Journal of Great Lakes Research 43 (2017) 689699

Contents lists available at ScienceDirect

Journal of Great Lakes Research

journal homepage: www.elsevier.com/locate/jglr

Wash-zone dynamics of the thermocline in Lake Simcoe, Ontario

R. Cossu a,d,, M.S. Ridgway c, J.Z. Li b, M.R. Chowdhury d,e, M.G. Wells d
a
School of Civil Engineering, University of Queensland, Brisbane, Queensland, Australia
b
Department of Physical and Environmental Sciences, University of Toronto, Ontario, Canada
c
Harkness Laboratory of Fisheries Research-Aquatic Research and Development Section, Ontario Ministry of Natural Resources, Peterborough, ON, Canada
d
University of Toronto Scarborough, Canada
e
Minnow Environmental Inc., Georgetown, Ontario, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Lake Simcoe exhibits substantial thermal variations during the stratied period and rapid temperature changes
Received 23 September 2015 along the zone where the thermocline intersects with the lake bed. The observed wind-driven thermocline
Accepted 13 May 2017 movements are analyzed and discussed based on high-frequency measurements made in August and September
Available online 16 June 2017
2011 and from a 30-year record of bi-weekly temperature proles. Winds with speeds exceeding 5 m s1 drive
large-amplitude internal seiches on an approximately weekly timescale. The projected wash-zone associated
Keywords:
Internal waves
with these thermocline motions stretches over a quarter of the total surface area where rapid changes in the ben-
Wash-zone of thermocline thic temperature of as much as 5 C/h occur. A major new nding of this work is that the mean depth of the ther-
Thermal habitat mocline scales favorably with the depth at which surface waves tend to redistribute only ner sediment and the
Ecotone associated substrate changes above and below this boundary. In lakes where this scaling applies, we hypothesize
that wash-zones can be associated with dynamic habitat bounds (ecotones) due to thermal variability and a sub-
strate type transition from muds to sand. Thus, thermocline movements in stratied, similar-sized lakes and the
links to physio-biological processes should be further scrutinized, in particular with ongoing climate change and
fast warming water bodies.
2017 The Authors. Published by Elsevier B.V. on behalf of International Association for Great Lakes Research. This
is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction
Upwelling of cold hypolimnetic waters and downwelling of warm
epilimnetic waters occur in all lakes due to internal waves and seiches
triggered by strong wind events (Hawley and Muzzi, 2003; Hlevca et
al., 2015; Troy et al., 2012; Wells and Parker, 2010). The motion of
these water masses by internal seiches results in important temperature
variability around the mean depth of the thermocline and can also in-
duce intensive benthic boundary mixing (Chowdhury et al., 2016). In
turn, these processes have a major control on physiological processes
and the ecology of the water body (Cyr, 2012, 2016; Evans et al.,
2008; Finlay et al., 2001; Magnuson et al., 1979; Spigarelli et al., 1982).
The physical signicance of temperature uctuations prevalent in tem-
perate lakes has previously been acknowledged, but few studies have
linked observations of thermal variability to biological processes
(Ludsin et al., 2014; Coulter et al., 2015a, 2015b). As water temperature
is a key ecological determinant for ectothermic organisms, we will doc-
ument how large the spatial extent of these persistent and important
Corresponding author at: School of Civil Engineering, University of Queensland,
Brisbane, Queensland, Australia.
E-mail address: r.cossu@uq.edu.au (R. Cossu).
temperature changes at the lake bed can be, which has been overlooked
in past studies of sh habitat usage.
The magnitude of internal seiche motions and associated tempera-
ture variability in stratied lakes depends on wind speed, wind dura-
tion, fetch length and strength of thermal stratication (Shintani et al.,
2010). During increased surface wind stress, the thermocline is de-
pressed at the downwind end of the lake, while upwelling occurs at
the upwind end (see Fig. 1; Fischer et al., 1979, and references therein).
When wind stress decreases, the position of the thermocline starts
to oscillate about its mean depth driving thermal variation and benthic
currents at the lake bed. The vertical scale of thermocline excursions can
be many meters in lakes, in particular when coupled with a gently
sloping lake bed (Hawley and Muzzi, 2003; Cossu and Wells, 2013).
The excursions can result in an extensive area where the
thermocline washes back and forth along the lake bed (hereinafter the
wash-zone).
In addition, with sufcient fetch, surface waves run ashore at the
downwind end of a lake (Fig. 1a). The wave base indicates the depth
at which wave-induced dynamics vanish so that erosion and sediment
transport is less prevalent beyond the wave base. This process is
known as sediment focusing and leads to accumulation of ner
sediments in deeper parts of lakes while coarse particles predominantly
remain in shallow areas (Blais and Kalff, 1995). It follows that the type of

http://dx.doi.org/10.1016/j.jglr.2017.05.002
0380-1330/ 2017 The Authors. Published by Elsevier B.V. on behalf of International Association for Great Lakes Research. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
690 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699

Fig. 1. a) Due to wind forcing, surface waves build at the downwind end of the lake. The orbital velocity is a function of wave height, and decreases with depth. The maximum depth at
which erosion occurs is known as the wave base. Thus, the wave base separates areas of erosion (green) and deposition (light-brown) as only ner sediments can be transported to the
deeper parts of the lake. b) The thermocline oscillates in the depth range between zthm zth, which is dened as the wash-zone (red). zthm is the average depth of the thermocline and zth
the excursions above and below this average depth. The range of the wash-zone indicates that temperature variability in a stratied system can occur in both the erosion and deposition
zones (i.e., between hard and soft substrates.

substrate depends on the location relative to the wave base and the pre-
vailing sediment transport regime.
We will show mathematically that the depth range of the thermo-
cline has an empirical scaling (Gorham and Boyce, 1989), which scales
similarly to the depth of wave bases in large lakes. This phenomenon
implies that these two important depth scales could overlap in many
lakes in the Great Lakes region (see next section), which is a novel
and general result not previously reported. This overlap underlines the
signicance of the location of the thermocline as an ecotone as upper
and lower thermal boundaries, as well as substrate changes, lie within
its physical location (Cadenasso et al., 2003; Strayer et al., 2003). Thus,
the position of the wash-zone could be important, in particular with re-
spect to ongoing climate change and its impact on thermal regimes in
temperate lakes (Ficke et al., 2007). Furthermore, the position of the
wash-zone may have strong implications for an ecosystem such as for
benthic secondary production and even to some extent to sh occupan-
cy (Szekeres et al., 2016). The variability and extension of the wash-
zone could also play a role for sh occupancy during near-bottom hyp-
oxia (Arend et al., 2011; Vanderploeg et al., 2009).
In this study, we focus predominantly on thermocline movements ob-
served in Lake Simcoe, Ontario, which is known for its dynamic thermo-
cline (Baird and Associates, 2006; Bouffard and Boegman, 2011;
Chowdhury et al., 2015) and large temperature variability in the wash-
zone (Cossu and Wells, 2013). We quantied the temperature variability
using both high-frequency data collected during a strongly stratied peri-
od in summer 2011 and long-term records (19802011) of the thermal
structure. Anoxia can often develop beneath strong thermoclines in
many lakes, so temperature variability in a wash-zone also implies dis-
solved oxygen variability. We discuss our ndings with respect to the im-
portance of linking physical and biological variables to improve and
advance future monitoring programmes in the Great Lakes region.
Methods
Field site
Our eld study was conducted in Lake Simcoe (4425 N, 7930 W),
a large dimictic lake located in Ontario, Canada (Fig. 2). It is the largest
inland lake of southern Ontario with a total surface area of 722 km2, a
total volume of 11.6 km3, and an elevation of 219 m above sea level.
The lake has one main basin with two side arms, Kempenfelt Bay in
the west and Cook's Bay in the southwest. The main basin is roughly cir-
cular with a gentle sloping shallow eastern half, and steep slopes on the
western side. Our eld site was located on the southeastern shore of the
main basin where there is a gentle slope of 1.5 m/km. The lake has a
maximum depth of 41 m, with a mean depth of 20 m and a maximum
effective fetch of 30 km. It is oligotrophic with a mean thermocline
depth of 10 m, where typical summertime surface temperatures reach
2022 C in the epilimnion and 810 C in the hypolimnion (Stainsby
et al., 2011). Stratication is strongest during hot summer months
(July to early September), which determined the period and duration
of our observations.
Lake Simcoe has suffered from water quality problems since the
1970s and has experienced severe environmental changes (North et
al., 2013), including problems with low hypolimnetic oxygen in deeper
regions favoured by the coldwater sh community. The Draft Lake
Simcoe Protection Act (2009) targets an ecologically sustainable state,
which is based on a self-sustaining coldwater sh community and a
deep-water dissolved oxygen target concentration of 7 mg L1. An im-
portant change in Lake Simcoe has been observed since zebra and quag-
ga mussels (Dreissena polymorpha and D. bugensis) started to colonize
the lake since 1995 and 2004, respectively, which has caused dramatic
changes to water clarity, benthic biomass, and food-web structure
(Evans et al., 2011). These mussels are most abundant on harder sub-
strate in depths between 5 and 15 m (Technical Progress Series in
Lake Monitoring: Report No. 2, 2014), and so they live near the mean
depth of the thermocline.
Instrumentation
Prevailing winds usually blow from the west and northwest during
late summer in Lake Simcoe. Wind speed and direction were measured
from the hourly records of the Environmental Canada meteorological
station located at the western shore of Lake Simcoe near Barrie airport
(4429 N 7933 W, Fig. 1).
 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699 691

Fig. 2. Location and bathymetry (contours every 5 m) of Lake Simcoe Ontario, Canada. Contours are shown in 4 m steps. Four thermistor chains were deployed: one at a depth of 20 m (C1)
and three benthic chains in shallow waters ranging from the lake bed to 1.2 m above the lake bed (T1T3). Also shown is the location of Barrie Airport (weather station) and the long-term
Ontario Ministry of the Environment monitoring sites K45 and S15.

To monitor the wash-zone dynamics, we deployed an array of four
temperature logger chains from 3rd of August corresponding to Day of
Year 216 (DOY 216) until 15th of September 2011 (DOY 257) on the
gently sloping SE side of Lake Simcoe. A vertical thermistor chain was
deployed at a depth of approximately 20 m, 3.4 km offshore near the
S15 long-term water quality monitoring station (Fig. 2). The thermistor
chain consisted of 16 individual temperature loggers (TR - 1060, RBR),
which were located between 5 and 20 m in depth, with a vertical spac-
ing of 1 m. The loggers had a response time of 3 s, an accuracy of 0.002
C, and a sampling interval of 4 s. In addition, three benthic vertical
thermistor chains were deployed near the lake bottom in a line between
the long thermistor chain and the shore (Fig. 2). Each thermistor chain
consisted of 10 individual temperature loggers (Seabird SBE 56), which
were located between 0.05 m and 1.30 m above the lake bed, with a ver-
tical spacing of 0.15 m. The accuracy of the thermistors was 0.002 C,
and the sampling interval was set to 2 s. The internal clocks of all loggers
were synchronized at the beginning of the sampling period. All data
presented here start at DOY 216, approximately 812 h after the instru-
ments were deployed in the water, and nish on DOY 257 of 2011.
Complementing the high-frequency measurements from 2011, we
also analyzed data from long-term water quality stations K45 and S15
that have been monitored since 1980 (Eimers et al., 2005; Stainsby et
al., 2011) by the Ontario Ministry of the Environment and Climate Change
(MOECC). Typically, the water temperatures and dissolved oxygen (DO)
levels were measured at the K45 and S15 sites every two weeks at
11 am and 2 pm (on the same day except during poor weather condi-
tions) for the entire ice-free season during May to October. These mea-
surements were made at 1-m depth intervals with a hand-held YSI probe.
Thermocline calculations
The depth of the thermocline in Lake Simcoe was calculated from the
temperature proles taken from either the high frequency 2011 data or
from the long-term MOECC data at sites S15 and K45. The density of the
water, , was computed using the non-linear freshwater equation of
state (Chen and Millero, 1977). The position of the seasonal thermocline
and location of maximum density gradient was established by
determining the rst moment of the density gradient,
ZH ZH
zth z dz= dz 1
z z
0 0
692 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699

with z being the vertical coordinate and the density of water
(Patterson et al., 1984). We were interested in correlating the deection
of the thermocline to wind-driven forcing. We determined the deec-
tion of the thermocline (zth) by the difference from the centered tem-
poral average of zth for 96 h for a continuous time-series as measured at
C1. We used this averaging period so that the thermocline oscillations
due to the basin-scale internal Kelvin waves (90 h) and Poincar
waves (14 h) did not inuence the calculated wind-driven thermocline
deection values (Bouffard and Boegman, 2011). The estimated V1H1
(vertical and horizontal mode 1) period is about 40 h with a fetch length
of 25 km. The magnitude of upwelling can be expressed with the
Wedderburn Number, W, which is dened as
0 2
g h1
W 2
u2 L
where g = g (2 1) / 1 is the reduced gravity with 1 and 2 being
the densities of the upper and lower water layer, h1 is the upper layer
thickness, u is the water-side shear velocity, and L is the length of the
lake basin (Shintani et al., 2010 and references therein).
Thermocline deections (zth) were also estimated from the long-
term monitoring records of thermal structure at sites K45 and S15. Be-
cause measurements at these two sites were made within 3 h of each
other, we could calculate the difference in thermocline depth between
the two stations and gain an estimate of the thermocline deection
(zth). This deection could then be compared with our high frequency
observations. As a simplication, the upwelling of the thermocline at
S15 was associated with a downwelling event at K45 so that the actual
thermocline depth across Lake Simcoe was assumed to be half of the
vertical deection determined at these sites. This simplication enabled
a comparison with the measured thermocline excursion at site C1.
We note that this assumption is based on a linear slope of the ther-
mocline across the basin of Lake Simcoe. However, this assumption
does not work in large lakes such as Lake Michigan, where the upwell-
ing front can result from surface layer Ekman transport (Troy et al.,
2012) and where Coriolis forces result in Poincare waves dominating
the internal seiches. Ekman pumping due to surface winds can also im-
pact the shape of the thermocline as observed in Lake Erie where a
bowl-shaped thermocline was reported by Beletsky et al. (2012).
However, in smaller lakes with less dominant Coriolis effects such as
Lake Simcoe, the assumption of a linear tilt to the thermocline offers a
reasonable approximation for a rst-order estimate of the variable
temperature zone, although bathymetry and stratication can cause a
more complex internal seiche structure (Fricker and Nepf, 2000) than
approximated here.
Correlation of erosion/deposition boundary with the depth of the thermocline
The depth of the erosion/deposition boundary in lakes is mainly set
by the magnitude of the surface waves. In most lakes, these surface
waves are limited in their growth by the size of the lake and the strength
of the wind. Consequently, larger lakes with a longer fetch length have
deeper erosion/deposition boundaries. For lakes at similar latitudes,
the depth of the thermocline is also related to the size of the lake and
the strength of the wind (Gorham and Boyce, 1989). Below, we will
show for the rst time that both the depth of the thermocline and the
depth of erosion/deposition boundary have the same scaling with lake
size and wind speed. Hence, the two zones will normally be co-located
at roughly the same depths in most lakes - deeper in large lakes, and
shallower in small lakes.
An estimate of the depth of the thermocline, hT, at the time of max-
imum heat content is given by Gorham and Boyce (1989) as
hT 2:0=g1=2 L1=2 3
where is the wind stress, g is gravity and is the density difference
between the epilimnion and hypolimnion. L is the fetch of the lake,
and for a roughly circular lake, L ~ Area1/2. The scaling is strictly only
for lakes of smaller surface area, but is similar to the scaling of Fee et
al. (1996) where hT ~ Area1/4, i.e., hT ~ L1/2, which was valid for lakes of
area up to 4848 km2. We note that this is an estimate for medium-size
lakes because, in small lakes, convection will have a larger inuence
on the depth of the mixed layer (Read et al., 2012). The wind stress
is also dened being proportional to the square of the wind speed U,
so that the above scaling can also be written as
hT  UL1=2 4

Fig. 3. a) Wind direction and magnitude for the Barrie Airport Weather Station (4429 N 7933W) during the observation period, day of year (DOY) 216257. Direction indicated is that
from which the wind is blowing. Concentric rings (2, 4, 6 and 8) show the percentage of the time wind is from that particular direction. The colours within a wedge represent the relative
proportion of time that the winds from that direction are of a certain magnitude. b) Wind direction and magnitude for wind events during August through September 20042012 at the
Barrie Airport station.
 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699 693

The depth to which surface wave-driven erosion in lakes occurs is gc.ca/isdm-gdsi/waves-vagues/data-donnees/data-donnees-eng.asp?

set by the depth at which waves orbital velocities rst become impor-
tant. This depth usually occurs at half the wavelength of a deep-water
wave, as below this depth the wave induced orbital velocities are negli-
gible. Thus, the erosion boundary is usually considered to be where
depth = 1/2 wavelength D. The relationship between wavelength
and frequency for deep-water waves is
2
D g=2 f
where f is wave frequency. Hence, low-frequency waves have larger
wavelengths. The wave-base or depth of inuence (Din) is the depth
where the orbital velocities have dropped substantially is usually
dened as
Din D =2
The wave periods in Lake Simcoe rarely exceed 3 s and periods of 4 s
are the absolute maximum ever observed during summer at Environ-
ment Canada wave buoy C45151 (http://www.meds-sdmm.dfo-mpo.
medsid=C45151). Thus, using Eqs. (5) and (6), the maximum depth of
the wave base (which is just half the wavelength) is about 12.5 m in
Lake Simcoe (using 4 s as a conservative upper boundary). The frequen-
cy of the surface waves can be predicted as a function of wind speed and
fetch. The empirical scaling developed by Donelan (1980) and Donelan
et al. (1985), based on data collected in the Great Lakes, suggested the
relationship between frequency (f), fetch, and wind speed to be
5
f 1:85g 0:77 L0:23 U cos 0:54 7
where U cos is the projection of the wind onto the fetch L and gravity g.
This scaling is similar to a larger dataset presented in Kahma and
Calkoen (1992). The exponents on the fetch and wind speed in this
6 equation are 0.23 and 0.54, which are close to 1/4 and 1/2, so
that we can approximate this equation as the following scaling relation-
ship of
f  L1=4 U 1=2 8

Fig. 4. Relation between wind speed, air temperature, thermocline movements, and benthic temperature variability during the observation period, day of year (DOY) 216257. Shaded
areas indicate periods of strong wind events. a) Recorded wind speed and wind direction at Barrie Airport. b) Comparison of air temperature (blue line) and water temperature (red
line) at the open-water location 5 m below the surface. c) Thermal variability with depth (z) at site C1 is shown with colour contours, with the position of the thermocline at the C1
site marked in black. d) Temperature stratication observed at T1 during up- and downwelling of the thermocline. Arrows indicate periods of rapid coldwater upwelling.
694 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699
In combination with Eqs. (5) and (6), we can estimate the depth of
erosion to be
Din  UL1=2
Importantly, this relationship has the same scaling on wind-speed U
and fetch L as the scaling for the depth of thermocline (Eq. (4)). The sim-
ilarity of these two simple scaling arguments demonstrates our impor-
tant and novel result that the depth where erosion occurs is
essentially the same scaling as the empirical scaling of the thermocline
depth. Both depths increase with wind speed and lake size (fetch
length), similar to scaling of mud depositional boundary of Rowan et
al. (1992).
Although the erosion/deposition boundary depth coincides with the
thermocline depth in Lake Simcoe (Electronic Supplementary Material
ESM Fig. S1), the movements of the thermocline do not directly cause
this erosion/deposition boundary. Rather the processes of wave forma-
tion and deepening of thermocline are driven by wind stress acting at
the lake surface. More importantly, our new scaling analysis suggests
that in most lakes, the ecotone formed by the wash-zone of the thermo-
cline occurs at the location of maximum change in substrate type.
Results
Meteorological data
The wind speed and direction are important as they excite thermo-
cline movements and have a major control on surface waves, which in
turn, can impact the deposition of substratum materials. The dominant
wind direction was from the northwest with a mean speed (Uwind) of
2.75 m s 1 (Fig. 3a). The largest wind velocities (exceeding Uwind =
10 m s1) only occurred 0.13% of the time. The meteorological data dur-
ing the eld observation period were in good agreement with long-term
weather patterns for August/September between 2004 and 2012 at Bar-
rie Airport Station (Fig. 3b). In addition, similar maximum wind speeds
N 10 m s1 only occurred 0.18% of the time and during the preceding
three decades (Chowdhury et al., 2015).
Fig. 5. a) Temperature variability at the bottom of the four thermistor chains at site C1 and T1T3 in Lake Simcoe during day of year (DOY) 216257. b) Frequencies of rate of temperature
change during the observations from all bottom temperature loggers at C1 and T1T3. The temporal changes in temperature were calculated over 30 min.
Thermal variability
Apart from air temperature, extreme wind events mainly determine
the thermal variability of lake waters. Several storm events with main
9 wind directions from west and northwest were observed during the
study (Fig. 4a). The epilimnetic waters showed a steady but slow decline
in temperature, from 24 C to 20 C between DOY 214 to DOY 257. This
nding was also supported by the air temperatures with diurnal varia-
tions where mean temperatures dropped from ~ 19 C to b 16 C (Fig.
4b). The thermal structure of the water column clearly showed warm
epilimnetic (N 20 C) and cold bottom waters (b10 C) (Fig. 4c). It is ev-
ident that periods with Uwind N 5 m s1 for 12 h correlated well with pe-
riods of strong thermocline excursions, for instance, observed shortly
after storm events at C1 at DOY 228, DOY 235, DOY 242 and DOY 248
(Fig. 4c).
Thermocline excursions during or shortly after strong winds were
also observed with the benthic thermistor chains in the near-bottom re-
gion. Temperatures varied at the lake bed, and ranged from N22 C to
approximately 10 C, as illustrated for the shallowest site T1 (Fig. 4d).
Periods with warm temperatures lasted for several days and alternated
with periods of cold temperatures associated with coldwater upwelling.
The upwelling and downwelling events were substantially shorter at
site C1 and more prolonged at site T1 (Fig. 4d). During downwelling
events of warm water, small vertical temperature gradients were evi-
dent at their conclusion. For instance, asymmetric temperature changes,
as monitored between DOY 223 and DOY 225, can be associated with
rapid coldwater upwelling (Cossu and Wells, 2013). Such rapid shifts
in temperature were noticeable at all thermistor chains near the bot-
tom. The intermittency of the temperature variability at the bottom of
the four thermistor chains at site C1 and T1T3 is depicted in Fig. 5a
and b. For 0.5% of the time, the rate of temperature change is as high
as 5 C/h. There were also several more extreme events (0.01% of
time), when the rate of thermal variability was larger than 610 C/
h (Fig. 5b).
A comparison of the monitored thermocline depths with the long-
term record in the period between DOY 200 and DOY 260 helps to clas-
sify the observed thermal structure with respect to general conditions in
Lake Simcoe's main basin. The depth of the mean thermocline zthm
matched well with the long-tern record between 1985 and 2012 (Fig.
 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699 695

6a). At site S15 the thermocline varied between zthm = 11.815.8 m at
station K45 the depth range was zthm between 12 m and 18 m. The av-
erage value of this 30-year record was zthm = 14 m, which is in good
agreement with the averaged depth of the thermocline (zthm =
13.8 m, red square marker) for the high-frequency measurements at
C1 in 2011 (Fig. 6a).
The Wedderburn number provides a good estimate of the relation
between wind, strength of the stratication, and resulting thermocline
movements. During times of increased wind speeds (N 5 m s 1), the
Wedderburn number showed minima of the order of O(101) (Fig.
6b) suggesting that upwelling occurred, which in turn relates to strong
thermocline excursions (Fig. 6c). For instance, vertical displacements of
the thermocline of 1 m were prevalent for Wedderburn numbers up to
the order of W ~ O(100), while large displacements (~4 m) occurred for
Wedderburn number W 1 (Fig. 6b and c).
The dynamics of the thermocline movements can be summarized
by comparing the vertical displacement of the thermocline (zth )
from the mean depth versus the frequency for relative and absolute
values (Fig. 6d and e, respectively). Similar relationships were ob-
served by estimating the vertical displacement as half the difference
in thermocline depth between station S15 and K45 from 30 years of
bi-weekly summer measurements made by the MOECC (Fig. 6f and g,
respectively). The most common thermocline deections were at the
order of 1 m from the neutral level. For instance, 1 m excursions oc-
curred ~ 50% of the time based on the 2011 dataset and 57% of the
time based on the long-term record; however, both datasets also
showed signicant excursions at the order of zth N 4 m, 1.5% during
the 2011 observation period and 0.94% for the long-term record. The
similar frequency distributions for the long-term data (variance of
2.34) and the high frequency data (with a variance of 1.14) demon-
strated that the observed short-term thermocline amplitudes were
representative of typical thermocline displacements in Lake Simcoe.
These large vertical movements can be linked to periods of strong
wind speeds as indicated in Fig. 4.

Fig. 6. (a) Seasonal average of the thermocline depth zth during 19822012 in Lake Simcoe measured at sites S15 (white circles) and K45 (grey shaded circles). The dashed line indicates an
average between these two stations which can be associated with an estimate of the overall mean thermocline depth in Lake Simcoe's main basin. The red square shows the estimated
position from the high frequency data in 2011. b) Wedderburn number estimated from the high-frequency data in 2011. c) Observed thermocline deection from its mean value, zth,
during the high-frequency observations in 2011. d) to g) Frequency (N) of positive (downwelling) and negative (upwelling) excursions of zth, for d) from the measuring location at
C1 and e) the long-term record at MOECC sites S15 and K45. f) Absolute values of the deection zth for measuring location at C1 and g) Absolute values of the deection for the long-
term record at MOECC sites S15 and K45. High-frequency data reect the position of the thermocline at 4 s intervals.
696 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699
Dissolved oxygen
Biweekly measurements of temperature and DO at site S15 allowed
further analysis of the thermocline position and oxygen conditions (Fig.
7a and b, respectively) for the period between DOY 125 to 325. The ther-
mocline at site C1 between DOY 216 and 257 represents the warmest
and most stratied period of the year (Fig. 7a). A few weeks after the
fast response thermistor chains were retrieved, the lake temperature
became vertically homogenous. There are only three times (i.e., at
DOY 221, DOY 235, and DOY 250) when the biweekly measurements
were taken during our period of high-frequency measurements. While
there was reasonable agreement between the position of the steepest
temperature gradient (or location of 12 C isotherm) based on both
datasets, it is clear that the interpolation of the bi-weekly dataset misses
most of the variability in thermocline movement.
The DO levels started to decrease with the onset of stratication (Fig.
7b) and were constantly below 10 mg L1 after DOY 200. At the end of
the summer, between DOY 250 and DOY 260, the hypolimnion in Lake
Simcoe had lower DO levels (b 5 mg L1), which is below the
7 mg L 1 target for Lake Simcoe (Draft Lake Simcoe Protection Act,
2009). Maximum DO values during the summer were found at the
depth of the thermocline (~15 m).
Wash-zone dynamics
A large fraction of the bed of Lake Simcoe is inuenced by wash-zone
dynamics, as a large fraction of the lake bed lies close to the mean depth
of the thermocline. The observations at site C1 were used to estimate
upper and lower vertical boundaries of the wash-zone. In combination
with the dimensionless slope s and vertical movement of the thermo-
cline (zth), the horizontal extent (X) of the wash-zone was estimated
(Fig. 8). For small slopes, the lateral extent of thermocline movement
can be approximated by X = h/s (Coman and Wells, 2012; Troy et
al., 2012). A lateral movement from the point where the thermocline in-
tersects the lake bed, Xm = 0.7 km, was estimated based on the
southern shore slope of s ~ 0.0015 and vertical thermocline movements
zthm = 1 m. Likewise, for extreme events with Uwind N 10 m s1 (W
b 1), thermocline movements of zthmax = 4 m the extent of the wash-
zone increases to Xmax = 2.5 km. More mild-sloping areas, such as
those characterizing most of eastern side of Lake Simcoe (with s ~
0.001), yield a larger horizontal extent of the wash-zone with Xmax
= 1 km and for zth = 1 m. For maximal thermocline excursions
(zthmax = 4 m), the horizontal extent increases to Xmax =
4 km, if winds from south and southwest are projected on this
slope. Such wind events and directions are also common for Lake
Simcoe in August (based on wind data measurements between 1981
and 2010; Chowdhury et al., 2015). In total, the estimated wash-zone
for largest observed excursions (zthmax = 4 m) extended over an
area of 170 km2. We note that this wash-zone area was based on mea-
surements made during August to September 2011. Thus, earlier in
the year, when the thermocline is shallower (see Fig. 7), a smaller
wash-zone in shallower waters could be expected. Nonetheless, 38%
of the lake bed lies between depths of 5 m and 15 m, while 52% lies be-
tween depths of 5 m and 20 m so that a large portion of the lake bed will
be directly impacted by wind-driven internal seiche motions.
Discussion
The physical process of wind-driven movements of the thermocline
occurs in all stratied lakes, and hence all stratied lakes will have a
wash-zone. In particular, many dimictic lakes in Southern Ontario
have similar thermal stratication patterns and experience similar
winds to those in Lake Simcoe. The magnitude of the upwelling depends
in part on the fetch. Thus, in general, larger lakes should experience
greater benthic thermal variability for the same winds. Our Eqs. (4)
and (9) apply to all lakes with similar thermal stratication, and our
novel conclusion is that the erosion depth correlates with thermocline
depth. Below, we specically discuss our results for Lake Simcoe, but
the same processes will be generalizable to many other lakes. This is
an important result and underlines that the wash-zone should be
thought of as an ecotone because, in this region, both the substrate
and temperature show rapid changes.
Physical implications
Substantial temperature variability existed during summer in Lake
Simcoe, which is attributed to vertical movements of the sharp thermo-
cline initiated by strong westerly wind events. Thermocline excursions
of up to 4 m imply that the sloping benthic boundary of Lake Simcoe
was frequently exposed to temperature uctuations during the strati-
ed season. The temporal and spatial scale of these thermocline move-
ments are substantial and could therefore have an impact on the lake
ecology (Cyr, 2016; Magnuson et al., 1979; Spigarelli et al., 1982) and
on biological processes (Ludsin et al., 2014; Coulter et al., 2015a, 2015b).
The position of the wash-zone is almost identical to the sand - mud
boundary in Lake Simcoe as illustrated in Fig. A1 (taken from Technical
Progress Series in Lake Monitoring, 2014). The lower bound of the ero-
sion zone is located near the wave base, i.e., where surface waves cease
to inuence sediment transport and beyond which only ner particles
are transported in suspension (Hkanson, 1977; Hilton et al., 1986;
Rowan et al., 1992). The wavelength of surface waves depends on the
fetch and wind duration and hence can be scaled with the lake surface
area (see Methods section). Based on data from Environment Canada
wave buoy C45151 in the centre of Lake Simcoe, the wave periods are
usually smaller than 4 s, resulting in a wave base with a maximum
depth of 12.5 m. This empirical scaling for the depth of the wave-base
agrees well with the empirical estimate for the thermocline depth as
both measures scale linearly with mean wind speeds and with the
square root of the fetch length. It can be concluded that large lakes
with strong winds have deeper wave bases and thermocline depths
than small lakes with lighter winds. This estimate implies that depths
at which important substrate composition changes occur (e.g., the
mud-sand boundary) are likely to overlap with large temperature vari-
ability observed in the wash-zone. Thus, our new prediction is that
wash-zones in many dimictic lakes can be associated with transitions
Fig. 7. a) Temperature (T) and b) dissolved oxygen (DO) data for site S15 during the
summer 2011 from the MOECC data. In both panels, the black line depicts the position
of the thermocline at the offshore location in 4 s intervals during the observations. The
red line in panel b represents the observed 7 mg L1 DO threshold at site S15. The
dashed white vertical lines indicate the times when biweekly proles were taken. The
three thick white dashed lines indicate the times of biweekly measurements during the
high-frequency measurements.
 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699 697

between both hard and soft substrates, but also warm- vs. cold-water
habitats.
Biological implications
Temperature has been coined as being the abiotic master factor
(Brett, 1971; Donaldson et al., 2008), and our analysis has shown
that the wash-zone introduces rapid temperature changes across a
large area including the substrate transition zone. Therefore, wash-
zone dynamics could be of substantial importance for aquatic organ-
isms, particularly in a warming climate with potentially stronger
thermoclines. Sub-daily temperature uctuations can affect physio-
logical processes (Coulter et al., 2015a, 2015b), and thermocline
movements have recently been suggested to potentially alter the be-
haviour of sh communities (Aspillaga et al., 2017). Thus, species
distributions are likely to shift as temperature denes habitats,

Fig. 8. The wash-zone in Lake Simcoe estimated by extrapolating thermocline deections at site C1. The red line depicts the mean height of the thermocline (zthm = 13.93 m below the
surface), the green lines show the wash-zone for zth = +/ 1 m and the blue lines depict the maximum excursion (zth_max/min = 4 m) of the thermocline. The wash-zone for
maximum excursion extends over an area of 170 km2.
698 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699
niche spaces, and species interactions (Magnuson et al., 1979; Ficke
et al., 2007; Aspillaga et al., 2017).
Global air temperatures are estimated to increase and many lakes
appear to be warming faster than this global air temperature trend
(Austin and Colman, 2007; O'Reilly et al., 2015). Consequently, thermal
habitats are expected to change in both extent and duration although
this variation in the thermal regime is not likely to affect all species
equally (Ficke et al., 2007; Cline et al., 2013). For instance, warmer sur-
face layers, will make sh in the littoral zones more susceptible to cold-
shock from upwelling of cold waters (Szekeres et al., 2016). In addition,
deeper thermoclines and larger wash-zones could mean that sh have
less habitat into which to escape if physical stressors increase. For in-
stance, the consequences of hypolimnetic hypoxia in Lake Erie reduced
sh occupancy and access to optimal growth conditions (Vanderploeg
et al., 2009; Arend et al., 2011).
Benthic communities are most sensitive to strong coldwater upwell-
ing, owing to their immobility and vulnerability to the effects of rapid
temperature changes (Wells and Parker, 2010), as well as potential
changes in DO due to movements of anoxic hypolimnetic waters.
Hence, it is likely that benthic communities should be most inuenced
by the rapid temperature or DO changes associated with a dynamic
wash-zone. The most dramatic temperature drops associated with up-
welling events in the Great Lakes have been observed to kill craysh
(Oronectes propinguus) (Emery, 1970). With the hypothesized changes
occurring due to wash-zone dynamics as benthic organisms must be ac-
climatized to the ubiquitous temperature variability. For instance, some
benthic organisms may benet from the water currents associated with
a fast-moving thermocline in the wash-zone. Chowdhury et al. (2016)
suggested that the movement of the thermocline washes the deep chlo-
rophyll maxima at the thermocline past the locations of maximum
dreissenid mussel density. These mussels are located near the erosion/
deposition boundary at the lake bed of Lake Ontario, as well as in Lake
Simcoe (Technical Progress Series in Lake Monitoring, 2014). These ex-
amples demonstrate that the role of the thermocline and predicted
changes of thermal regimes in lakes are an important factor that should
be taken into consideration when assessing the effects of multiple
stressors for lake ecosystems in the future.
Implication of the ecotone approach: the wash-zone as a dynamic habitat
Our results demonstrated that the wash-zone in Lake Simcoe meets
all denitions of an ecotone, as upper and lower thermal boundaries as
well as substrate changes, lie within its physical location (Cadenasso et
al., 2003; Strayer et al., 2003). Ongoing climate change has already al-
tered the dynamics of Lake Simcoe, evident in intensied and longer
stratied periods (Stainsby et al., 2011). Consequently, a further change
of vertical and horizontal extension of the wash-zone in Lake Simcoe
could have a major impact on density and productivity of benthic food
webs (Cyr, 2016), sh occupancy (Cline et al., 2013), and foraging capa-
bilities (Webb, 1978). This has two important implications when con-
sidering the wash-zone as dynamic habitat in lakes. First, the wash-
zone is potentially a key habitat that couples pelagic and benthic pro-
ductivity because of physical contact during stratication periods. Sec-
ond, the wash-zone is likely to affect the short-term movements of
organisms by driving spatiotemporal variability in suitable habitat.
Sorting out the importance of the wash-zone as a dynamic habitat will
likely require approaches incorporating biotelemetry and targeted,
ner scale sampling to capture organism distribution and productivity
within and outside the ecotone.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jglr.2017.05.002.
Acknowledgements
We acknowledge funding from Environment Canada through the
Lake Simcoe cleanup fund. Logistic support was provided by the Ontario
Ministry of the Environment and Climate Change. MGW received
funding from NSERC (2016-06542) and the Ontario Ministry of Innova-
tions ERA (ER10-07-023) program. We also acknowledge helpful dis-
cussions with Brian Shuter and Ken Minns. We thank the associate
editor Stuart Ludsin and three anonymous reviewers for their feedback
and help to strengthen this work.
References
Arend, K.K., Beletsky, D., Depinto, J.V., Ludsin, S.A., Roberts, J.J., Rucinski, D.K., Scavia, D.,
Schwab, D.J., Hk, T.O., 2011. Seasonal and interannual effects of hypoxia on sh
habitat quality in central Lake Erie. Freshw. Biol. 56:366383. http://dx.doi.org/10.
1111/j.1365-2427.2010.02504.x.
Aspillaga, E., Bartumeus, F., Starr, R.M., Lpez-Sanz, ., Linares, C., Daz, D., Garrabou, J.,
Zabala, M., Hereu, B., 2017. Thermal stratication drives movement of a coastal
apex predator. Sci. Rep. 7 (1), 526.
Austin, J.A., Colman, S.M., 2007. Lake Superior summer water temperatures are increasing
more rapidly than regional air temperatures: a positive ice-albedo feedback.
Geophys. Res. Lett. 34, L06604. http://dx.doi.org/10.1029/2006GL029021.
Baird and Associates, 2006. Lake Simcoe hydrodynamic and water quality model. A Re-
port to the Lake Simcoe Region Conservation Authority. WF Baird and Associates,
Oakville (ONT) (Report# 10940).
Beletsky, D., Hawley, N., Rao, Y.R., Vanderploeg, H.A., Beletsky, R., Schwab, D.J., Ruberg,
S.A., 2012. Summer thermal structure and anticyclonic circulation of Lake Erie.
Geophys. Res. Lett. 39 (6).
Blais, J.M., Kalff, J., 1995. The inuence of lake morphometry on sediment focusing.
Limnol. Oceanogr. 40, 582588.
Bouffard, D., Boegman, L., 2011. Spatio-temporal dynamics of the basin scale internal
waves in Lake Simcoe. Proceedings 7th Int. Symp. on Stratied Flows, Rome, Italy, Au-
gust 2226, 2011, p. 6.
Brett, J.R., 1971. Energetic responses of salmon to temperature. A study of some thermal
relations in the physiology and freshwater ecology of sockeye salmon (Oncorhynchus
nerka). Am. Zool. 99113.
Cadenasso, M.L., Pickett, S.T.A., Weathers, K.C., Jones, C.G., 2003. A framework for a theory
of ecological boundaries. Bioscience 53, 750758.
Chen, C.T., Millero, F.J., 1977. Effect of salt content on the temperature of maximum den-
sity and on static stability in Lake Ontario. Limnol. Oceanogr. 22:158159. http://dx.
doi.org/10.4319/lo.1977.22.1.0158.
Chowdhury, M., Wells, M.G., Cossu, R., 2015. Observations and environmental implica-
tions of variability in the vertical turbulent mixing in Lake Simcoe. J. Great Lakes
Res. 41:9951009. http://dx.doi.org/10.1016/j.jglr.2015.07.008.
Chowdhury, M., Wells, M.G., Howell, T., 2016. Movements of the thermocline lead to high
variability in benthic mixing in the nearshore of large lake. Water Resour. Res. 52:
30193039. http://dx.doi.org/10.1002/2015WR017725.
Cline, T.J., Bennington, V., Kitchell, J.F., 2013. Climate change expands the spatial extent
and duration of preferred thermal habitat for Lake Superior shes. PLoS One 8 (4),
e62279. http://dx.doi.org/10.1371/journal.pone.0062279.
Coman, M., Wells, M.G., 2012. Temperature variability in the near shore benthic boundary
layer of Lake Opeongo is due to wind driven upwelling events. Can. J. Fish. Aquat. Sci.
69:282296. http://dx.doi.org/10.1139/f2011-167.
Cossu, R., Wells, M.G., 2013. The interaction of large amplitude internal seiches with a
shallow sloping lake bed: observations of benthic turbulence in lake Simcoe, Ontario,
Canada. PLoS One 8 (3), e57444.
Coulter, D.P., Hk, T.O., Mahapatra, C.T., Guffey, S.C., Sepulveda, M.S., 2015a. Fluctuating
water temperatures affect development, physiological responses and cause sex rever-
sal in fathead minnows. Environ. Sci. Technol. 49:19211928. http://dx.doi.org/10.
1111/eff.12227.
Coulter, D.P., Sepulveda, M.S., Troy, C.D., Hk, T.O., 2015b. Species-specic effects of
subdaily temperature uctuations on consumption, growth and stress responses in
two physiologically similar sh species. Ecol. Freshw. Fish. http://dx.doi.org/10.
1111/eff.12227.
Cyr, H., 2012. Temperature variability in shallow littoral sediments of Lake Opeongo (Can-
ada). Freshw. Sci. 31, 895907.
Cyr, H., 2016. Wind-driven thermocline movements affect the colonisation and growth of
Achnanthidium minutissimum, a ubiquitous benthic diatom in lakes. Freshw. Biol. 61
(10), 16551670.
Donaldson, M.R., Cooke, S.J., Patterson, D.A., Macdonald, J.S., 2008. Cold shock and sh.
J. Fish Biol. 73, 14911530.
Donelan, M.A., 1980. Similarity theory applied to the forecasting of wave heights, periods,
and directions. Proc. of the 1980 Canadian Coastal Conference. Natl. Res. Council, Can-
ada, pp. 4761.
Donelan, M.A., Hamilton, J., Hui, W., 1985. Directional spectra of wind-generated waves.
Philos. Trans. R. Soc. A 315 (1534), 509562.
Draft Lake Simcoe Protection Plan, 2009. Ontario Ministry of the Environment, Toronto,
Ontario (January).
Eimers, M.C., Winter, J.G., Scheider, W.A., Watmough, S.A., Nicholls, K.H., 2005. Recent
changes and patterns in the water chemistry of Lake Simcoe. J. Great Lakes Res. 31,
322332.
Emery, A.R., 1970. Fish and craysh mortalities due to an internal seiche in Georgian Bay,
Lake Huron. J. Fish. Res. Board Can. 27, 11651168.
Evans, M.A., MacIntyre, S., Kling, G.W., 2008. Internal wave effects on photosynthesis: ex-
periments, theory, and modeling. Limnol. Oceanogr. 53, 339353.
Evans, D.O., Skinner, J.A., Ron Allen, R., McMurtry, M.J., 2011. Invasion of zebra mussel,
Dreissena polymorpha, in Lake Simcoe. J. Great Lakes Res. 37, 3645.
 R. Cossu et al. / Journal of Great Lakes Research 43 (2017) 689699
Fee, E.J., Hecky, R.E., Kasian, S.E.M., Cruikshank, D.R., 1996. Effects of lake size, water clar-
ity, and climatic variability on mixing depths in Canadian Shield lakes. Limnol.
Oceanogr. 41 (5), 912920.
Ficke, A.D., Myrick, C.A., Hansen, L.J., 2007. Potential impacts of global climate change on
freshwater sheries. Rev. Fish Biol. Fish. 17 (4), 581613.
Finlay, K., Cyr, H., Shuter, B.J., 2001. Spatial and temporal variability in water temperatures
in the littoral zone of a multibasin lake. Can. J. Fish. Aquat. Sci. 58, 609619.
Fischer, H.B., List, E.J., Koh, R.C.Y., Imberger, J., Brooks, N.H., 1979. Mixing in Inland and
Coastal Waters. Academic Press, New York.
Fricker, P.D., Nepf, H.M., 2000. Bathymetry, stratication, and internal seiche structure.
J. Geophys. Res. Oceans 105 (C6), 1423714251.
Gorham, E., Boyce, F.M., 1989. Inuence of lake surface area and depth upon thermal
stratication and the depth of the summer thermocline. J. Great Lakes Res. 15 (2),
233245.
Hkanson, L., 1977. The inuence of wind, fetch, and water depth on the distribution of
sediments in Lake Vanern, Sweden. Can. J. Earth Sci. 14, 397.412.
Hawley, N., Muzzi, R.W., 2003. Observations of nepheloid layers made with an autono-
mous vertical proler. J. Great Lakes Res. 29, 124133.
Hilton, J., Lishman, J.P., Allen, P.V., 1986. The dominant processes of sediment distribution
and focusing in a small, eutrophic, monomictic lake. Limnol. Oceanogr. 31, 125133.
Hlevca, B., Doka, S., Cooke, S., Midwood, J., Wells, M.G., 2015. Characterization of water
temperature variability within a harbour connected to a large lake. J. Great Lakes
Res. 41 (4), 10101023.
Kahma, K.K., Calkoen, C.J., 1992. Reconciling discrepancies in the observed growth of
wind-generated waves. J. Phys. Oceanogr. 22 (12), 13891405.
Ludsin, S.A., DeVanna, K.M., Smith, R.E.H., 2014. Physical-biological coupling and the chal-
lenge of understanding sh recruitment in freshwater lakes. Can. J. Fish. Aquat. Sci.
71, 775794.
Magnuson, J.J., Crowder, L.B., Medvick, P.A., 1979. Temperature as an ecological resource.
Am. Zool. 19, 331343.
North, R.L., Barton, D., Crowe, A.S., Dillon, P.J., Dolson, R.M.L., Evans, D.O., Ginn, B.K.,
Hkanson, L., Hawryshyn, J., Jarjanazi, H., King, J.W., La Rose, J.K.L., Len, L., Lewis,
C.F.M., Liddle, G.E., Lin, Z.H., Longstaffe, F.J., Macdonald, R.A., Molot, L., Ozersky, T.,
Palmer, M.E., Quinlan, R., Rennie, M.D., Robillard, M.M., Rod, D., Rhland, K.M.,
Schwalb, A., Smol, J.P., Stainsby, E., Trumpickas, J.J., Winter, J.G., Young, J.D., 2013.
The state of Lake Simcoe (Ontario, Canada): the effects of multiple stressors on phos-
phorus and oxygen dynamics. Inland Waters 3 (1), 5174.
O'Reilly, C., Sharma, M.S., Gray, D.K., Hampton, S.E., Read, J.S., Rowley, R.J., Schneider, P.,
Lenters, J.D., McIntyre, P.B., Kraemer, B.M., et al., 2015. Rapid and highly variable
warming of lake surface waters around the globe. Geophys. Res. Lett. 42:
1077310781. http://dx.doi.org/10.1002/2015GL066235.
699
Patterson, J.C., Hamblin, P.F., Imberger, J., 1984. Classication and dynamic simulation of
the vertical density structure of lakes. Limnol. Oceanogr. 29:845861. http://dx.doi.
org/10.4319/lo.1984.29.4.0845.
Read, J.S., Hamilton, D.P., Desai, A.R., Rose, K.C., MacIntyre, S., Lenters, J.D., Smyth, R.L.,
Hanson, P.C., Cole, J.J., Staehr, P.A., Rusak, J.A., 2012. Lake-size dependency of wind
shear and convection as controls on gas exchange. Geophys. Res. Lett. 39 (9):
L09405. http://dx.doi.org/10.1029/2012GL051886.
Rowan, D.J., Kalff, J., Rasmussen, J.B., 1992. Estimating the mud deposition boundary depth
in lakes from wave theory. Can. J. Fish. Aquat. Sci. 49 (12), 24902497.
Shintani, T., de la Fuente, A., Nino, Y., Imberger, J., 2010. Generalizations of the
Wedderburn number: parameterizing upwelling in stratied lakes. Limnol. Oceanogr.
55, 13771389.
Spigarelli, S.A., Thommes, M.M., Prepejchal, W., 1982. Feeding, growth, and fat deposition
by brown trout in constant and uctuating temperatures. Trans. Am. Fish. Soc. 111,
199209.
Stainsby, E.A., Winter, J.G., Jarjanazi, H., Paterson, A.M., Evans, D.O., Young, J.D., 2011.
Changes in the thermal stability of Lake Simcoe from 1980 to 2008. J. Great Lakes
Res. 37, 5562.
Strayer, D.L., Power, M.E., Fagan, W.F., Pickett, S.T.A., Belnap, J., 2003. A classication of
ecological boundaries. Bioscience 53, 723729.
Szekeres, P., Eliason, E.J., Lapointe, D., Donaldson, M.R., Brownscombe, J.W., Cooke, S.J.,
2016. On the neglected cold side of climate change and what it means to sh. Clim.
Res. 69 (3), 239245.
Technical Progress Series in Lake Monitoring, 2014. Invasive Dreissena spp. mussels in
Lake Simcoe: distribution, trends, and potential impacts. Technical Progress Series
in Lake Monitoring: Report No. 2 (39 pp.).
Troy, C.D., Ahmed, S., Hawley, N., Goodwell, A., 2012. Cross-shelf thermal variability in
southern Lake Michigan during the stratied periods. J. Geophys. Res. 117, C02028
(16 PP). http://dx.doi.org/10.1029/2011JC007148.
Vanderploeg, H.A., Ludsin, S.A., Ruberg, S.A., Hk, T.O., Pothoven, S.A., Brandt, S.B., Lang,
G.A., Liebig, J.R., Cavaletto, J.F., 2009. Hypoxia affects spatial distributions and overlap
of pelagic sh, zooplankton, and phytoplankton in Lake Erie. J. Exp. Mar. Biol. Ecol.
381, S92S107.
Webb, P.W., 1978. Temperature effects on acceleration of rainbow-trout, salmo-gairdneri.
J. Fish. Res. Board Can. 35 (11), 14171422.
Wells, M.G., Parker, S., 2010. The thermal variability of the waters of Fathom Five National
Marine Park, Lake Huron. J. Great Lakes Res. 36, 570576.