Satellite-based sensors provide synoptic measurements of surface parameters useful in detecting physical and biological conditions of the Great Lakes. Satellite surface temperature measurements using infrared spectra are compatible with buoy measurements and the time series now covers more than two decades in length. This time-series tracks seasonal warming patterns and localized upwelling events. The use of visible spectra for remote sensing of water clarity and particle composition is improving with new algorithms to separate chlorophyll-a, inorganic particles, and dissolved organic matter. The use of satellites to measure these variables holds promise for future quantification of phytoplankton production, calcite precipitation (whiting), and suspended sediment from rivers and resuspension events. Satellite imagery has also been useful for interpreting ship-collected data such as those associated with the bi-national Lake Ontario Lower Foodweb Assessment (LOLA) in 2003.

Introduction

The launch of the Landsat-1 satellite in July 1972 and the NOAA-2 satellite in October 1972 added a new view of spatial patterns of the Great Lakes that greatly expanded the data provided by extensive ship-based sampling efforts such as the 1972 International Field Year of the Great Lakes (IFYGL) sampling of Lake Ontario. Infrared sensors (detecting 10,500–12,500 nm wavelength band) on the NOAA-2 satellite detected gradients of surface water temperature, enabling the mapping of the extensive upwelling of cold water along the coast of Lake Michigan in the summer of 1973 (Strong et al., 1974). This seasonal phenomenon had been observed for many of the Great Lakes using in situ thermal measurements on ships, buoys, and docks, but never had such a instantaneous view captured its full spatial extent and dynamic nature. Landsat-1 sensors measured the 500–600 nm wavelength band (green) and differentiated green warm water masses adjacent to clearer upwelled water. Stumpf and Strong (1975) used the interplay of these water masses to chart surface currents in a manner consistent with the results of drifter-based studies. In situ data later confirmed that the greenish hue of the warm water was the result of the precipitation of calcium carbonate (known as “whiting events”). Satellite images collected from 1972–1975 observed whiting events in Lakes Ontario, Erie, and Michigan, but not Lakes Superior, nor Huron (Strong and Eadie, 1978).

This early success in the use of satellite imagery for the Great Lakes encouraged other applications as sensor technology improved. In operation from 1978 to 1986, the Coastal Zone Color Scanner (CZCS) added additional bands of the visible spectrum at a finer resolution. The CZCS sensor collected four channels in the visible spectrum including blue (443 nm), blue-green (520 nm), green (550 nm), and red (670 nm) with a bandwidth of 20 nm. An initial hope in using the CZCS data for the Great Lakes was to quantify phytoplankton production in Lake Michigan (Mortimer, 1988). This goal had been attained in open ocean systems using relatively simple algorithms (e.g. ratio of blue-green and green bands) to calculate the concentration of the phytoplankton pigment chlorophyll-a (chl a) (O'Reilly et al., 1998). However, large inland seas such as the Great Lakes present a much more complex test than the open ocean. Open ocean waters (known as Case 1) contain primarily phytoplankton pigments, while inland and coastal waters (known as Case 2) contain color producing agents (CPAs) other than phytoplankton including inorganic suspended material and colored dissolved organic matter (CDOM) (Morel and Prieur, 1977). In Case 2 conditions, chl a concentration cannot be calculated without considering the additive effects of the other two components. Great Lakes waters at times have high concentrations of sediment and precipitated calcium carbonate that are highly reflective. Atmospheric corrections are also more complex for lakes because of scattering in the near infrared bands.

Despite its disappointing application to chl a concentration, the CZCS images were very effective in documenting resuspension of sediment in southern Lake Michigan during the development of the spring thermocline (Mortimer, 1988). This observation led to one of the best examples of the use of satellite imagery in a Great Lakes research project. The Episodic Events-Great Lakes Experiment (EEGLE) revisited the phenomenon and coordinated satellite imagery with shipboard and buoy-based measurements from 1998–2000 (Chen et al., 2004). Satellite images containing regions of high remote sensing reflectance (RSR) of sediment plumes and buoy-based wind measurements were key tools in observing the phenomenon on March 1998 and March 1999. The study confirmed that the resuspension events in southern Lake Michigan were caused by southward episodic wind events having a period of 5–7 days. Satellite images captured spatial differences between the years- in 1998 sediment plumes were present on both the eastern and western shores, while in 1999 plumes were restricted to a region on the southwest shore. In models, the light limiting conditions of sediment plumes led to a shift from phytoplankton production to enhanced heterotrophic production (Chen et al., 2004).

Satellite imagery may help in monitoring of Great Lakes ecosystems. Predictions of global climate change have important ramifications. Surface waters of the Great Lakes are expected to warm in the future with a longer period of summer stratification. For the time period 1963–2001, maximum ice coverage was lowest in the winters of 1998–2001 for four of the five Great Lakes (Assel et al., 2003). Buoy data indicate that summer surface waters of Lake Superior warmed by 2.5°C during the period from 1979–2006 and the date of initial stratification advanced by one half day per year (Austin and Colman, 2007). Times-series also suggest that the surface of Lake Michigan and Lake Huron has similarly warmed.

Warmer surface waters of the Great Lakes are expected to alter the timing of phytoplankton blooms. Decreased physical mixing could lead to lower primary productivity and shifts from diatoms to blue-green algae. Invasive species also impact phytoplankton density as seen in grazing by dreissenid mussels. Lake-wide changes in temperature and phytoplankton productivity are difficult to detect using traditional shipboard measurements because of spatial and temporal data gaps. This paper reviews recent advances in the use of satellite-based sensors in monitoring the Great Lakes and applies such imagery to a recent shipboard monitoring effort.

Satellite Based Surface Temperature Products

Satellites measure surface water temperature via infrared radiation from the lake surface. A continuous series of satellites equipped with an AVHRR (Very High Resolution Radiometer) have been launched since 1972 with the latest deployment of the NOAA-19 satellite in 2009. Temperature measurements from satellites agree with ground truth buoys within 0.5°C on average (Schwab et al., 1999). Lake-wide averages for surface temperature are possible because satellites simultaneously measure the entire lake surface with a 1 km resolution. The primary weaknesses of satellite data are data gaps from cloud cover, coastal boundaries, and flyover paths.

The Great Lakes Surface Environmental Analysis (GLSEA) is an automated system for calculating the average surface temperature for each of the Great Lakes from satellite data (Schwab et al., 1999). It was developed at NOAA's Great Lakes Environmental Research Laboratory (GLERL). The analysis interpolates across periods of cloud coverage for each 1 km pixel. An initial analysis described surface temperature variation for each lake over a six-year period (1992–1997) and compared it to an earlier AVHRR-based climatology (Schneider et al., 1993). GLSEA temperatures are continually posted on the website http://coastwatch.glerl.noaa.gov and include seasonal animations and averages (Figure 1).

Figure 1.

Satellite derived average surface temperatures for Lake Ontario in 2003. Data from Great Lakes Surface Environmental Analysis (GLSEA) (access http://coastwatch.glerl.noaa.gov).

Figure 1.

Satellite derived average surface temperatures for Lake Ontario in 2003. Data from Great Lakes Surface Environmental Analysis (GLSEA) (access http://coastwatch.glerl.noaa.gov).

Several temperature-based data products have been derived from satellite data. An automated method for detecting temperature fronts using gradient analysis was developed by Ullman et al., 1998. This method delineates features such as thermal bars (warm coastal water) and upwelling (cold coastal water) events. Coastal upwelling events can also be identified as a cold tail in the lake-wide temperature distribution (Plattner et al., 2006). The number of pixels represented by a cold tails defines the spatial extent of upwelling events. Seasonal data can be used to quantify the frequency of occurrence of upwelling for particular coastal sites. Both of these temperature-based tools have been useful in interpreting biological distributions. For example, catch rates of steelhead in Lake Michigan were highest in regions of temperature fronts defined by satellite data (Hook et al., 2004). Upwelling frequency was a useful parameter for describing the abundance of living dreissenid mussels in Lake Ontario (Wilson et al., 2006).

Use of the Visible Spectrum in the Great Lakes

The Sea-viewing Wide Field-of-View Sensor (SeaWiFS) was launched in 1997 and the Moderate Resolution Imaging Spectrometer (MODIS) was added in 2002. These instruments improved on the CZCS sensor for application in Case II waters such as the Great Lakes. For example, the MODIS sensor has 9 bands in the visible spectrum and includes two narrow bands (673–683 nm and 662–672 nm) for specifically detecting fluorescence of chl a (International Ocean-Colour Coordinating Group (IOCCG), 2000). However, improvements have not alleviated the problem of separating the multiple color components including: chl a, suspended sediment, and dissolved organic matter.

Secchi depth is influenced by all three components and can be generally inferred from water-leaving radiance in the green portion of the visible spectrum (555 nm) detected by satellite sensors. Binding et al. (2007) utilized this approach to look at changes in secchi depth using a time series of CZCS (1979–1985) and SeaWiFS (1998–2005) data. They found that spring secchi disk depths have doubled in Lake Ontario and in the eastern basin of Lake Erie, yet decreased by half in the western basin of Lake Erie. The expansion of the photic zone is consistent with long-term ground-based studies and consistent with lower nutrient loading and the filter feeding of invasive zebra mussels. The seasonal pattern of secchi depth includes a late-summer minimum caused by whiting events. Binding et al. (2007) identified a one-month delay in whiting events in more recent SeaWiFS imagery relative to CZCS imagery in Lakes Ontario and Erie. The occurrence of whiting events also decreased in Lake Ontario, perhaps due to calcium uptake by dreissenid mussels.

Despite this success, the quantification of chl a concentration in the Great Lakes by satellites still faces several obstacles. Chl a, inorganic suspended material, and dissolved organic matter all have overlapping absorption and scattering spectra. An important step in separating chl a from inorganic suspended material and dissolved organic matter is to quantify the three component contributions simultaneously (IOCCG, 2000). Soon after the deployment of the CZCS, Bukata et al. (1981a, b) developed an optical model for Lake Ontario based on in situ measurements. He successfully used inverse modeling to quantify the three components from surface reflectance (Bukata et al., 1991). However, the problem of atmospheric correction and computational complexity were two significant barriers in applying this model to satellite data. The development of new sensors (such as SeaWiFS and MODIS) that detect more wavelength bands at a finer resolution is a key step in developing more effective algorithms (IOCCG, 2000). Measurement of the optical properties of the Great Lakes continues using ship-of-opportunity platforms (Lesht et al., 2003).

A recent algorithm combines Bukata's optical model from Lake Ontario with complex computational methods to separate chl a from the other components in Lake Michigan (Pozdnyakov et al., 2005). This model employs non-linear optimization techniques including Levenberg-Marquardt multivariate optimization and neural network emulation. These techniques essentially minimize the differences of modeled and measured irradiances by varying the concentrations of the three components. Seven years of SeaWiFS data have identified seasonal patterns and interannual change in the abundance of the three components in Lake Michigan (Shuchman et al., 2006). Chl a concentration exhibited a bimodal pattern with spring and fall peaks representing seasonal blooms. However, suspended inorganic matter concentration reflected two different phenomena. In February, the concentration of suspended material was correlated to the number of days with a strong northern wind component and reflected wind-driven resuspension of bottom sediments. In August, suspended sediment was correlated to surface water temperature and instead reflected suspended calcite that precipitates at warm water temperatures. This analysis is the first step toward measuring chl a patterns for other Great Lakes including Lake Ontario. George Leshkewitz of GLERL and Upstate Freshwater Institute of Syracuse, NY measured the optical properties of Lake Ontario in August, 2007 and May, 2008 toward updating such models.

Use of satellite data for comparison to LOLA 2003

Direct comparison to in situ data is a key validation of satellite imagery. Each of the Great Lakes is the focus of lake-wide monitoring efforts once every five years, and in 2003 Lake Ontario was sampled by a team of Canadian and US scientists in a program known as the Lake Ontario Lower Foodweb Assessment (LOLA) (Holeck et al., 2008). Water chemistry, phytoplankton, zooplankton, and benthic samples were collected during three lake-wide cruises aboard the EPA vessel Lake Guardian and the Canadian Coast Guard vessel Limnos. These surveys correspond to spring (April 28–May 3, 2003), summer (August 10–11 and August 19–21, 2003), and fall (September 21–25, 2003) seasons and sampled sites along four north-south transects (Figure 2). The program was closely modeled on the Lake Ontario Trophic Transfer Study (LOTT) of 1995 and 1990.

Figure 2.

LOLA sampling sites in 2003. Chl a measurements were made on integrated water samples from the upper 20 m of the water column for each site. The gray area represents the 0–30 m depth interval.

Figure 2.

LOLA sampling sites in 2003. Chl a measurements were made on integrated water samples from the upper 20 m of the water column for each site. The gray area represents the 0–30 m depth interval.

The analysis of LOLA shipboard data documented historic lows for total phosphorus, chl a, and zooplankton biomass in Lake Ontario for 2003 (Holeck et al., 2008). The average lake-wide chl aconcentration did not exceed 2.7 μg l−1. Average chl a concentration was significantly higher in the fall (2.7 μg l−1) than in summer (1.9 μg l−1) and spring (1.3 μg l−1). The fall chl a concentration increase was consistent with a decrease in secchi depth from an average of 9.8 m in spring to 6.7 m in the fall. The shipboard data exhibited few examples of spatial heterogeneity represented by east/west or nearshore/offshore differences. One notable exception was that chl a concentration was significantly higher in the offshore sites (2.9 μg l−1) than for the nearshore sites (2.0 μg l−1) during the fall.

Satellite images are available for comparison to these sampling periods. A researcher can apply to be an authorized SeaWiFS user by following the instructions on the website http://oceancolor.gsfc.nasa.gov/SeaWiFS/LICENSE/checklist.html. Once granted permission, the user may download data for Lake Ontario at the website http://oceancolor.gsfc.nasa.gov/seadas/. At this website the user can also download the software program SEADAS (latest version of 6.1 available for use on several platforms) for use in viewing and processing the files. This study downloaded cloud-free MODIS satellite images (Level 2 processed for surface water temperature and chl a) from May 3, August 18, and September 26, 2003 for comparison to the three LOLA sampling surveys (Figure 3).

Figure 3.

Moderate Resolution Imaging Spectrometer (MODIS) satellite imagery for surface water temperature of Lake Ontario for May 3, August 18, and September 26, 2003. Plotted with NASA's SEADAS software. Scale bar in degree C.

Figure 3.

Moderate Resolution Imaging Spectrometer (MODIS) satellite imagery for surface water temperature of Lake Ontario for May 3, August 18, and September 26, 2003. Plotted with NASA's SEADAS software. Scale bar in degree C.

Satellite based surface temperature measurements agreed well with shipboard measurements, with an overall r2 = 0.98 for the three sampling seasons (Watkins, 2009). Satellite images provide a comprehensive view of the lake surface. Surface water temperature was cool and uniform in the well-mixed conditions of early May (Figure 3A). The water column stratified with a warm surface layer by August (Figure 3B). The warm epilimnion began to break down and cool by late September (Figure 3C). GLSEA analysis (http://coastwatch.glerl.noaa.gov) confirms that August 18 was the time of peak average surface water temperature (23.6°C) for 2003 in Lake Ontario (Figure 1). Wind-driven upwelling events, atmospheric cooling and the passage of Hurricane Isabel decreased average surface water temperature by 7°C between August 18 and September 26, 2003.

In the offshore region, satellite based chl a measurements based on the OC4 algorithm compared well with shipboard measurements with an r2 = 0.62 (Watkins, 2009). As in the case of the shipboard data, the offshore region had elevated chl a concentration in September relative to August and May (Figure 4). However, satellite measurements often grossly overestimated chl a concentration of shallow sites < 30 m. This overestimate was probably due to the added reflectance of suspended sediment and dissolved organic matter. In August, the nearshore region off the southern shore of Lake Ontario had particularly high reflectance and exhibits features such as river plumes, eddies, and offshore jets (Figure 4B). By September the nearshore/offshore gradient in chl a concentration was weaker (Figure 4C). This general success is comparable to a similar comparison of the 1995 Lake Ontario Trophic Transfer (LOTT) monitoring effort to data from the European Space Agency satellite Along Track Scanning Radiometer (ATSR-2) (Johnson et al., 1997).

Figure 4.

MODIS satellite imagery for surface chl a of Lake Ontario for May 3, August 18, and September 26, 2003. Based on OC4 algorithm. Scale bar in μg l−1.

Figure 4.

MODIS satellite imagery for surface chl a of Lake Ontario for May 3, August 18, and September 26, 2003. Based on OC4 algorithm. Scale bar in μg l−1.

Additional satellite images fill the gaps in time between the three shipboard surveys. For examples, satellite temperature images from June 15, 2003 show the development of the thermal bar with a sharp gradient from a band of warmer nearshore water to cold offshore water (Figure 5A). The nearshore water mass also had high chl a concentration (Figure 5B). The measurements of chl a concentration for this water mass exceed the values validated by our comparison of in situ data. Therefore, sediment resuspension may contribute to overestimates for chl a.

Figure 5.

MODIS satellite imagery for surface temperature and chl a of Lake Ontario for June 15, 2003 illustrating development of thermal bar.

Figure 5.

MODIS satellite imagery for surface temperature and chl a of Lake Ontario for June 15, 2003 illustrating development of thermal bar.

Conclusions

Satellite imagery including surface water temperature and chl a are easily and freely accessible for comparison to shipboard monitoring efforts of the Great Lakes from the website http://oceancolor.gsfc.nasa.gov/seadas/. Temperature measurements are very reliable as long as images are free of cloud cover. Chl a algorithms perform well in offshore conditions with low chl a and at times of low atmospheric interference, but fail closer to shore. A recently developed algorithm that separates the contributions of chl a, particles, and dissolved organic matter in Lake Michigan is a very promising advancement. In situ optical measurements have been collected towards developing a similar algorithm for Lake Ontario. This algorithm will permit us to better analyze the archived SEAWiFS and MODIS data for past trends of phytoplankton abundance.

Acknowledgements

I thank Mohi Munawar for inviting me to speak at the IAGLR session in spring 2007 and contribute to this special issue. I thank my advisors Ed Mills and Lars Rudstam for supporting my pursuit of this side-interest during our analysis of LOLA data. I also acknowledge the SeaDAS Development Group at NASA GSFC for the use of SeaDAS software and the Ocean Color Group for providing access to MODIS data. This is contribution number 272 of the Cornell Biological Field Station.

References

Assel, R., Cronk, K. and Norton, D.
2003
.
Recent trends in Laurentian Great Lakes ice cover
.
Climatic Change
,
57
:
185
204
.
Austin, J. A. and Colman, S.
2007
.
Lake Superior summer water temperatures are increasing more rapidly than regional air temperatures: a positive ice-albedo feedback
.
Geophys. Res. Lett.
,
34
:
L06604
doi:10.1029/2006GL029021
Binding, C. E., Jerome, J. H., Bukata, R. P. and Booty, W. G.
2007
.
Trends in water clarity of the lower Great Lakes from remotely sensed aquatic color
.
J. Great Lakes Res.
,
33
:
828
841
.
Bukata, R. P., Jerome, J. H., Bruton, J. E., Jain, S. C. and Zwick, H. H.
1981a
.
Optical water quality model for Lake Ontario. 1: Determination of the optical cross sections of organic and inorganic particulates in Lake Ontario
.
Applied Optics
,
20
(
9
):
1696
1703
.
Bukata, R. P., Bruton, J. E., Jain, S. C. and Zwick, H. H.
1981b
.
Optical water quality model for Lake Ontario. 2: Determination of chlorophyll a and suspended mineral concentrations of natural waters from submersible and low altitude optical sensors
.
Applied Optics
,
20
(
9
):
1704
1714
.
Bukata, R. P., Jerome, J. H., Kondratyev, K. Y. and Pozdnyakov, D. V.
1991
.
Estimation of organic and inorganic matter in inland waters: optical cross sections of Lakes Ontario and Ladoga
.
J. Great Lakes Res.
,
17
(
4
):
461
469
.
Chen, C., Wang, L., Ji, R., Budd, J. W., Schwab, D. J., Beletsky, D., Fahnenstiel, G. L., Vanderploeg, H., Eadie, B. and Cotner, J.
2004
.
Impacts of suspended sediment on the ecosystem in Lake Michigan: A comparison between the 1998 and 1999 plume events
.
J. Geophys. Res.
,
109
:
C10S05
doi:10.1029/2002JC001687
Holeck, K. T., Watkins, J. M., Mills, E. L., Johannsson, O. E., Millard, E. S., Richardson, V. and Bowen, K.
2008
.
Spatial and long-term temporal assessment of Lake Ontario nutrients, chlorophyll a, and zooplankton
.
Aquatic Ecosystem Health and Management
,
11
(
4
):
377
391
.
Hook, T. O., Rutherford, E. S., Brines, S. J., Schwab, D. J. and McCormick, M. J.
2004
.
Relationship between surface temperature and steelhead distributions in Lake Michigan
.
North Am. J. Fisheries Management
,
24
(
1
):
211
221
.
IOCCG
.
2000
.
Remote Sensing of Ocean Colour in Coastal, and Other Optically Complex Waters
, Edited by: Sathyendranath, S.
Dartmouth, , Canada
:
IOCCG
.
Reports of the International Ocean-Colour Coordinating Group, No. 3
Johnson, G. M., Millington, A. and Harper, D.
1997
.
Observations of phytoplankton dynamics in Lake Ontario using ATSR-2 Imagery
Lesht, B. M., Leshkevich, G. A., Stroud, J. R., Schwab, D. J. and Stein, M. L.
2003
. “
Application of Satellite Observations of Water Color to Study of the Great Lakes
”. In
Global Threats to Large Lakes: Managing in an Environment of Instability and Unpredictability
,
101
102
.
Chicago, IL
:
International Association of Great Lakes Research
.
Morel, A. and Prieur, L.
1977
.
Analysis of variations in ocean colour
.
Limnology and Oceanography
,
22
:
709
722
.
Mortimer, C. H.
1988
.
Discoveries and testable hypotheses arising from Coastal Zone Color Scanner imagery of southern Lake Michigan
.
Limnology and Oceanography
,
33
(
2
):
203
226
.
O'Reilly, J. E., Maritorena, S., Mitchell, B. G., Siegel, D. A., Carder, K. L. and Garver, S. A.
1998
.
Ocean color chlorophyll algorithms for SeaWiFS
.
Journal of Geophysical Research
,
103
(
C11
):
2493
2495
.
Plattner, S., Mason, D. M., Leshkevich, G. A., Schwab, D. J. and Rutherford, E. S.
2006
.
Classifying and forecasting coastal upwellings in Lake Michigan using satellite derived temperature images and buoy data
.
J. Great Lakes Res.
,
32
(
1
):
63
76
.
Pozdnyakov, D., Shuchman, R., Korosov, A. and Hatt, C.
2005
.
Operational algorithm for the retrieval of water quality in the Great Lakes
.
Remote Sensing of the Environment
,
97
(
3
):
352
370
.
Schneider, K., Assel, R. A. and Croley, T. E.
1993
.
Normal water temperature and ice cover of the Laurentian Great Lakes: A computer animation, data base, and analysis tool
,
Ann Arbor, MI
:
NOAA Great Lakes Env. Res. Lab.
.
NOAA Tech. Memo. ERL GLERL-81
Shuchman, R., Korosov, A., Hatt, C., Pozdnyakov, D., Means, J. and Meadows, G.
2006
.
Verification and application of a bio-optical algorithm for Lake Michigan using SEAWIFS: a 7-year inter-annual analysis
.
J. Great Lakes Res.
,
32
(
2
):
258
279
.
Schwab, D. J., Leshkevich, G. A. and Muhr, G. C.
1999
.
Automated mapping of surface water temperature in the Great Lakes
.
J. Great Lakes Res.
,
25
(
3
):
468
481
.
Strong, A. E. and Eadie, B. J.
1978
.
Satellite observations of calcium carbonate precipitations in the Great Lakes
.
Limnology and Oceanography 23
,
5
:
877
887
.
Strong, A. E., Stumpf, H. G., Hart, J. L. and Pritchard, J. A.
1974
. “
Extensive summer upwelling on Lake Michigan during 1973 observed by NOAA-2 and ERTS-1 satellites
”. In
Remote sensing of the environment
,
923
932
.
Ann Arbor
:
Environ. Res. Inst. Mich.
.
Proc. Int. Symp. (9th)
Stumpf, H. G. and Strong, A. E.
1975
. “
Surface circulation in the Great Lakes as observed by Landsat-1 August 1972–December 1973: Southern Lake Michigan
”. In
NASA Earth Resources Survey
1973
1988
.
Proc. Symp. Johnson Space Center, Houston. TMX-58168
Ullman, D., Brown, J., Cornillon, P. and Mavor, T.
1998
.
Surface temperature fronts in the Great Lakes
.
J. Great Lakes Res.
,
24
(
4
):
753
775
.
Watkins, J.
2009
.
Comparison of shipboard and satellite measurements of surface water temperature and chl a in Lake Ontario
.
Aquatic Ecosystem Health and Management
,
12
(
3
):
271
280
.
Wilson, K. A., Howell, E. T. and Jackson, D. A.
2006
.
Replacement of zebra mussels by quagga mussels in the Canadian nearshore of Lake Ontario: the importance of substrate, round goby abundance, and upwelling frequency
.
J. Great Lakes Res.
,
32
(
1
):
11
28
.