Oxygen availability is decreasing in many lakes and reservoirs worldwide, raising the urgency for understanding how anoxia (low oxygen) affects coupled biogeochemical cycling, which has major implications for water quality, food webs, and ecosystem functioning. Although the increasing magnitude and prevalence of anoxia has been documented in freshwaters globally, the challenges of disentangling oxygen and temperature responses have hindered assessment of the effects of anoxia on carbon, nitrogen, and phosphorus concentrations, stoichiometry (chemical ratios), and retention in freshwaters. The consequences of anoxia are likely severe and may be irreversible, necessitating ecosystem-scale experimental investigation of decreasing freshwater oxygen availability. To address this gap, we devised and conducted REDOX (the Reservoir Ecosystem Dynamic Oxygenation eXperiment), an unprecedented, seven-year experiment in which we manipulated and modeled bottom-water (hypolimnetic) oxygen availability at the whole-ecosystem scale in a eutrophic reservoir. Seven years of data reveal that anoxia significantly increased hypolimnetic carbon, nitrogen, and phosphorus concentrations and altered elemental stoichiometry by factors of 2-5 relative to oxic periods. Importantly, prolonged summer anoxia increased nitrogen export from the reservoir by six-fold and changed the reservoir from a net sink to a net source of phosphorus and organic carbon downstream. While low oxygen in freshwaters is thought of as a response to land use and climate change, results from REDOX demonstrate that low oxygen can also be a driver of major changes to freshwater biogeochemical cycling, which may serve as an intensifying feedback that increases anoxia in downstream waterbodies. Consequently, as climate and land use change continue to increase the prevalence of anoxia in lakes and reservoirs globally, it is likely that anoxia will have major effects on freshwater carbon, nitrogen, and phosphorus budgets as well as water quality and ecosystem functioning.

The National Ecological Observatory Network (NEON)’s standardized monitoring program provides an unprecedented opportunity for comparing the predictability of ecosystems. To harness the power of NEON data for examining environmental predictability, we scaled a near-term, iterative water temperature forecasting system to all six conterminous NEON lakes. We generated 1 to 35-day ahead forecasts using a process-based hydrodynamic model that was updated with observations as they became available. Forecasts were more accurate than a null model up to 35-days ahead among lakes, with an aggregated 1-day ahead RMSE (root-mean square error) of 0.60℃ and 35-days ahead RMSE of 2.17℃. Water temperature forecast accuracy was positively associated with lake depth and water clarity, and negatively associated with catchment size and fetch. Our results suggest that lake characteristics interact with weather to control the predictability of thermal structure. Our work provides some of the first probabilistic forecasts of NEON sites and a framework for examining continental-scale predictability.

1. Freshwater phytoplankton communities are currently experiencing multiple global change stressors, including increasing frequency and intensity of storms. An important mechanism by which storms affect lake and reservoir phytoplankton is by altering the water column’s thermal structure (e.g., changes to thermocline depth). However, little is known about the effects of intermittent thermocline deepening on phytoplankton community vertical distribution and composition or the consistency of phytoplankton responses to varying frequency of these disturbances over multiple years. 2. We conducted whole-ecosystem thermocline deepening manipulations in a small reservoir. We used an epilimnetic mixing system to experimentally deepen the thermocline in two summers, simulating potential responses to storms, and did not manipulate thermocline depth in two succeeding summers. We collected weekly depth profiles of water temperature, light, nutrients, and phytoplankton biomass as well as discrete samples to assess phytoplankton community composition. We then used time-series analysis and multivariate ordination to assess the effects of intermittent thermocline deepening due to both our experimental manipulations and naturally-occurring storms on phytoplankton community structure. 3. We observed inter-annual and intra-annual variability in phytoplankton community response to thermocline deepening. We found that peak phytoplankton biomass was significantly deeper in years with a higher frequency of thermocline deepening events (i.e., years with both manipulations and natural storms) due to weaker thermal stratification and deeper depth distributions of soluble reactive phosphorus. Furthermore, we found that the depth of peak phytoplankton biomass was linked to phytoplankton community composition, with certain taxa being associated with deep or shallow biomass peaks, often according to functional traits such as optimal growth temperature, mixotrophy, and low-light tolerance. 4. Our results demonstrate that abrupt thermocline deepening due to water column mixing affects both phytoplankton depth distribution and community structure via alteration of physical and chemical gradients. In addition, our work supports previous research that phytoplankton depth distribution and community composition interact at inter-annual and intra-annual timescales. 5. Variability in the inter-annual and intra-annual responses of phytoplankton to abrupt thermocline deepening indicates that antecedent conditions and the seasonal timing of surface water mixing may mediate these responses. Our findings emphasize that phytoplankton depth distributions are sensitive to global change stressors and effects on depth distributions should be taken into account when predicting phytoplankton responses to increased storms under global change.

Globally-significant quantities of carbon (C), nitrogen (N), and phosphorus (P) enter freshwater reservoirs each year. These inputs can be buried in sediments, respired, taken up by organisms, emitted to the atmosphere, or exported downstream. While much is known about reservoir-scale biogeochemical processing, less is known about spatial and temporal variability of biogeochemistry within a reservoir along the continuum from inflowing streams to the dam. To address this gap, we examined longitudinal variability in surface water biogeochemistry (C, N, and P) in two small reservoirs throughout a thermally-stratified season. We sampled total and dissolved fractions of C, N, and P, and chlorophyll-a from each reservoir’s major inflows to the dam. We found that time was generally a more important driver of heterogeneity in biogeochemical concentrations than space. However, dissolved nutrient and organic carbon concentrations had high site-to-site variability within both reservoirs, potentially as a result of shifting biological activity or environmental conditions. When considering spatially explicit processing, we found that certain locations within the reservoir, most often the stream-reservoir interface, acted as ‘hotspots’ of change in biogeochemical concentrations. Our study suggests that spatially explicit metrics of biogeochemical processing could help constrain the role of reservoirs in C, N, and P cycles in the landscape. Ultimately, our results highlight that biogeochemical heterogeneity in small reservoirs is driven more by seasonality than longitudinal spatial gradients, and that some sites within reservoirs play critically important roles in whole-ecosystem biogeochemical processing.

As climate and land use increase the variability of many ecosystems, forecasts of ecological variables are needed to inform management and use of ecosystem services. In particular, forecasts of phytoplankton would be especially useful for drinking water management, as phytoplankton populations are exhibiting greater fluctuations due to human activities. While phytoplankton forecasts are increasing in number, many questions remain regarding the optimal model time step (the temporal frequency of the forecast model output), time horizon (the length of time into the future a prediction is made) for maximizing forecast performance, as well as what factors contribute to uncertainty in forecasts and their scalability among sites. To answer these questions, we developed near-term, iterative forecasts of phytoplankton 1 to 14 days into the future using forecast models with three different time steps (daily, weekly, fortnightly), that included a full uncertainty partitioning analysis at two drinking water reservoirs. We found that forecast accuracy varies with model time step and forecast horizon, and that forecast models can outperform null estimates under most conditions. Weekly and fortnightly forecasts consistently outperformed daily forecasts at 7-day and 14-day horizons, a trend which increased up to the 14-day forecast horizon. Importantly, our work suggests that forecast accuracy can be increased by matching the forecast model time step to the forecast horizon for which predictions are needed. We found that model process uncertainty was the primary source of uncertainty in our phytoplankton forecasts over the forecast period, but parameter uncertainty increased during phytoplankton blooms and when scaling the forecast model to a new site. Overall, our scalability analysis shows promising results that simple models can be transferred to produce forecasts at additional sites. Altogether, our study advances our understanding of how forecast model time step and forecast horizon influence the forecastability of phytoplankton dynamics in aquatic systems, and adds to the growing body of work regarding the predictability of ecological systems broadly.

Near-term ecological forecasts provide resource managers advance notice of changes in ecosystem services, such as fisheries stocks, timber yields, or water and air quality. Importantly, ecological forecasts can identify where uncertainty enters the forecasting system, which is necessary to refine and improve forecast skill and guide interpretation of forecast results. Uncertainty partitioning identifies the relative contributions to total forecast variance (uncertainty) introduced by different sources, including specification of the model structure, errors in driver data, and estimation of initial state conditions. Uncertainty partitioning could be particularly useful in improving forecasts of high-density cyanobacterial events, which are difficult to predict and present a persistent challenge for lake managers. Cyanobacteria can produce toxic or unsightly surface scums and advance warning of these events could help managers mitigate water quality issues. Here, we calibrate fourteen Bayesian state-space models to evaluate different hypotheses about cyanobacterial growth using data from eight summers of weekly cyanobacteria density samples in an oligotrophic (low nutrient) lake that experiences sporadic surface scums of the toxin-producing cyanobacterium, Gloeotrichia echinulata. We identify dominant sources of uncertainty for near-term (one-week to four-week) forecasts of G. echinulata densities over two years. Water temperature was an important predictor in calibration and at the four-week forecast horizon. However, no environmental covariates improved over a simple autoregressive (AR) model at the one-week horizon. Even the best fit models exhibited large variance in forecasted cyanobacterial densities and often did not capture rare peak density occurrences, indicating that significant explanatory variables in calibration are not always effective for near-term forecasting of low-frequency events. Uncertainty partitioning revealed that model process specification and initial conditions uncertainty dominated forecasts at both time horizons. These findings suggest that observed densities result from both growth and movement of G. echinulata, and that imperfect observations as well as spatial misalignment of environmental data and cyanobacteria observations affect forecast skill. Future research efforts should prioritize long-term studies to refine process understanding and increased sampling frequency and replication to better define initial conditions. Our results emphasize the importance of ecological forecasting principles and uncertainty partitioning to refine and understand predictive capacity across ecosystems.

Lakes are traditionally classified based on their thermal regime and trophic status. While this classification adequately captures many lakes, it is not sufficient to understand seasonally ice-covered lakes, the most common lake type on Earth. Here, we propose an additional classification to differentiate under-ice stratification. When ice forms in smaller and deeper lakes, inverse stratification will form with a thin buoyant layer of cold water (near 0oC) below the ice, which remains above a deeper 4oC layer. In contrast, the entire water column can cool to ~0oC in larger and shallower lakes. We suggest these alternative conditions for dimictic lakes be termed “cryostratified” and “cryomictic.” We describe the inverse thermal stratification in 19 highly varying lakes and derive a model that predicts the temperature profile as a function of wind stress, area, and depth. The model opens up for a more precise prediction of lake responses to a warming climate.