Lake Mendota Research Articles

Coassembly and binning of a twenty-year metagenomic time-series from Lake Mendota

The North Temperate Lakes Long-Term Ecological Research (NTL-LTER) program has been extensively used to improve understanding of how aquatic ecosystems respond to environmental stressors, climate fluctuations, and human activities. Here, we report on the metagenomes of samples collected between 2000 and 2019 from Lake Mendota, a freshwater eutrophic lake within the NTL-LTER site. We utilized the distributed metagenome assembler MetaHipMer to coassemble over 10 terabases (Tbp) of data from 471 individual Illumina-sequenced metagenomes. A total of 95,523,664 contigs were assembled and binned to generate 1,894 non-redundant metagenome-assembled genomes (MAGs) with ≥50% completeness and ≤10% contamination. Phylogenomic analysis revealed that the MAGs were nearly exclusively bacterial, dominated by Pseudomonadota (Proteobacteria, N = 623) and Bacteroidota (N = 321). Nine eukaryotic MAGs were identified by eukCC with six assigned to the phylum Chlorophyta. Additionally, 6,350 high-quality viral sequences were identified by geNomad with the majority classified in the phylum Uroviricota. This expansive coassembled metagenomic dataset provides an unprecedented foundation to advance understanding of microbial communities in freshwater ecosystems and explore temporal ecosystem dynamics.

Time-series metagenomics reveals changing protistan ecology of a temperate dimictic lake

BackgroundProtists, single-celled eukaryotic organisms, are critical to food web ecology, contributing to primary productivity and connecting small bacteria and archaea to higher trophic levels. Lake Mendota is a large, eutrophic natural lake that is a Long-Term Ecological Research site and among the world’s best-studied freshwater systems. Metagenomic samples have been collected and shotgun sequenced from Lake Mendota for the last 20 years. Here, we analyze this comprehensive time series to infer changes to the structure and function of the protistan community and to hypothesize about their interactions with bacteria.ResultsBased on small subunit rRNA genes extracted from the metagenomes and metagenome-assembled genomes of microeukaryotes, we identify shifts in the eukaryotic phytoplankton community over time, which we predict to be a consequence of reduced zooplankton grazing pressures after the invasion of a invasive predator (the spiny water flea) to the lake. The metagenomic data also reveal the presence of the spiny water flea and the zebra mussel, a second invasive species to Lake Mendota, prior to their visual identification during routine monitoring. Furthermore, we use species co-occurrence and co-abundance analysis to connect the protistan community with bacterial taxa. Correlation analysis suggests that protists and bacteria may interact or respond similarly to environmental conditions. Cryptophytes declined in the second decade of the timeseries, while many alveolate groups (e.g., ciliates and dinoflagellates) and diatoms increased in abundance, changes that have implications for food web efficiency in Lake Mendota.ConclusionsWe demonstrate that metagenomic sequence-based community analysis can complement existing efforts to monitor protists in Lake Mendota based on microscopy-based count surveys. We observed patterns of seasonal abundance in microeukaryotes in Lake Mendota that corroborated expectations from other systems, including high abundance of cryptophytes in winter and diatoms in fall and spring, but with much higher resolution than previous surveys. Our study identified long-term changes in the abundance of eukaryotic microbes and provided context for the known establishment of an invasive species that catalyzes a trophic cascade involving protists. Our findings are important for decoding potential long-term consequences of human interventions, including invasive species introduction.BaNPxYigHXA7H5uqkH8sryVideo

Stochastic dynamics of phycocyanin in years of contrasting phosphorus load

AbstractResilience, measured by the distribution of passage times between alternate states, indicates persistence of a state in stochastic dynamic systems such as blooms of cyanobacteria in lakes. We used high‐frequency datasets to compare the resilience of low and high states of phycocyanin, a pigment indicator of cyanobacteria, in Lake Mendota, Wisconsin, USA, for three growing seasons that ranged sevenfold in external phosphorus (P) load. Each year we observed 139–265 passage times across the unstable threshold that separated the low‐ from high‐phycocyanin states. Each sample of passage times is highly skewed with low median, larger mean, much larger SD, and wide tails extending to long lifetimes of a state. About 25% of events, whether low or high phycocyanin, lasted a day or more. Among these 3 years of contrasting external P load, there were no discernible differences in the resilience of either ecosystem state. We attribute this lack of contrast to the sustained recycling of P from sediments and the high stochasticity of phycocyanin in this lake.

A Machine Learning and Remote Sensing‐Based Model for Algae Pigment and Dissolved Oxygen Retrieval on a Small Inland Lake

AbstractExcessive algae growth can lead to negative consequences for ecosystem function, economic opportunity, and human and animal health. Due to the cost‐effectiveness and temporal availability of satellite imagery, remote sensing has become a powerful tool for water quality monitoring. The use of remotely sensed products to monitor water quality related to algae and cyanobacteria productivity during a bloom event may help inform management strategies for inland waters. To evaluate the ability of satellite imagery to monitor algae pigments and dissolved oxygen conditions in a small inland lake, chlorophyll‐a, phycocyanin, and dissolved oxygen concentrations are measured using a YSI EXO2 sonde during Sentinel‐2 and Sentinel‐3 overpasses from 2019 to 2022 on Lake Mendota, WI. Machine learning methods are implemented with existing algorithms to model chlorophyll‐a, phycocyanin, and Pc:Chla. A novel machine learning‐based dissolved oxygen modeling approach is developed using algae pigment concentrations as predictors. Best model results based on Sentinel‐2 (Sentinel‐3) imagery achieved R2 scores of 0.47 (0.42) for chlorophyll‐a, 0.69 (0.22) for phycocyanin, and 0.70 (0.41) for Pc:Chla. Dissolved oxygen models achieved an R2 of 0.68 (0.36) when applied to Sentinel‐2 (Sentinel‐3) imagery, and Pc:Chla is found to be the most important predictive feature. Random forest models are better suited to water quality estimations in this system given built in methods for feature selection and a relatively small data set. Use of these approaches for estimation of Pc:Chla and dissolved oxygen can increase the water quality information extracted from satellite imagery and improve characterization of algae conditions among inland waters.

Legacy Phosphorus and Ecosystem Memory Control Future Water Quality in a Eutrophic Lake

AbstractLake water clarity, phytoplankton biomass, and hypolimnetic oxygen concentration are metrics of water quality that are highly degraded in eutrophic systems. Eutrophication is linked to legacy nutrients stored in catchment soils and in lake sediments. Long lags in water quality improvement under scenarios of nutrient load reduction to lakes indicate an apparent ecosystem memory tied to the interactions between water biogeochemistry and lake sediment nutrients. To investigate how nutrient legacies and ecosystem memory control lake water quality dynamics, we coupled nutrient cycling and lake metabolism in a model to recreate long‐term water quality of a eutrophic lake (Lake Mendota, Wisconsin, USA). We modeled long‐term recovery of water quality under scenarios of nutrient load reduction and found that the rates and patterns of water quality improvement depended on changes in phosphorus (P) and organic carbon storage in the water column and sediments. Through scenarios of water quality improvement, we showed that water quality variables have distinct phases of change determined by the turnover rates of storage pools—an initial and rapid water quality improvement due to water column flushing, followed by a much longer and slower improvement as sediment P pools were slowly reduced. Water clarity, phytoplankton biomass, and hypolimnetic dissolved oxygen differed in their time responses. Water clarity and algal biomass improved within years of nutrient reductions, but hypolimnetic oxygen took decades to improve. Even with reduced catchment loading, recovery of Lake Mendota to a mesotrophic state may require decades due to nutrient legacies and long ecosystem memory.

Quantifying co‐benefits of water quality policies: An integrated assessment model of land and nitrogen management

AbstractDue to the nature of nitrogen cycling, policies designed to address water quality concerns have the potential to provide benefits beyond the targeted water quality improvements. For example, actions to protect water quality by reducing nitrate leaching from agriculture also reduce emissions of nitrous oxide, a potent greenhouse gas. These positive effects, which are incidental to the regulation's intended target, are termed “co‐benefits.” To quantify the co‐benefits associated with reduced nitrate leaching, we integrate an economic model of farmer decision making with a model of terrestrial nitrogen cycling for the watershed surrounding Lake Mendota, Wisconsin, USA. Our modeling approach provides a framework that links air and water pollutants in an agri‐environmental system and offers a direction for future studies. Our model results highlight the finding that the co‐benefits from nitrous oxide abatement are substantial, and their inclusion increases the benefit–cost ratio of water quality policies. Consideration of these co‐benefits has the potential to reverse the conclusions of benefit–cost analysis in the assessment of current water quality policies.

Physiological and genomic evidence of cysteine degradation and aerobic hydrogen sulfide production in freshwater bacteria.

The sulfur-containing amino acid cysteine is abundant in the environment, including in freshwater lakes. Biological cysteine degradation can result in hydrogen sulfide (H2S), a toxic and ecologically relevant compound that is a central player in biogeochemical cycling in aquatic environments. Here, we investigated the ecological significance of cysteine in oxic freshwater, using isolated cultures, controlled experiments, and multiomics. We screened bacterial isolates enriched from natural lake water for their ability to produce H2S when provided cysteine. We identified 29 isolates (Bacteroidota, Proteobacteria, and Actinobacteria) that produced H2S. To understand the genomic and genetic basis for cysteine degradation and H2S production, we further characterized three isolates using whole-genome sequencing (using a combination of short-read and long-read sequencing) and tracked cysteine and H2S levels over their growth ranges: Stenotrophomonas maltophilia (Gammaproteobacteria), S. bentonitica (Gammaproteobacteria), and Chryseobacterium piscium (Bacteroidota). Cysteine decreased and H2S increased, and all three genomes had genes involved in cysteine degradation. Finally, to assess the presence of these organisms and genes in the environment, we surveyed a 5-year time series of metagenomic data from the same isolation source (Lake Mendota, Madison, WI, USA) and identified their presence throughout the time series. Overall, our study shows that diverse isolated bacterial strains can use cysteine and produce H2S under oxic conditions, and we show evidence using metagenomic data that this process may occur more broadly in natural freshwater lakes. Future considerations of sulfur cycling and biogeochemistry in oxic environments should account for H2S production from the degradation of organosulfur compounds. IMPORTANCE Hydrogen sulfide (H2S), a naturally occurring gas with both biological and abiotic origins, can be toxic to living organisms. In aquatic environments, H2S production typically originates from anoxic (lacking oxygen) environments, such as sediments, or the bottom layers of thermally stratified lakes. However, the degradation of sulfur-containing amino acids such as cysteine, which all cells and life forms rely on, can be a source of ammonia and H2S in the environment. Unlike other approaches for biological H2S production such as dissimilatory sulfate reduction, cysteine degradation can occur in the presence of oxygen. Yet, little is known about how cysteine degradation influences sulfur availability and cycling in freshwater lakes. In our study, we identified diverse bacteria from a freshwater lake that can produce H2S in the presence of O2. Our study highlights the ecological importance of oxic H2S production in natural ecosystems and necessitates a change in our outlook on sulfur biogeochemistry.

Long-range dependence and extreme values of precipitation, phosphorus load, and Cyanobacteria

Extreme daily values of precipitation (1939-2021), discharge (1991-2021), phosphorus (P) load (1994-2021), and phycocyanin, a pigment of Cyanobacteria (June 1-September 15 of 2008-2021) are clustered as multi-day events for Lake Mendota, Wisconsin. Long-range dependence, or memory, is the shortest for precipitation and the longest for phycocyanin. Extremes are clustered for all variates and those of P load and phycocyanin are most strongly clustered. Extremes of P load are predictable from extremes of precipitation, and precipitation and P load are correlated with later concentrations of phycocyanin. However, time delays from 1 to 60 d were found between P load extremes and the next extreme phycocyanin event within the same year of observation. Although most of the lake's P enters in extreme events, blooms of Cyanobacteria may be sustained by recycling and food web processes.

Diagnosing Discrepancies between Observations and Models of Surface Energy Fluxes in a Midlatitude Lake

Abstract The physical processes of heat exchange between lakes and the surrounding atmosphere are important in simulating and predicting terrestrial surface energy balance. Latent and sensible heat fluxes are the dominant physical process controlling ice growth and decay on the lake surface, as well as having influence on regional climate. While one-dimensional lake models have been used in simulating environmental changes in ice dynamics and water temperature, understanding the seasonal to daily cycles of lake surface energy balance and its relationship to lake thermal properties, atmospheric conditions, and how those are represented in models is still an open area of research. We evaluated a pair of one-dimensional lake models, Freshwater Lake (FLake) and the General Lake Model (GLM), to compare modeled latent and sensible heat fluxes against observed data collected by an eddy covariance tower during a 1-yr period in 2017, using Lake Mendota in Madison, Wisconsin, as our study site. We hypothesized transitional periods of ice cover as a leading source of model uncertainty, and we instead found that the models failed to simulate accurate values for large positive heat fluxes that occurred from late August into late December. Our results ultimately showed that one-dimensional models are effective in simulating sensible heat fluxes but are considerably less sensitive to latent heat fluxes than the observed relationships of latent heat flux to environmental drivers. These results can be used to focus future improvement of these lake models especially if they are to be used for surface boundary conditions in regional numerical weather models. Significance Statement While lakes consist of a small amount of Earth’s surface, they have a large impact on local climate and weather. A large amount of energy is stored in lakes during the spring and summer, and then removed from lakes before winter. The effect is particularly noticeable in high latitudes, when the seasonal temperature difference is larger. Modeling this lake energy exchange is important for weather models and measuring this energy exchange is challenging. Here we compare modeled and observed energy exchange, and we show there are large amounts of energy exchange happening in the fall, which models struggle to capture well. During periods of partial ice coverage in early winter, lake behavior can change rapidly.

Evolutionary change in metabolic rate of Daphnia pulicaria following invasion by the predator Bythotrephes longimanus.

Metabolic rate is a trait that may evolve in response to the direct and indirect effects of predator‐induced mortality. Predators may indirectly alter selection by lowering prey densities and increasing resource availability or by intensifying resource limitation through changes in prey behavior (e.g., use of less productive areas). In the current study, we quantify the evolution of metabolic rate in the zooplankton Daphnia pulicaria following an invasive event by the predator Bythotrephes longimanus in Lake Mendota, Wisconsin, US. This invasion has been shown to dramatically impact D. pulicaria, causing a ~60% decline in their biomass. Using a resurrection ecology approach, we compared the metabolic rate of D. pulicaria clones originating prior to the Bythotrephes invasion with that of clones having evolved in the presence of Bythotrephes. We observed a 7.4% reduction in metabolic rate among post‐invasive clones compared to pre‐invasive clones and discuss the potential roles of direct and indirect selection in driving this change.

Algal bloom forecasting with time-frequency analysis: A hybrid deep learning approach

The rapid emergence of deep learning long-short-term-memory (LSTM) technique presents a promising solution to algal bloom forecasting. However, the discontinuous and non-stationary processes within algal dynamics still largely limit the functions of LSTMs. To overcome this challenge, an advanced time-frequency wavelet analysis (WA) technique was introduced to enhance the prediction accuracy of LSTMs. Herein, the novel hybrid approach (named WLSTM) successfully decreased the algal forecasting inaccuracy of classic LSTMs by 41% ± 8% in Lake Mendota (Wisconsin, USA), with powerful one-step-ahead predictions at hourly, daily, and monthly time resolutions (R2 = 0.976, 0.878, and 0.814, respectively). In addition, the WLSTM outperformed the other two widely used algal forecasting approaches - deep neural network (DNN), and autoregressive-integrated-moving-average (ARIMA) model, represented by average 72% and 85% decrease in root-mean-square-error, respectively. Furthermore, the WLSTM was implemented in an experimentally fertilized lake (Lake Tuesday, Michigan) for a multi-step forecasting examination. It satisfactorily forecasted the algal fluctuations involving substantial peak and extreme values (average R2 > 0.900) and presented accurate judgment outcomes to their bloom levels with high accuracy > 95% on average. This work highlighted the utility of deep learning approaches in effective early-warning for algal blooms, and demonstrated an important direction for improving the adaptability of conventional deep learning approaches to the aquatic problems.

Early changes in the benthic community of a eutrophic lake following zebra mussel (Dreissena polymorpha) invasion

ABSTRACT In the context-dependent world of biological invasions, biologists understand few general patterns of spread and impact. One possible exception is the zebra mussel (Dreissena polymorpha), an invader that routinely restructures food webs through an ecosystem engineering process termed “benthification.” By efficiently consuming phytoplankton, zebra mussels can increase light penetration and nutrient concentrations in the benthos of a lake, thereby stimulating growth of benthic periphyton (phytobenthos) and macroinvertebrates (zoobenthos). Few studies monitor the response of these benthic communities to invasion. We documented early changes in phytobenthos and zoobenthos as zebra mussels invaded eutrophic Lake Mendota (Wisconsin, USA). From 2015 to 2018, the number of zebra mussel individuals reached densities >30 000 m−2 on hard substrates and 3000 m−2 in macrophyte beds. Community data showed classic signs of benthification, including 300% increases in (non-zebra mussel) zoobenthos and phytobenthos abundance on average across a depth gradient, including significant increases at depths where zebra mussels did colonize. Deep macrophyte biomass increased 900%, but water clarity showed no significant rapid increase. We speculate that nutrient enrichment may be more strongly responsible than increased light penetration for the benthic response of Lake Mendota. Continued integration of benthic production and processes into our study of lake ecosystems will be critical to understanding whole ecosystem function, especially as zebra mussels continue to “benthify” lakes within their invaded range.

Spatial and Temporal Variability of Dissolved Organic Matter Molecular Composition in a Stratified Eutrophic Lake

AbstractDissolved organic matter (DOM) is an intermediate between organic carbon formed by primary producers and carbon dioxide (CO2) produced through respiration, making it a key component of the carbon cycle in aquatic ecosystems. Its composition influences the routes of mineralization. Here, we evaluate DOM composition as a function of time and depth in Lake Mendota, a highly productive eutrophic lake that stratifies in warm months and is located in Madison, Wisconsin, USA. Dissolved organic carbon concentrations and optical properties are presented for 73 samples collected at a single location at varying depths within the water column from June to November. A subset of samples is analyzed by Fourier transform‐ion cyclotron resonance mass spectrometry (FT‐ICR MS) to investigate DOM composition at the molecular level. Temporally, increases in more oxidized formulas are observed in both the epilimnion and hypolimnion. At the surface, correlations between DOM formulas and both chlorophyll concentrations and light intensity show that photochemical reactions contribute to DOM oxidation. In the hypolimnion, redox conditions and interactions with sediments likely influence temporal compositional change. Our results show DOM composition varies with depth with more highly oxidized formulas identified deeper in the water column. However, DOM composition varies more temporally than by location within the water column. This work has implications for climate change as DOM photooxidation in lakes represents an understudied flux of CO2 to the atmosphere. Additionally, lake eutrophication is increasing due to warming temperatures and this data set yields detailed molecular information about DOM composition and processing in such lakes.

Impact of salinization on lake stratification and spring mixing

AbstractAnthropogenic freshwater salinization affects thousands of lakes worldwide, and yet little is known about how salt loading may shift timing of lake stratification and spring mixing in dimictic lakes. Here, we investigate the impact of salinization on mixing in Lakes Mendota and Monona, Wisconsin, by deploying under‐ice buoys to record salinity gradients, using an analytical approach to quantify salinity thresholds that prevent spring mixing, and running an ensemble of vertical one‐dimensional hydrodynamic lake models (GLM, GOTM, and Simstrat) to investigate the long‐term impact of winter salt loading on mixing and stratification. We found that spring salinity gradients between surface and bottom waters persist up to a month after ice‐off, and that theory predicts a salinity gradient of 1.3–1.4 g kg−1 would prevent spring mixing. Numerical models project that salt loading delays spring mixing and increases water column stability, with ramifications for oxygenation of bottom waters, biogeochemistry, and lake habitability.

Data inaccessibility at sub-county scale limits implementation of manuresheds.

The manureshed concept aims to rebalance surplus manure nutrients produced at animal feeding operations (sources) and the demands from nutrient-deficient croplands (sinks) to reduce negative environmental impacts and utilize nutrients more efficiently. Due to water quality implications, studies focused on this rebalancing have typically created domain boundaries that match a particular watershed. However, a majority of agricultural datasets that are used to inform these analyses-specifically, livestock populations-are only available at the county scale, which generally does not match watershed boundaries. The common method used to address this mismatch is to weight the county statistics based on the proportion of watershed area within the county. However, these straightforward assumptions imply that animal density is uniform across a county, which can be highly problematic, especially in an era of increasing concentration of livestock production on a smaller land area. We present a case study of the Lake Mendota watershed in south-central Wisconsin using both a typical county-based downscaled dataset as well as a more spatially explicit dataset of livestock counts from the Census of Agriculture that aggregates a set of zip codes that best matches the watershed boundary. This comparison reveals a substantial difference in estimated livestock numbers and their associated manure production that is due to a concentration of dairy operations within the watershed compared with the rest of the county. We argue that sub-county scale data need to become more available and integrated into nutrient and water quality management efforts so that manuresheds can be more effectively delineated and implemented.

Development of a sub-seasonal cyanobacteria prediction model by leveraging local and global scale predictors

In recent decades, cultural eutrophication of coastal waters and inland lakes around the world has contributed to a rapid expansion of potentially toxic cyanobacteria, threatening aquatic and human systems. For many locations, a complex array of physical, chemical, and biological variables leads to significant inter-annual variability of cyanobacteria biomass, modulated by local and large-scale climate phenomena. Currently, however, minimal information regarding expected summertime cyanobacteria biomass conditions is available prior to the season, limiting proactive management and preparedness strategies for lake and beach safety. To address this, sub-seasonal (two-month) cyanobacteria biomass prediction models are developed, drawing on pre-season predictors including stream discharge, phosphorus loads, a floating algae index, and large-scale sea-surface temperature regions, with an application to Lake Mendota in Wisconsin. A two-phase statistical modeling approach is adopted to reflect identified asymmetric relationships between predictors (drivers of inter-annual variability) and cyanobacteria biomass levels. The model illustrates promising performance overall, with particular skill in predicting above normal cyanobacteria biomass conditions which are of primary importance to lake and beach managers.

Lake thermal structure drives interannual variability in summer anoxia dynamics in a eutrophic lake over 37 years

Abstract. The concentration of oxygen is fundamental to lake water quality and ecosystem functioning through its control over habitat availability for organisms, redox reactions, and recycling of organic material. In many eutrophic lakes, oxygen depletion in the bottom layer (hypolimnion) occurs annually during summer stratification. The temporal and spatial extent of summer hypolimnetic anoxia is determined by interactions between the lake and its external drivers (e.g., catchment characteristics, nutrient loads, meteorology) as well as internal feedback mechanisms (e.g., organic matter recycling, phytoplankton blooms). How these drivers interact to control the evolution of lake anoxia over decadal timescales will determine, in part, the future lake water quality. In this study, we used a vertical one-dimensional hydrodynamic–ecological model (GLM-AED2) coupled with a calibrated hydrological catchment model (PIHM-Lake) to simulate the thermal and water quality dynamics of the eutrophic Lake Mendota (USA) over a 37 year period. The calibration and validation of the lake model consisted of a global sensitivity evaluation as well as the application of an optimization algorithm to improve the fit between observed and simulated data. We calculated stability indices (Schmidt stability, Birgean work, stored internal heat), identified spring mixing and summer stratification periods, and quantified the energy required for stratification and mixing. To qualify which external and internal factors were most important in driving the interannual variation in summer anoxia, we applied a random-forest classifier and multiple linear regressions to modeled ecosystem variables (e.g., stratification onset and offset, ice duration, gross primary production). Lake Mendota exhibited prolonged hypolimnetic anoxia each summer, lasting between 50–60 d. The summer heat budget, the timing of thermal stratification, and the gross primary production in the epilimnion prior to summer stratification were the most important predictors of the spatial and temporal extent of summer anoxia periods in Lake Mendota. Interannual variability in anoxia was largely driven by physical factors: earlier onset of thermal stratification in combination with a higher vertical stability strongly affected the duration and spatial extent of summer anoxia. A measured step change upward in summer anoxia in 2010 was unexplained by the GLM-AED2 model. Although the cause remains unknown, possible factors include invasion by the predacious zooplankton Bythotrephes longimanus. As the heat budget depended primarily on external meteorological conditions, the spatial and temporal extent of summer anoxia in Lake Mendota is likely to increase in the near future as a result of projected climate change in the region.

Climate and food web effects on the spring clear‐water phase in two north‐temperate eutrophic lakes

AbstractAlthough climate change has shifted the phenological timing of plankton in lakes, few studies have explicitly addressed the relative contributions of climate change and other factors, including planktivory and nutrient availability. The spring clear‐water phase is a period of marked reduction in algal biomass and increased water transparency observed in many lakes. Here, we quantified the phenological patterns in the start date, maximum date, duration, and magnitude of the clear‐water phase over 38 yr in Lakes Mendota and Monona, and examined the effects of water temperature, total phosphorus, and food web structure (proportion of large‐bodied Daphnia pulicaria and density of invasive Bythotrephes) and interactions between temperature and other predictors on these clear‐water phase metrics. We found that climate and food web structure affected the clear‐water phase, but the effects differed among the metrics. Higher water temperature led to earlier clear‐water phase start dates and maximum dates in both lakes. The proportion of D. pulicaria affected all clear‐water phase metrics in both lakes. When D. pulicaria proportion was higher, the clear‐water phase occurred earlier, lasted longer, and the water was clearer. Moreover, high Bythotrephes density delayed clear‐water phase start dates (both lakes), and decreased clear‐water phase duration (Lake Mendota) in the following year. These results suggest that variation in food web structure changes the full phenological dynamics of the clear‐water phase, while variation in climate condition affects clear‐water phase timing only. Our findings highlight the importance of large‐bodied grazers for managing water quality under climate change.

Predicting lake surface water phosphorus dynamics using process-guided machine learning

ABSTRACT Phosphorus (P) loading to lakes is degrading the quality and usability of water globally. Accurate predictions of lake P dynamics are needed to understand whole-ecosystem P budgets, as well as the consequences of changing lake P concentrations for water quality. However, complex biophysical processes within lakes, along with limited observational data, challenge our capacity to reproduce short-term lake dynamics needed for water quality predictions, as well as long-term dynamics needed to understand broad scale controls over lake P. Here we use an emerging paradigm in modeling, process-guided machine learning (PGML), to produce a phosphorus budget for Lake Mendota (Wisconsin, USA) and to accurately predict epilimnetic phosphorus over a time range of days to decades. In our implementation of PGML, which we term a Process-Guided Recurrent Neural Network (PGRNN), we combine a process-based model for lake P with a recurrent neural network, and then constrain the predictions with ecological principles. We test independently the process-based model, the recurrent neural network, and the PGRNN to evaluate the overall approach. The process-based model accounted for most of the observed pattern in lake P; however it missed the long-term trend in lake P and had the worst performance in predicting winter and summer P in surface waters. The root mean square error (RMSE) for the process-based model, the recurrent neural network, and the PGRNN was 33.0 μg P L−1, 22.7 μg P L−1, and 20.7 μg P L−1, respectively. All models performed better during summer, with RMSE values for the three models (same order) equal to 14.3 μg P L−1, 10.9 μg P L−1, and 10.7 μg P L−1. Although the PGRNN had only marginally better RMSE during summer, it had lower bias and reproduced long-term decreases in lake P missed by the other two models. For all seasons and all years, the recurrent neural network had better predictions than process alone, with root mean square error (RMSE) of 23.8 μg P L−1 and 28.0 μg P L−1, respectively. The output of PGRNN indicated that new processes related to water temperature, thermal stratification, and long term changes in external loads are needed to improve the process model. By using ecological knowledge, as well as the information content of complex data, PGML shows promise as a technique for accurate prediction in messy, real-world ecological dynamics, while providing valuable information that can improve our understanding of process.

Molecular markers obtained from draft genomic sequence data characterise an isolate of Oedogonium (Chlorophyceae) used for biomass applications

ABSTRACT Local isolates of the species-rich, fast-growing chlorophyte genus Oedogonium have been employed in outdoor biomass cultivation for industrial applications such as wastewater remediation, but such isolates may not express structural traits needed for definitive species identification by microscopy alone. We used maximum likelihood and Bayesian phylogenetic methods to evaluate molecular marker gene sequences derived from genomic sequences to characterise an industrial isolate, Oedogonium sp. Lake Mendota strain, which could not be reliably classified at species level using only morphology. Molecular markers evaluated were rbcL, 23S rDNA, 28S rDNA, 18S rDNA, and a sequence that included ITS1+5.8S rDNA+ITS2. All five markers distinguished the Lake Mendota strain from other Oedogonium accessions in GenBank. Using full genome sequences allowed the efficient acquisition of multiple types of marker sequences whose sequencing depth was known. In the absence of a classical species determination, these markers provided a taxonomic reference background for future genomic, biochemical, and other investigations of the local isolate. Local isolates are advantageous for outdoor industrial biomass cultivation applications because such isolates are already adapted to climatic conditions prevailing at cultivation sites. The results should encourage industrial use of local isolates of Oedogonium because these can be characterised by molecular markers even if structural features needed for accurate species determinations are not expressed.