI had greater trophic diversity when compared to the other two sites. Our results suggest that diverse OM sources and in situ, chemolithoautotrophic OM production can support complex groundwater food webs and increase species richness. Chemolithoautotrophy has been fundamental for the long-term maintenance of species diversity, trophic complexity, and community stability in this subterranean ecosystem, especially during periods of decreased photosynthetic production and groundwater recharge that have occurred over geologic time scales.