Lake Erie Nutrient Loads, 1970–1980

The Buffalo District, Corps of Engineers’ Lake Erie Wastewater Management Study and Heidelberg College's Water Quality Laboratory supported a tributary water quality monitoring program from 1974 to 1980 of the major United States tributaries to Lake Erie. This program was designed to measure nutrient loads by monitoring concentration changes occurring in association with increased streamflow. Soluble orthophosphate loads, chloride loads, and silica loads decreased from 1970 to 1980. Nitrogen species were highly variable and increased over the period. Total phosphorus loads to Lake Erie have decreased during the period as a result of phosphorus removal at wastewater treatment plants. The effect of the phosphorus reductions can be seen in the lake concentrations and were predicted by a three-basin phosphorus budget model developed in the early 1970s. The results show that phosphorus removal programs are having the predicted effect on Lake Erie water quality.     

Updated phosphorus loads from Lake Huron and the Detroit River: Implications

The binational Great Lakes Water Quality Agreement (GLWQA) revised Lake Erie’s phosphorus (P) loading targets, including a 40% western and central basin total P (TP) load reduction from 2008 levels. Because the Detroit and Maumee River loads are roughly equal and contribute almost 90% of the TP load to the western basin and 54% to the whole lake, they have drawn significant policy attention. The Maumee is the primary driver of western basin harmful algal blooms, and the Detroit and Maumee rivers are key drivers of central basin hypoxia and overall western and central basin eutrophication. So, accurate estimates of those loads are particularly important. While daily measurements constrain Maumee load estimates, complex flows near the Detroit River mouth, along with varying Lake Erie water levels and corresponding back flows, make measurements there a questionable representation of loading conditions. Because of this, the Detroit River load is generally estimated by adding loads from Lake Huron to those from the watersheds of the St. Clair and Detroit rivers and Lake St. Clair. However, recent research showed the load from Lake Huron has been significantly underestimated. Herein, I compare different load estimates from Lake Huron and the Detroit River, justify revised higher loads from Lake Huron with a historical reconstruction, and discuss the implications for Lake Erie models and loading targets.     

Scenario-testing of agricultural best management practices in Lake Erie watersheds

Current research has shown that reductions in nonpoint nutrient loading are needed to reduce the incidence of harmful algal blooms and hypoxia in the western and central basins of Lake Erie. We used the Soil and Water Assessment Tool (SWAT) to test various sediment and nutrient load reduction strategies, including agricultural best management practice (BMP) implementation and source reduction in various combinations for six watersheds. These watersheds, in order of decreasing phosphorus loads, include the Maumee, Sandusky, Cuyahoga, Raisin, Grand, and Huron, and together comprise 53% of the binational Lake Erie Basin area. Hypothetical pristine nutrient yields, after eliminating all anthropogenic influences, were estimated to be an order of magnitude lower than current yields, underscoring the need for stronger management actions. However, cover crops, filter strips, and no-till BMPs, when implemented at levels considered feasible, were minimally effective, reducing sediment and nutrient yields by only 0–11% relative to current values. Sediment yield reduction was greater than nutrient yield reduction, and the greatest reduction was found when all three BMPs were implemented simultaneously. When BMPs were targeted at specific locations rather than at random, greater reduction in nutrient yields was achieved with BMPs placed in high source locations, whereas reduction in sediment yields was greatest when BMPs were located near the river outlet. Modest nutrient source reduction also was minimally effective in reducing yields. Our model results indicate that an “all-of-above” strategy is needed to substantially reduce nutrient yields and that BMPs should be much more widely implemented.     

Economic viability of proactive harmful algal bloom mitigation through attached algal growth

Harmful algal blooms (HABs) affect fresh and saltwater bodies around the world, causing a variety of damages to the surrounding communities. The primary driver of HABs is nutrient pollution. One novel HAB mitigation solution is to employ large-scale attached algal growth systems that consume nutrients from the water and prevent downstream nutrient accumulation and large scale HAB events. The feasibility of technology deployment is dependent on the economic viability. A model was created to estimate HAB-related economic damages, with or without non-point source nutrient reduction solutions, using Lake Erie as a case study. In the model, HAB severity is predicted based on nutrient loads and is tied to economic indicators to estimate the HAB-related economic losses. Results show that on average, Lake Erie communities lose $142 M (± $29 M) year^?1 from HABs without mitigation technology. Use of attached algal systems show an average net savings ranging between $29–42 M year^?1 from HAB mitigation depending on the system configuration. Attached algal systems show greater positive cash flows when compared to farm-based best management practices, thus warranting additional testing and consideration as a potential HAB mitigation strategy. Additional considerations included stochastic uncertainty and increasing dissolved reactive phosphorus concentrations, which increased the effectiveness of attached algal systems.     

Systemic,early-season Microcystis blooms in western Lake Erie and two of its major agricultural tributaries (Maumee and Sandusky rivers)

Recurrent, massive cyanobacterial blooms composed mainly of the genus Microcystis indicate a broad-scale re-eutrophication of Lake Erie. In the past, ameliorating eutrophication relied on intentionally decreasing point-source tributary nutrient, especially phosphorus, loads to the lake. However, recent research has shown that tributaries load not only nutrients but also bloom-levels of phytoplankton, including Microcystis. We built on this previous work by sampling earlier in the year and in much smaller tributaries in both the Maumee and Sandusky systems. We found Microcystis wet biomasses in these tributaries averaged 3.16 mg/L (± 0.59 mg/L, one standard error of the mean) in 2009 and 3.42 mg/L (± 0.55 mg/L) in 2010. Importantly, we found Microcystis in small ditches in March, much earlier than previously observed. Microcystis biomass did not directly correspond to measured phosphorus, chlorophyll, or phycocyanin concentrations likely reflecting complexities associated with lagged physiological responses and/or non-linear growth relationships. Consequently, our findings emphasize that Microcystis blooms form a more broad-scale problem than previously documented, occurring far upstream much earlier in the year.     

St. Clair-Detroit River system: Phosphorus mass balance and implications for Lake Erie load reduction,monitoring, and climate change

To support the 2012 Great Lakes Water Quality Agreement on reducing Lake Erie's phosphorus inputs, we integrated US and Canadian data to update and extend total phosphorus (TP) loads into and out of the St. Clair-Detroit River System for 1998–2016. The most significant changes were decreased loads from Lake Huron caused by mussel-induced oligotrophication of the lake, and decreased loads from upgraded Great Lakes Water Authority sewage treatment facilities in Detroit. By comparing Lake St. Clair inputs and outputs, we demonstrated that on average the lake retains 20% of its TP inputs. We also identified for the first time that loads from resuspended Lake Huron sediment were likely not always detected in US and Canadian monitoring programs due to mismatches in sampling and resuspension event frequencies, substantially underestimating the load. This additional load increased over time due to climate-induced decreases in Lake Huron ice cover and increases in winter storm frequencies. Given this more complete load inventory, we estimated that to reach a 40% reduction in the Detroit River TP load to Lake Erie, accounting for the missed load, point and non-point sources other than that coming from Lake Huron and the atmosphere would have to be reduced by at least 50%. We also discuss the implications of discontinuous monitoring efforts.     

Great Lakes total phosphorus revisited: 1. Loading analysis and update (1994–2008)

Phosphorus load estimates have been updated for all of the Great Lakes with an emphasis on lakes Superior, Michigan, Huron and Ontario for 1994–2008. Lake Erie phosphorus loads have been kept current with previous work and for completeness are reported here. A combination of modeling and data analysis is employed to evaluate whether target loads established by the Great Lakes Water Quality Agreement (GLWQA, 1978, Annex 3) have been and are currently being met. Data from federal, state, and provincial agencies were assembled and processed to yield annual estimates for all lakes and sources. A mass-balance model was used to check the consistency of loads and to estimate interlake transport. The analysis suggests that the GLWQA target loads have been consistently met for the main bodies of lakes Superior, Michigan and Huron. However, exceedances still persist for Saginaw Bay. For lakes Erie and Ontario, loadings are currently estimated to be at or just under the target (with some notable exceptions). Because interannual variability is high, the target loads have not been met consistently for the lower Great Lakes. The analysis also indicates that, because of decreasing TP concentrations in the lakes, interlake transport of TP has declined significantly since the mid-1970s. Thus, it is important that these changes be included in future assessments of compliance with TP load targets. Finally, detailed tables of the yearly (1994–2008) estimates are provided, as well as annual summaries by lake tributary basin (in Supplementary Information).     

Application of continuous turbidity sensors to supplement estimates of total phosphorus concentrations in the Grand River,Ontario, Canada

Accurate estimates of total phosphorus (TP) loadings to eastern Lake Erie are critical for developing load reduction targets and for determining if commitments are being met under the Great Lakes Water Quality Agreement, 2012 (GLWQA). Currently, loading calculations from Canadian priority tributaries are supported by year-round event-focused water quality sampling using automated samplers and laboratory water quality measurements. Here we evaluate the suitability of continuously-measured parameters, namely turbidity and flow, to supplement or enhance knowledge about TP concentrations in the Grand River, ON, by providing continuous data alongside event-focused sample measurements. A series of simple and multiple linear regression models were evaluated and compared with respect to their ability to predict TP water concentrations as a function of different combinations of explanatory variables. Explanatory variables included turbidity, flow, season and flow condition (i.e. hysteresis). The models that performed best explained 63–65% of the variation of TP which is comparable to surrogate model applications in the U. S and elsewhere. Additional model calibration work is needed due to gaps in turbidity data particularly during high flow events. We emphasize the need for continued automated, event-focused water quality sampling. However, provided that operational challenges are overcome, our results indicate that sensor-derived water quality parameters to predict TP concentrations is a promising technique that may supplement and improve nutrient loading estimates in the Grand River into the future and provides guidance for the utilization of this method in other tributaries.     

Philopatry and vagrancy of white bass (Morone chrysops) spawning in the Sandusky River: Evidence of metapopulation structure in western Lake Erie using otolith chemistry

Although natal homing and philopatry are well studied in anadromous salmon, few studies have investigated philopatric behavior in large, freshwater systems. In western Lake Erie, white bass (Morone chrysops) undergo seasonal spawning migrations from the open-water regions of Lake Erie to nearshore reef complexes and tributaries. The three primary spawning locations in Lake Erie are within 80 km of each other and are not separated by physical barriers. We used naturally occurring differences in otolith strontium concentrations among major spawning locations to address philopatry and vagrancy to the Sandusky River spawning location. Most individuals spawning in the Sandusky River were natal to this river (73%). No statistically significant differences in the extent of homing by sex or age of spawning were found, although a potential pattern of decreased homing with increased age of fish was observed. Given the proportion of vagrant individuals we found spawning in the Sandusky River (27%), it is unlikely that Lake Erie white bass spawning populations are genetically distinct. Furthermore, the white bass population in Lake Erie appears to be structured as a metapopulation, with non-philopatric individuals serving as a link between spawning populations.     

Nutrient concentrations and loadings in the St. Clair River–Detroit River Great Lakes Interconnecting Channel

Long-term (2001–2015) water quality monitoring data for the St. Clair River are presented with data from studies in the Detroit River in 2014 and 2015 to provide the most complete information available about nutrient concentrations and loadings in the Lake Huron–Lake Erie interconnecting corridor. Concentrations of total phosphorus (TP) in the St. Clair River have reflected declines in Lake Huron. We demonstrate that St. Clair River TP concentrations are higher than offshore Lake Huron values. The recent average (2014 and 2015) incoming TP load from the upstream Great Lakes is measured here to be 980 metric tonnes per annum (MTA), which is roughly three times greater than previous estimates. Significant TP load increases are also indicated along the St. Clair River. We treat the lower Detroit River as three channels to sample water quality as part of a two year monitoring campaign that included winter sampling and SRP in the parameter suite. We found concentrations of many parameters are higher near the shorelines, with the main Mid-River channel resembling water quality upstream measured at the mouth of the St. Clair River. Comparison with past estimates indicates both concentrations and loadings of TP have dramatically declined since 2007 in the Trenton Channel, while those in the Mid-River and in the Amherstburg Channel have remained similar or have possibly increased. The data demonstrate that the TP load exiting the mouth of the Detroit River into Lake Erie is currently in the range of 3740 (in 2014) to 2610 (2015) MTA.     

Re-eutrophication of Lake Erie: Correlations between tributary nutrient loads and phytoplankton biomass

Both abiotic and biotic explanations have been proposed to explain recent recurrent nuisance/harmful algal blooms in the western basin and central basin of Lake Erie. We used two long-term (> 10 years) datasets to test (1) whether Lake Erie total phytoplankton biomass and cyanobacterial biomass changed over time and (2) whether phytoplankton abundance was influenced by soluble reactive phosphorus or nitrate loading from agriculturally-dominated tributaries (Maumee and Sandusky rivers). We found that whereas total phytoplankton biomass decreased in Lake Erie's western basin from 1970 to 1987, it increased starting in the mid-1990s. Total phytoplankton and cyanobacterial seasonal (May–October) arithmetic mean wet-weight biomasses each significantly increased with increased water-year total soluble reactive phosphorus load from the Maumee River and the sum of soluble reactive phosphorus load from the Maumee and Sandusky rivers, but not for the Sandusky River alone during 1996–2006. During this same time period, neither total phytoplankton nor cyanobacterial biomass was correlated with nitrate load. Consequently, recently increased tributary soluble reactive phosphorus loads from the Maumee River likely contributed greatly to increased western basin and (central basin) cyanobacterial biomass and more frequent occurrence of harmful algal blooms. Managers thus must incorporate the form of and source location from which nutrients are delivered to lakes into their management plans, rather than solely considering total (both in terms of form and amount) nutrient load to the whole lake. Further, future studies need to address the relative contributions of not only external loads, but also sources of internal loading.     

Trends in Total Phosphorus in Canadian Near–Shore Waters of the Laurentian Great Lakes: 1976–1999

Beginning as early as 1976 at many locations, total phosphorus concentrations (TP) were measured weekly in samples collected year-round in the intake water of 18 municipal water treatment plants in Canadian (Ontario) waters of the Laurentian Great Lakes. No consistent long-term trends were evident at two north-shore Lake Superior sampling locations, but there were significant long-term declines in TP measured at all three Lake Huron locations; however, concentrations there have remained relatively constant during the past decade. Declines in TP averaging about 1 μg/L/yr during 1976 to 1990 were prevalent at lower Great Lakes sampling locations and by the early 1990s TP had declined to 15–25 μg/L in Lake Erie and 10–20 μg/L in Lake Ontario. Declines generally levelled out in Lake Ontario after 1990, but TP increased substantially at some Lake Erie locations in the late 1990s. Recent (1996 to 1999) total phosphorus concentrations in north-shore Lake Erie locations in the range of 20 to 30 μg/L were 2 to 3 times higher than at Lake Ontario near-shore locations in the 8 to 11 μg/L range. Rates of decline of TP were generally highest for the March–April period (−1.88, −1.61, and −1.34 μg/L/yr in Lakes Ontario, Erie, and Huron, respectively for 1976 to 1990). The March–April Lake Ontario near-shore rate of TP decline was nearly twice as high as that reported previously for off-shore Lake Ontario (attributed to proximity to P loading sources and to lower net sedimentation losses of P in the near-shore environment). There were substantial declines in chlorophyll-to-TP ratios and in the slopes and Y-intercepts of chlorophyll-TP regressions for both Lake Erie and Lake Ontario following the establishment of dreissenid mussels.     

Tributary and Point Source Total Phosphorus Loading to Lake Erie

Total phosphorus data from tributaries, municipal and industrial sources, connecting channels, and atmospheric deposition have been evaluated to estimate the loading of phosphorus to Lake Erie. The annual total load from these sources has declined over the study period from a high of 27,944 tonnes in 1968 to a low of 10,452 tonnes in 1981. Of all sources the Detroit River represented the single largest component of the total load. Reductions in the Detroit River annual load, which is a complement of both U.S. and Canadian loads, from 17,822 tonnes to 3,541 tonnes was principally responsible for the overall total phosphorus loading decrease. These decreases were attributed principally to the phosphorus abatement program on municipal sources and the restriction of phosphate concentrations in detergents. In 1978 the contribution to the total load from tributary sources for the first time exceeded the load of phosphorus from the Detroit River, with U.S. tributary sources representing approximately 47 percent of the total lake load.     

A hydrodynamic approach to modeling phosphorus distribution in Lake Erie

The purpose of this paper is to show how a high-resolution numerical circulation model of Lake Erie can be used to gain insight into the spatial and temporal variability of phosphorus (and by inference, other components of the lower food web) in the lake. The computer model simulates the detailed spatial and temporal distribution of total phosphorus in Lake Erie during 1994 based on tributary and atmospheric loading, hydrodynamic transport, and basin-dependent net apparent settling. Phosphorus loads to the lake in 1994 were relatively low, about 30% lower than the average loads for the past 30 years. Results of the model simulations are presented in terms of maps of 1) annually averaged phosphorus concentration, 2) temporal variability of phosphorus concentration, and 3) relative contribution of annual phosphorus load from specific tributaries. Model results illustrate that significant nearshore to offshore gradients occur in the vicinity of tributary mouths and their along-shore plumes. For instance, the annually averaged phosphorus concentration can vary by a factor of 10 from one end of the lake to the other. Phosphorus levels at some points in the lake can change by a factor of 10 in a matter of hours. Variance in phosphorus levels is up to 100 times higher near major tributary mouths than it is in offshore waters. The model is also used to estimate the spatial distribution of phosphorus variability and to produce maps of the relative contribution of individual tributaries to the annual average concentration at each point in the lake.     

Detroit River load estimation; the need for a new monitoring approach

The Great Lakes Water Quality Agreement (GLWQA) established new Lake Erie phosphorus loading targets, including a 40% total phosphorus load reduction to its western and central basins. The Detroit and Maumee rivers’ loads are roughly equal and contribute about 90% of the load to the western basin and 54% to the whole lake. They are key drivers of central basin hypoxia and western basin algal production. So, accurate estimates of the Detroit River load are important. Direct measurement of that load near its mouth is difficult due to requiring real-time knowledge of flows around islands and the influence of Lake Erie’s seiches. Consequently, most estimates sum the loads to the St. Clair/Detroit River system. But this approach is complicated by uncertainties in the Lake Huron load and load retention in Lake St. Clair. Routine GLWQA reassessments will confirm or adjust over time the goals, loading targets, and approaches based on evolving information. So, there is a need to improve monitoring approaches that ensure accurate Detroit River loads. New approaches should take into account both the characteristics of this dynamic connecting channel and the uses of monitoring results: 1) determining the Detroit River loads to drive models, develop mass balances, set load reduction targets, and track progress; and 2) assessing the sources and processing of the loads to help guide reduction strategies. Herein, we review temporal and spatial variability in the St. Clair/Detroit River system, and suggest adjustments to monitoring that address those variabilities and both uses.     

Status of the major aquaculture carps of China in the Laurentian Great Lakes Basin

There is concern of economic and environmental damage occuring if any of the four major aquacultured carp species of China, black carp Mylopharyngodon piceus, bighead carp Hypophthalmichthys nobilis, silver carp H. molitrix, or grass carp Ctenopharyngodon idella, were to establish in the Laurentian Great Lakes. All four are reproducing in the Mississippi River Basin. We review the status of these fishes in relation to the Great Lakes and their proximity to pathways into the Great Lakes, based on captures and collections of eggs and larvae. No black carp have been captured in the Great Lakes Basin. One silver carp and one bighead carp were captured within the Chicago Area Waterway System, on the Great Lakes side of electric barriers designed to keep carp from entering the Great Lakes from the greater Mississippi River Basin. Three bighead carp were captured in Lake Erie, none later than the year 2000. By December 2019, at least 650 grass carps had been captured in the Great Lakes Basin, most in western Lake Erie, but none in Lake Superior. Grass carp reproduction has been documented in the Sandusky and Maumee rivers in Ohio, tributaries of Lake Erie. We also discuss environmental DNA (eDNA) results as an early detection and monitoring tool for bighead and silver carps. Detection of eDNA does not necessarily indicate presence of live fish, but bigheaded carp eDNA has been detected on the Great Lakes side of the barriers and in a small proportion of samples from the western basin of Lake Erie.     

Modelling streamflow and phosphorus fluxes in the Lake of the Woods watershed

This study provides an evaluation of streamflow and the spatial and temporal variability of phosphorus (P) fluxes for the transboundary Lake of the Woods (LoW) watershed using the Canadian version of the Soil and Water Assessment Tool (CanSWAT). The model calibration and validation generally indicate good performance for the simulated flow, especially for the Rainy River, the main tributary to LoW, while the sediment and nutrient calibration performance was satisfactory. Model results indicated Rainy River is the primary source of total phosphorus (TP), contributing about 88% of the external non-point source (NPS) and point source P loads to LoW, with the majority being NPS. Simulated TP loads varied seasonally with over approximately 60% occurring during the spring period and varied spatially across the LoW watershed. TP yields tended to be lower upstream of Rainy Lake in the Precambrian Shield (a.k.a., Canadian Shield) and higher downstream of Rainy Lake in the Glacial Lake Agassiz lakebed, particularly in the Lower Rainy and Little Fork sub-watersheds. Point sources along the Rainy River constituted the largest anthropogenic TP source. Tributary P loads estimated by the model were also used in a simple lake mass-balance model that suggested 32–46% of TP load to LoW was retained within the lake.     

Urban total phosphorus loads to the St. Clair-Detroit River System

The St. Clair-Detroit River System watershed is a large, binational watershed draining into the connecting channel between lakes Huron and Erie. In addition to extensive agricultural lands, it contains large urban areas that discharge phosphorus from point source facilities, runoff of impervious surfaces, and overflows of combined sewers. To help guide actions to reduce phosphorus input to Lake Erie, we analyzed the spatial and temporal dynamics of loads from the three largest urban areas in the watershed (southeast Michigan; Windsor, Ontario; and London, Ontario), and used a previously calibrated storm water management model (SWMM) to explore options for reducing loads around metro Detroit. Point sources in these three urban areas contribute, on average, 81% of the total urban load and 19% of the Detroit River’s total phosphorus (TP) load to Lake Erie, while combined sewer overflows and runoff both contribute about 10% each to the urban load and about 2.5% each to the Detroit River’s load to Lake Erie. Most of the urban load (56%) comes from a single point source, the wastewater treatment facility in Detroit; however, TP loads from that facility have decreased by about 51% since 2008 due to improvements in wastewater treatment. Model simulations suggest that increasing pervious land area or implementing green infrastructure could help reduce combined sewer overflows in certain upper portions of the metro Detroit sewer system, but reductions were much less expressed for wet-weather discharge from the system.     

Great Lakes chloride trends: Long-term mass balance and loading analysis

Surveillance data collected over the past 150 years are compiled and analyzed to identify chloride trends in the Laurentian Great Lakes. These data indicate that chloride levels started rising in the mid-19th century and began accelerating in the early twentieth century. Lake Superior's and Lake Michigan's concentrations have continued to increase steadily and currently stand at their maximum recorded levels. In contrast, lakes Huron, Erie and Ontario reached peak levels between 1965 and 1975, but then began to decline. However, recent data indicate that the chloride concentrations in these lakes are now increasing again. Because loading data are not readily available, a mass-balance model is employed to estimate the chloride inputs required to account for the concentration trends. This inverse analysis yields computed load reductions that are consistent with reported industrial load reductions during the last three decades of the 20th century. Hence, it appears that the improvements were for the most part attributable to industrial controls. The model is also used to predict that if loads are held fixed at 2006 levels, concentrations in all lakes will continue to increase with the most dramatic rise occurring in Lake Michigan which will ultimately approach the level of Lake Erie.