Physical measurements and nearshore nested hydrodynamic modeling for Lake Ontario nearshore nutrient study
Highlights
► Data are presented for three nearshore areas sampled in Lake Ontario in 2008. ► Nested hydrodynamic models are developed, which are linked with a whole-lake model. ► Model results and data are evaluated for potential use in nearshore management. ► Resolution on the order of tens of meters is needed to simulate thermal gradients. ► Initial conditions remain in the model for much longer than previously thought.
Introduction
The first Great Lakes Water Quality Agreement (GLWQA), signed by the United States and Canada in 1972, was a response to a wave of scientific and public concern about pollution and significant water quality problems in Lake Erie (NGOs, 2007). The main goal of the GLWQA was to reduce eutrophication by decreasing phosphorus levels. Average open lake phosphorus concentrations in Lake Ontario were successfully reduced from a peak of about 25 μg/L in 1971 to below the target concentration of 10 μg/L by the mid 1980s. Offshore phosphorus levels have maintained a declining trend, and concentrations are now approximately 5–7 μg/L (EPA, Environmental Protection Agency, 2006). Despite this success, resurgence of the benthic filamentous algae Cladophora glomerata, which was widespread in shallower regions of the lakes in the 1960s and 1970s, and other nuisance algae problems indicate nutrient enrichment in nearshore areas (EPA, Environmental Protection Agency, 2006). Specific studies documenting Cladophora blooms and associated problems in Lakes Michigan, Erie and Ontario are referenced by Higgins et al. (2012), and sloughed Cladophora has led to frequent beach closings at two of our study areas—Oak Orchard and Rochester (see below). Anecdotally, Cladophora has also been tied to cooling water intake clogging at power plants along the southern shore of Lake Ontario.
A key component in setting target phosphorus loadings to achieve water quality objectives was the application of phosphorus mass balance models. These models ranged in complexity from primarily empirical loading relationships (e.g., Vollenweider, 1972) to more sophisticated process-based models (Thomann et al., 1975, Thomann et al., 1976), or combinations of empirical and mechanistic approaches (e.g., Chapra, 1977, Chapra, 1980). Models were generally applied with lake-wide or basin-wide resolution, and did not consider differences between nearshore and offshore locations. In order to examine in better detail water quality in nearshore areas, particularly in contrast with offshore waters, models are needed with appropriate spatial and process resolutions. Water quality models require advection and diffusion values, which are derived from a hydrodynamic model. A major challenge then is to provide the hydrodynamic forcing at a sufficiently fine scale in a nearshore area of interest, while incorporating whole-lake processes that may affect the nearshore, such as upwelling and downwelling, and at the same time avoiding excessive computational burden by having an overly fine scale for the whole lake.
These requirements have led to the use of nested modeling, where a finer-scale grid for a region of interest is embedded within a coarser scale grid for the entire lake. For example, Murthy et al. (1986) applied a semi-empirical method to predict pollutant transport problems in the Pickering area along the northern shore of Lake Ontario. This same area is the site of a more mechanistic model application coupling a 2 km whole-lake grid with a 100 m finer grid, simulating both hydrodynamics and water quality, with an emphasis on nuisance algae (Leon et al., 2012). Shen et al. (1995) used several different resolutions to simulate pollutant dispersion from a small creek in the Toronto waterfront area under isothermal or stratified conditions, using an iterative scheme to match boundary conditions between the nearshore and whole-lake models. Hayashida et al. (2000) used a variable grid finite element approach to enable fine-scale analysis of flow patterns in the Niagara River outlet region, with grid dimensions varying from about 50 m to several kilometers. However, their model was vertically integrated, and therefore not able to simulate a full range of (three-dimensional) motions in the lake. More recently, Sheng and Rao (2006) applied a nested ocean circulation model to simulate circulation and temperatures in Georgian Bay, Lake Huron, reporting reasonable comparisons with observations.
The goal of the present work was to test the application of a well-known three-dimensional hydrodynamic code to the Great Lakes (the Princeton Ocean Model, or POM), using a nested model approach. This application is planned to serve as a basic component of a more comprehensive modeling framework to simulate nutrients, contaminants, sediment and lower food web dynamics to support management decision making for nearshore regions in Lake Ontario. It is similar to the approach used by Leon et al. (2012), except it uses a public domain model, to allow greater flexibility over a range of applications. The approach here is based on the nested modeling capability developed recently at the Great Lakes Environmental Research Laboratory (GLERL, D. Schwab, pers. comm.), and takes advantage of a field data set collected in 2008 as part of the Lake Ontario Nearshore Nutrient Study (LONNS), of which this study was a part (Higgins et al., 2012).
The underlying conceptual model guiding the LONNS is the nearshore shunt hypothesis (Hecky et al., 2004), which suggests nearshore nutrient sequestration and cycling as causes of nearshore eutrophication. These processes may be enhanced by nearshore circulation phenomena such as the thermal bar (Rodgers and Anderson, 1963, Rodgers, 1966, Rodgers, 1968, Rodgers and Sato, 1970) and upwelling and downwelling (Rao and Schwab, 2007). The spring thermal bar, for example, usually coincides with a period of relatively large runoffs and nutrient loadings due to urban, agricultural, and other sources. Cross-frontal exchange coefficients are reduced in the thermal bar region, with values typically an order of magnitude smaller than alongshore exchange coefficients (Gbah and Murthy, 1998). Rao et al. (2004) also suggested that mixing across the bar was reduced, at least when the bar was relatively close to shore. Upwelling and downwelling during stratified periods also play a significant role for nearshore biological and chemical processes, serving as major transport pathways moving material on- and off-shore (Rao and Schwab, 2007). All these processes alone, however, do not explain the present nearshore water quality conditions—the difference between current conditions and those with which the lake ecology evolved is the relatively recent introduction of invasive mussels. To investigate the cause of these problems, models are needed that can be used to interpret physical influences on biological and chemical processes on a finer spatial scale than is typical in whole-lake models.
The LONNS was initiated to investigate possible causes of observed water quality differences between nearshore and offshore waters in Lake Ontario, and to better understand reasons for the resurgence of Cladophora in the nearshore region. The field component of this study focused on three nearshore areas along the New York shoreline: (1) Oak Orchard Creek, (2) Rochester Embayment, and (3) Sandy Creek (also referred to as Mexico Bay) (Fig. 1). Sampling polygons with dimensions of approximately 5 km × 20 km were defined for each of these areas, and sampling cruises were conducted in late May to early June, early to mid-August, and late September to early October, to capture conditions representative of late spring, mid-summer, and early fall. The present study focuses on the physical data, which include surface temperatures obtained continuously at a depth of 1 m during the sampling cruises, depth-variable temperatures from CTD casts at specified locations, and acoustic Doppler profiler (ADP) velocity profiles taken at approximately the same locations as the CTD casts. NOAA buoy data and satellite-derived surface temperatures available from NOAA Coastwatch were also used for additional comparisons with modeled temperatures.
Section snippets
Field methods and site descriptions
For this study the nominal working definition of the nearshore was taken as the region between the shoreline and the 30 m isobath. Various definitions for the nearshore region are possible, depending on specific interests of a given study, and these include average depth of the thermocline (e.g., Edsall and Charlton, 1997), wave energy dissipation depth, photic depth, or substrate characteristics (EC&EPA, Environment Canada and the U.S. Environmental Protection Agency, 2009). The 30 m isobath
Model development and application
The POM solves the three-dimensional, time-dependent equations for continuity, Reynolds-averaged momentum, temperature, and salinity, and includes Smagorinski diffusivity for horizontal mixing and a Mellor–Yamada turbulence closure scheme for vertical mixing (Mellor and Yamada, 1982, Blumberg and Mellor, 1987). Vertical resolution is described using sigma coordinates, which divide the total depth into an equal number of layers at all locations in the model domain, with layer depth proportional
Whole-lake model
Because of its importance in driving the nested regions, results from the whole-lake model were first examined. Fig. 2 displays comparisons of model results with near-surface temperatures recorded at the three lake buoys, buoy 45012 in the central part of the lake (Fig. 2a), buoy 45135 in the eastern basin (Fig. 2b), and buoy 45139 in the western basin (Fig. 2c)—see Fig. 1 for buoy locations. Table 1 lists a comparison of measured and computed monthly average temperatures at these three buoys.
Discussion
While demonstrating the feasibility of using a nested approach for hydrodynamic calculations, this study also shows that an even finer resolution would be needed to reproduce certain features such as observed thermal fronts. Further analysis is needed to better understand the physical mechanisms that maintain these fronts, especially in the absence of thermal bar dynamics. A finer scale also would be needed to accommodate tributary inflows, with a width that may be on the order of tens of
Conclusion
The model and data comparisons for the nested regions are promising, but three main issues should be addressed in developing this work further: (1) finer resolution, probably on the order of one to several tens of meters, is needed to resolve certain observations such as thermal fronts and local inflows; (2) the impact of different initial conditions should be explored in greater detail; and (3) alternative mixing formulations should be evaluated to better simulate mixing in areas very close to
Acknowledgments
We thank David Schwab and Gregory Lang at the Great Lakes Environmental Research Laboratory for their invaluable assistance in teaching us how to use the nested POM modeling framework, and in helping us obtain meteorological and bathymetric data needed for the model runs. We are also indebted to Scott Brown for his ideas and numerous discussions during the development of this modeling approach. Ted Lewis created and archived the data base for the LONNS project, which was managed by Joseph
References (38)
- et al.
Mean circulation in the Great Lakes
J. Great Lakes Res.
(1999) Simulation of recent and projected total phosphorus trends in Lake Ontario
J. Great Lakes Res.
(1980)- et al.
Urban influences on Cladophora blooms in Lake Ontario
J. Great Lakes Res.
(2012) - et al.
Nested 3D modeling of the spatial dynamics of nutrients and phytoplankton in a Lake Ontario nearshore zone
J. Great Lakes Res.
(2012) - et al.
A semi-Lagrangian study of circulation and transport in Lake Ontario
J. Great Lakes Res.
(2007) - et al.
Transport and mixing between the coastal and offshore waters in the Great Lakes: a review
J. Great Lakes Res.
(2007) - et al.
A three-dimensional nested hydrodynamic/pollutant transport simulation model for the nearshore areas of Lake Ontario
J. Great Lakes Res.
(1995) - et al.
Circulation and thermal structure in Lake Huron and Georgian Bay: application of a nested-grid hydrodynamic model
Cont. Shelf Res.
(2006) - et al.
Surface temperature fronts in the Great Lakes
J. Great Lakes Res.
(1998) - et al.
IFYGL—The International Field Year for the Great Lakes
(1981)
Radiation boundary conditions for wave-like equations
Comm. Pure Appl. Math.
Modeling circulation and thermal structure in Lake Michigan: annual cycle and interannual variability
J. Geophys. Res.
A description of a three-dimensional coastal ocean circulation model
Total phosphorus model for the Great Lakes
J. Environ. Eng. Div. ASCE
ARL models of global solar radiation
Nearshore areas of the Great Lakes 2009
State of the Lakes Ecosystem Conference 2008—Background Paper, EPA 905-R-09-013, Cat No. En164-19/2009E-PDF
Nearshore waters of the Great Lakes
State of the Lakes Ecosystem Conference 1996 Background Paper
Lake Ontario Lakewide Management Plan Status
3.3.2 Nutrients in open waters, 4.5.4 Eutrophication or Undesirable Algae, April 22, 2006
Cited by (11)
A modeling study to determine the contribution of interbasin versus intrabasin phosphorus loads on the southwestern nearshore of Lake Ontario
2022, Journal of Great Lakes ResearchMass balance analysis and calculation of wind effects on heat fluxes and water temperature in a large lake
2018, Journal of Great Lakes ResearchCitation Excerpt :Velocities were set to zero. It is expected that the initial condition is “washed out” within about one to two months (Atkinson et al., 2012), so the specified initial conditions should not affect results in summer. The initial water elevation was set as 0.75 m, relative to the Lake Ontario low water datum (74.2 m IGLD85), based on observation on April 1st 2008.
Utilizing intensive monitoring and simulations for identifying sources of phosphorus and sediment and for directing, siting, and assessing BMPs: The Genesee River example
2015, Journal of Great Lakes ResearchCitation Excerpt :In the future, this may take the form of integrated land-use models, such as the Soil and Water Assessment Tool (SWAT) with nearshore models linked to the offshore and the utilization of these models to direct and target remediation efforts in the watershed. Both Atkinson et al. (2012) and Prakash et al. (2007) have begun development of an integrated coastal ecosystem model (ICEM) to synthesize data and evaluate system responses to multiple stressors (physical, chemical, and biological) acting in concert, as well as to provide predictive capabilities for further hypothesis formulation and testing. With the development of ICEMs, results from the GRSWAT model could be used to test “what if” scenarios evaluating various BMPs to determine those most effective in reducing P loads to the Rochester Embayment and the nearshore zone of Lake Ontario.
The Lake Ontario Nearshore Study: Introduction and summary
2012, Journal of Great Lakes ResearchCitation Excerpt :The model was able, without special parameter tuning, to usefully predict major water quality variables (temperature, light, TP, SRP) that would be important to nuisance algae growth. In a similar approach, Atkinson et al. (2012) developed a nested hydrodynamic model using a public domain model code (the Princeton Ocean Model or POM) that has been used for many years in Great Lakes applications to consider physical transport processes in each of the three nearshore sites along the New York coast. As with the north shore work, the higher resolution models demonstrate greater detail in flow and temperature structure, providing output on a scale more comparable to the observed data.
Physical and chemical characteristics of the nearshore zone of Lake Ontario
2012, Journal of Great Lakes ResearchCitation Excerpt :There is reason to believe that this occurs throughout the spring, summer, and autumn. The extension of the nearshore zone into the open waters of a large lake varies with season, thermal bar development, proximity of rivers, and likely with locally induced wind-generated currents (Atkinson et al., 2012). The high variability in nutrient levels (Figs. 5 and 6) observed within the nearshore zone is a result of several factors including: the degree of wind and wave exposure, land use in the watershed, river discharge and nutrient loading, shoreline development along the coast, thermal bar development, the occurrence of zebra mussels, etc.
- 1
Tel.: + 1 716 286 8251; fax: + 1 716 286 8254.