Basin-Scale Gyres and Mesoscale Eddies in Large La

https://doi.org/10.
5194/gmd-2022-104
Preprint. Discussion started: 19 July 2022
c Author(s) 2022. CC BY 4.0 License.
Basin-scale gyres and mesoscale eddies in large lakes: A novel

procedure for their detection and characterization, assessed in Lake
Geneva
Seyed Mahmood Hamze-Ziabari1, Ulrich Lemmin1, Frédéric Soulignac1,2, Mehrshad Foroughan1 and
5 David Andrew Barry1
1
Ecological Engineering Laboratory (ECOL), Environmental Engineering Institute (IIE), Faculty of Architecture, Civil and
Environmental Engineering (ENAC), Ecole Polytechnique Fédérale de Lausanne (EPFL), 1015 Lausanne, Switzerland
2
Commission Internationale pour la Protection des Eaux du Léman (CIPEL), Nyon, Switzerland
Correspondence to: Seyed Mahmood Hamze-Ziabari ([email protected])
10 Abstract. Gyres and eddies, i.e., large-scale rotating coherent water masses, are prominent features of large lakes and oceans,
are formed due to the interplay between Coriolis force and wind stress. Understanding their dynamics is important as they are
known to play a crucial role in spreading bio-chemical materials and energy throughout lakes and oceans. Since field
observations in large lakes are sparse in time and location, often limited to a few moorings, they cannot provide a
comprehensive validation dataset for such large-scale current systems. Previous numerical studies suggested the presence of
15 different and complex gyre systems in many large lakes, however none were confirmed with detailed field measurements. In
order to assess the spatial and temporal extent of gyres and eddies, their dynamics and vertical structure, as well as validate
their prediction in numerical simulation results, transect field observations should be carried out. However, at present it is
difficult to forecast when and where such transect field observations should be taken. To overcome this problem, a novel
procedure combining 3D numerical simulations, statistical analyses, and remote sensing data was developed that permits
20 determination of the spatial and temporal patterns of basin-scale gyres during different seasons. The efficiency and robustness
of the proposed procedure was validated in Lake Geneva. For the first time in a lake, detailed field evidence of the existence
of basin-scale gyres and (sub)mesoscale eddies was provided by data collected along transects whose locations were
predetermined by the proposed procedure. The close correspondence between field observations and detailed numerical results
further confirms the validity of the model for capturing large-scale current circulations as well as submesoscale eddies. The
25 procedure can be applied to other large lakes and can be extended to the interaction of biological-chemical-physical processes.
1 Introduction
Pressure on large lakes, as major freshwater resources, is increasing due to a combination of changing climate and human
activities. To a large extent, lake water quality is controlled by the pathways along which nutrients, phytoplankton and
contaminants are transported/redistributed on meso- or basin-scales (Ralph, 2002); gyres and eddies play an important role in
30 this current pattern. Basin-scale gyres are usually restricted in the vertical by the thermocline depth (Ishikawa et al., 2002; Ji
1
https://doi.org/10.5194/gmd-2022-104
and Jin, 2006), which is also subject to high spatial and temporal variability due to the presence of mesoscale or submesoscale
circulation (Nõges and Kangro, 2005; Ostrovsky and Sukenik, 2008; Corman et al., 2010).
Understanding gyre dynamics in large lakes has mainly been advanced by Csanady’s theoretical analyses (Csanady,
1973, 1975). Wind forcing (the primary driver), the Earth’s rotation (Coriolis force), and lakebed morphology are the main
35 controlling parameters (Birchfield, 1967; Rao and Murty, 1970; Csanady, 1973; Pickett and Rao, 1977). In order to provide
and accumulate enough energy to generate basin-scale circulation such as gyres, wind with a relatively constant direction must
blow over most of the water surface for a certain time (Csanady, 1975). It has been shown that in the Laurentian Great Lakes
in North America, a downwind flow in shallower nearshore regions and an upwind return flow in the deeper part of a lake
basin enhance the formation of such circulation cells. Stratification and the width of downwind flow, which is also known as
40 a “coastal jet”, will increase and confine the flow to the surface layer (Bennett, 1974). However, these studies were conducted
under simplified boundary conditions, using bathymetry and atmospheric forcing data with limited resolution.
Theoretical studies have indicated that the response of a depth-variable lake to a strong uniform wind often leads to
the formation of two counter-rotating gyres, also known as a double-gyre or dipole (Csanady, 1973; Bennett, 1974; Shilo et
al., 2007), as was confirmed by early numerical simulations (Simons, 1980). With increased computational power, three-
45 dimensional (3D) hydrodynamic models with higher accuracy, stability and resolution have since been developed and have
significantly contributed to the understanding of lake circulation, especially that driven by wind (e.g., Beletsky et al., 1999;
Laval et al., 2005; Beletsky and Schwab, 2008; Bai et al., 2013; Mao and Xia, 2020; Baracchini et al., 2020; Lin et al. 2022;
Wu et al. 2022). Recent numerical simulations have also highlighted the role of baroclinicity induced by gradients of surface
buoyancy (mainly surface heating and cooling), which can enhance the large-scale circulation in the Great Lakes (Schwab and
50 Beletsky, 2003; Bennington et al., 2010; Verburg et al., 2011; McKinney et al., 2012).
Based on long-term current data of the Laurentian Great Lakes from a limited number of moorings, Beletsky et al.
(1999) suggested that during the non-stratified season, a single counterclockwise rotating gyre exists in the larger lakes (Lake
Huron, Lake Michigan, Lake Superior), whereas a two-gyre pattern is established in the smaller ones (Lake Ontario, Lake
Erie). Stratification may modify this pattern. A two-gyre pattern was also observed in Lake Okeechobee (USA), a shallow (3.2
55 m mean depth) large lake (Ji and Jin, 2006). In Lake Biwa (Japan), topography, stratification and non-linearity were found to
affect gyre development and could result in a three-gyre pattern (Akitomo et al., 2004). Determining the direct influence of
atmospheric forcing on individual large-scale current systems in large lakes is quite challenging due to the complexity and
potential range of hydrodynamic responses.
In parallel to increased computational capacity, high-resolution satellite images allow direct observations of gyres
60 and mesoscale and submesoscale eddies (Steissberg et al., 2005; Zhan et al., 2014). In particular, Synthetic Aperture Radar
(SAR) imagery is applied to detect and characterize eddies in oceans (e.g., Johannessen et al., 1996; DiGiacomo and Holt,
2001; Marmorino et al., 2010; Qazi et al., 2014). Although small coastal cyclonic eddies were identified in Lake Superior with
SAR imagery (McKinney et al., 2012), it has not yet been used to detect basin-scale gyres in lakes.
2
https://doi.org/10.5194/gmd-2022-104
Numerical simulations of large-scale motions in Lake Geneva (e.g., Umlauf and Lemmin, 2005; Perroud et al., 2009;
65 Le Thi et al., 2012; Razmi et al., 2013; Cimatoribus et al., 2018, 2019; Baracchini et al., 2020; Reiss et al., 2020) show that
the Coriolis force is important in the force balance, and that gyres form. Although the existence of different gyre systems was
suggested by these studies, none were confirmed with detailed field measurements. Based on long-term mooring data, albeit
limited, Lemmin and D’Adamo (1996) suggested the existence of a basin-scale gyre. The formation of smaller scale eddies in
two embayments was observed, affected by the embayment geometry (Razmi et al., 2013) and by meteorological conditions
70 (Razmi et al., 2017).
In past studies, numerical modeling results and observations from moored current meters were analyzed for the
presence of large-scale gyres mainly by studying individual events. However, since field observations in large lakes are
generally sparse in time and space, often limited to a few moorings, they cannot provide a detailed description of the large-
scale gyre pattern (Beletsky et al., 2013; Hui et al., 2021) and, in particular, determine whether these patterns are “typical” and
75 thus important in the long-term development of the lake flow system. Therefore, in order to improve the understanding of the
general circulation in large lakes, an algorithm that allows identifying and tracking basin-scale and mesoscale water mass
movements in a lake is needed. In this context, one of the most popular methods in oceanography to identify and track eddies
from Sea Level Anomaly maps and numerical modeling results is based on the Okubu-Weiss (OW) parameter (e.g., Isern-
Fontanet et al., 2004; Xiu et al., 2010; Chang and Oey, 2014). The OW parameter allows separation of flow fields into vorticity-
80 dominated regions (gyres and eddies) and strain-dominated regions (the ambient flow field) (Okubo, 1970; Weiss, 1991).
In the present study, high-resolution 3D hydrodynamic model results from Lake Geneva are combined with the OW parameter
and Empirical Orthogonal Function (EOF) analysis in a novel procedure with the following objectives:
(i) To detect large-scale coherent flow features in large lakes, in particular large-scale gyres and mesoscale eddies.
(ii) To design strategies for field campaigns that can confirm the existence of these features and thus, for the first
85 time, provide unambiguous detailed field evidence of basin-scale gyres and mesoscale eddies in lakes.
(iii) To identify links between atmospheric forcing patterns and the dominant variability of basin-scale flow
systems that can help to predict the occurrence and the lifetime of the dominant flow patterns.
The Supplementary Information (SI) section provides additional figures, tables, and clarifications and details, denoted with
prefix S.
90 2 Materials and Methods
2.1 Study site
Lake Geneva (Lac Léman), the largest lake in Western Europe, is a crescent-shaped pre-alpine lake, situated between
Switzerland and France (Figure 1). It is composed of an eastern basin, the Grand Lac, with a maximum depth of 309 m and a
western shallow and narrow basin, the Petit Lac, with a maximum depth of nearly 70 m. The lake is approximately 70-km
95 long, has a maximum width of 14 km, a surface area of 580 km2 and a volume of 89 km3. The lake is located between the Alps
3
https://doi.org/10.5194/gmd-2022-104
to the south and east, and the Jura to the northwest. Two strong, dominant wind fields, namely the Bise (coming from the
northeast) and the Vent (coming from the southwest), are guided by the surrounding topography (Wanner and Furger, 1990;
Lemmin and D’Adamo, 1996). The central and western parts of the lake can experience strong wind events, which may last
from several hours to several days. The eastern part of the lake is sheltered from strong wind events by the surrounding high
100 mountains (Rahaghi et al., 2018; Lemmin, 2020).
Coriolis force plays an important role in the formation of gyres, since the width of the lake is much larger than the internal
Rossby radius of deformation; the typical range of the Rossby radius is O(5 km) during strongly stratified seasons, and O(1
km) during weakly stratified seasons (Lemmin and D’Adamo, 1996; Cimatoribus et al., 2018). Numerical simulations show
that basin-scale gyres in Lake Geneva can rotate clockwise (anticyclonic) or counterclockwise (cyclonic) due to the
105 surrounding topography and the two dominant wind fields, Bise and Vent (Figure 1a) (Razmi et al., 2017; Cimatoribus et al.,
2018, 2019).
Lake Geneva is well suited for studying gyre dynamics and testing the new procedure, since it exhibits the full range of
hydrodynamic complexity expected in a large, oligomictic lake with a strong summer and weak winter stratification; occasional
full convective mixing only occurs during exceptionally cold winters. Furthermore, all background data needed for the
110 development of this procedure are available in the public domain.
4
https://doi.org/10.5194/gmd-2022-104
5
https://doi.org/10.5194/gmd-2022-104
Figure 1. (a) Lake Geneva and surrounding topography, adapted from a public domain satellite image (NASA World Wind,
last accessed 2 August 2019) and bathymetry data from SwissTopo (last accessed 2 August 2019). The colorbar indicates the
water depth. SHL2 (yellow cross) is a long-term CIPEL monitoring station where data on physical and biological parameters
are measured. The red square shows the location of the EPFL Buchillon meteorological station (100-m offshore). The black
115 dashed line approximately delimits the two basins of Lake Geneva called the Petit Lac and the Grand Lac. The thick red
arrows indicate the direction of the two strong dominant winds, called the Bise (coming from the northeast) and the Vent
(coming from the southwest). (b) A schematic view of transects (T) selected for the different field campaigns in this study (see
text and Table S1 in SI for details). The background satellite image is taken from Sentinel-1A.
2.2 Hydrodynamic modeling
120 In this study, MITgcm (Marshall et al., 1997), which solves the 3D Boussinesq, hydrostatic Navier-Stokes equations (including
the Coriolis force), was used in a series of numerical simulations for Lake Geneva for 2018 and 2019, based on the model
setup of Cimatoribus et al. (2018). The model was forced by meteorological data (wind fields, atmospheric temperature,
humidity, solar radiation) extracted from the COnsortium for Small-scale MOdeling (COSMO) atmospheric model of
MeteoSwiss (last accessed 6 June 2021) (Voudouri et al., 2017). The first modeling step was a Low Resolution (LR) model
125 (horizontal resolution 173 to 260 m, 35 depth layers, integration time step 20 s), which was initialized from rest using the
temperature profile from CIPEL station SHL2 (CIPEL, 2019) measured on 25 October 2017 and 19 December 2018,
respectively (calm weather conditions prevailed on both dates). For each run, the LR model spin up was ~180 d. The 3D
interpolated results of the LR modeling were then used to initialize the High Resolution (HR) version of the model (horizontal
resolution 113 m and 50 depth layers, with layer thicknesses ranging from 0.30 m at the surface to approximately 12 m for the
130 deepest layer, integration time step of 6 s). For each year, the HR model was run for two seasons: under weakly stratified
conditions and under strongly stratified conditions.
The model’s capability to realistically reproduce stratification, mean flow, and internal seiche variability in Lake Geneva was
demonstrated by Cimatoribus et al. (2018, 2019). To further assess the model’s ability to reproduce seasonal stratification, a
separate validation was carried out in the present study using temperature profiles measured in 2019 at SHL2 (Figure S1 in
135 SI), located at the center of the Grand Lac (Figure 1a).
2.3 Transect field measurements
Based on the proposed procedure, six transects were selected in different parts of the Grand Lac (Figure 1b), i.e., one transect
in the eastern part, and five in the central and western parts (details below). The large-scale gyre in the central part of the
Grand Lac was the main focus of the transect field measurements due to its importance and persistence. Along each transect,
140 ten profiles of current velocity spaced 1 km apart were measured, each for at least 10 min using an Acoustic Doppler Current
Profiler (ADCP, Teledyne Marine Workhorse Sentinel) with the transducer located at 0.5-m depth. The ADCP was equipped
6
https://doi.org/10.5194/gmd-2022-104
with a bottom-tracking module, was set up for 100 1-m bins (blanking distance of 2 m), and operated in high-resolution
processing mode. Tilt and heading angles were derived from sensors located inside the instrument.
2.4 SAR remote sensing
145 Synthetic Aperture Radar (SAR) is frequently used to detect oceanic surface features under light wind conditions (Johannessen
et al., 1996; Wang et al., 2019). The main advantages of SAR imagery are that: (i) it functions day and night under all weather
conditions, (ii) it has high sensitivity to small scale variability of the water surface, and (iii) it provides high resolution images.
The patterns observed in SAR images are due to the change of water surface roughness, which is influenced by wave/current
interactions, natural surface films and spatial variations of the local wind field (Johannessen et al., 2005). Gyres or eddies in
150 SAR images are indicated by dark spiral features called “black” or “classical” eddies (Karimova, 2012). More details are given
in the section S1 in SI. For the present study, C-band SAR data were obtained from the European Space Agency’s (ESA)
Sentinel-1A and Sentinel-1B satellites. The co-polarized VV (Vertical transmit, Vertical receive SAR polarization) data were
used because noise restricts the application of VH (Vertical transmit, Horizontal receive SAR polarization) data (Gao et al.,
2019). The spatial resolution of SAR data varies between 5 and 20 m for a ground sampling distance of 10 m.
155 2.5 Okubo-Weiss parameter and Empirical Orthogonal Functions
The Okubo-Weiss (OW) parameter describes the local strain-vorticity balance in the horizontal flow field of a shallow fluid
layer. It allows separating vorticity-dominated regions associated with basin-scale gyres or mesoscale eddies from strain-
dominated ambient regions (Okubo, 1970; Weiss, 1991) and can be written as:
OW = 𝑆𝑛2 + 𝑆𝑠2 − 𝜔𝑧2 (1)
160 where Sn is the normal component of strain, Ss is the shear component of strain, and ωz is the relative vorticity of the flow,
defined respectively by:
𝜕𝑢 𝜕𝑣
𝑆𝑛 = − (2)
𝜕𝑥 𝜕𝑦
𝜕𝑣 𝜕𝑢
𝑆𝑠 = + (3)
𝜕𝑥 𝜕𝑦
𝜕𝑣 𝜕𝑢
𝜔𝑧 = − (4)
𝜕𝑥 𝜕𝑦
165 where (u,v) are Cartesian components of the horizontal flow field at different depths and times.
For a given data set, EOF analysis identifies the main spatial patterns (E) and their temporal evolution (P) (Wang and
𝑘
An, 2005). Hourly-averaged OW values derived from numerical simulations were decomposed into the basis function 𝐸OW (𝑋)
𝑘
and the principal component coefficient 𝑃OW (𝑡) as:
𝑘 𝑘
OW(𝑋, 𝑡) = ∑𝑁
𝑘=1 𝐸OW (𝑋)𝑃OW (𝑡) (5)
170 where k is the EOF mode, which varies from 1 to the maximum N, X = (x,y) is the position under consideration and t is time.
𝑘 𝑘
The 𝐸OW (𝑋) modes identify the spatial patterns of the OW parameter. The time evolution of each 𝐸OW (𝑋) mode can be
7
https://doi.org/10.5194/gmd-2022-104
𝑘
obtained from the time series of principal component coefficients, indicated as 𝑃OW (𝑡). Positive OW values relate to strain-
dominated regions, whereas negative OW values identify vorticity-dominated regions.
3 Procedure for detecting gyres
175 The proposed gyre identification procedure consists of four steps (Figure 2): (i) data pre-processing, (ii) extracting dominant
patterns using EOF analysis of OW fields, (iii) defining the 3D structure of a gyre, and (iv) finding the correlation between the
dominant gyre pattern and environmental forcing.
In step (i), the OW values are computed for selected time windows using the horizontal velocity fields (u,v) generated
by the 3D numerical modeling. Vorticity-dominated and strain-dominated regions can be identified from OW values, and can
180 then, for step (ii), be used to locate and characterize the gyre flow field, in particular, gyre centers and the outer limits of gyres,
located where vorticity and strain are approximately in balance. This is further supported/confirmed by patterns detected in
SAR images. This OW/EOF analysis is repeated for several depths in step (iii), in order to verify that this surface pattern
documents a gyre extending over the upper layers with its depth limited by the thermocline. Step (iv) links the observed
dominant OW pattern to the forcing which generated it, thereby identifying forcing patterns, mainly wind events, preceding
185 the formation of gyre patterns. Scrutinizing actual meteorological data for such forcing patterns will then allow planning when
and where to take detailed transect field measurements of the gyre pattern in the lake.
3.1 Detection of the core of a gyre or eddy
The separation of the OW field in terms of its sign can be used to detect cores of complex fluid flows (McWilliams, 1984;
Elhmaïdi et al., 1993). These cores are characterized by negative OW values below a given (negative) threshold, OW T
190 (Pasquero et al., 2001; Isern-Fontanet et al., 2006; Henson and Thomas, 2008; Liu et al., 2021). For this purpose, a threshold
value of OWT = 0.2σow was defined, where σow is the root-mean-square fluctuation of OW in the epilimnion with the same sign
of the vorticity of the gyre/eddy core (Pasquero et al., 2001; Isern-Fontanet et al., 2006; Henson and Thomas, 2008).
In the data pre-processing stage, the hourly computed OW values were normalized by hourly values of OW T such that the
normalized OW parameter, OWN, partitions the topology of the gyre/eddy field into three regions: elliptic regions, OW N < -1,
195 hyperbolic regions, OWN > 1, and a background field,│OWN│ ≤ 1 (Pasquero et al., 2001; Isern-Fontanet et al., 2006). As
shown in Figure 2, elliptic regions, which represent the center of gyres/eddies, are significantly more pronounced than the
other regions. In lakes, basin-scale gyres are mainly restricted by the lake basin geometry and also impacted by the surrounding
topography. Therefore, the regions with OW N > -1 can be influenced by the interaction between gyres and nearshore
boundaries. As a result, a wide spatial variability in the OW N values for these regions was observed, whereas regions with
200 OWN < -1 were spatially rather stable. To eliminate such processes from the spatial and temporal identification of gyres/eddies,
the regions with OWN > -1 were filtered before implementing the EOF analysis.
For the proposed procedure (Figure 2), hourly filtered OW N values of different depth layers were computed for each
month. The EOF analysis was then applied to detect the main modes (spatial pattern) and the corresponding principal
8
https://doi.org/10.5194/gmd-2022-104
component time series. From different modes of the EOF results, signatures of gyre or eddy cores were identified based on the
205 following criteria:
1. A local extreme value exists in the spatial modes of OW N values
2. At least one closed line exists around each local extreme value
3. The core edge is identified as the closed line where the sign of 𝐸OWN × 𝑃OWN changes
4. A vertical and horizontal coherence of a gyre or eddy signature must exist between different depth layers (bounded
210 below by the thermocline), i.e., criteria 1-3 above must be spatially consistent for different depth layers.
By applying these criteria, the location of gyre and eddy centers can be identified.
3.2 Detecting the outer boundary of a gyre or an eddy
Gyre boundaries are located where vorticity and strain are approximately in balance, i.e., │OWN│ ≤ 1. However, as discussed
above, the outer gyre boundaries cannot be completely resolved by the OW analysis. Furthermore, the noise in the resulting
215 fields of OWN makes the detection of coherent flow structures difficult (Souza et al., 2011). Defining vertical and horizontal
coherence criteria that can confirm the coherent 3D pattern of gyres can significantly reduce this limitation of the OW
parameter analysis. This threshold-based boundary detection method can be complemented by a geometry-based method using
contour lines in Synthetic Aperture Radar (SAR) images. Since basin-scale gyres and large eddies can be detected in SAR
images, information about the outer boundaries in support of the OW N analysis results can be obtained from SAR imagery.
220 3.3 Finding the link between the temporal variation of a gyre and environmental forcing
Wind stress and surface buoyancy flux (due to heating/cooling) are the processes controlling gyre circulations and their
variability. To find a link between the pattern of external forcing and the computed spatial pattern of the OW N parameter
presented in the previous sections, monthly EOF analyses of the total wind stress ( √𝜏𝑥2 + 𝜏𝑦2 ) and the net upward heat flux
extracted from the COSMO atmospheric data are implemented. In the final stage of the procedure, the lagged cross-correlations
225 between the spatial mode and the corresponding principal component time series of the environment forcing and the OWN are
examined.
9
https://doi.org/10.5194/gmd-2022-104
Figure 2. Flowchart of the proposed procedure for detecting basin-scale gyres and eddies applied to Lake Geneva.
10
https://doi.org/10.5194/gmd-2022-104
4 Results
In this section, the performance of the procedure for detecting the temporal and spatial variations of gyres/eddies is first
230 evaluated for September 2018. Thereafter, the robustness of the proposed procedure is assessed by comparing results of transect
field campaigns that were carried out based on the proposed procedure with numerical results during weakly and strongly
stratified months.
4.1 Detecting gyres and eddies: Okubo-Weiss parameter and EOF analysis
4.1.1 Detection of the location of gyre centers
235 Results of numerical 3D modeling for the month of September 2018 are used to demonstrate the performance of the proposed
procedure. The first and second modes of the OW N variations in different depth layers (1, 5, 10, 15, 20 and 30 m) are given in
Figures 3a,b. Due to the quick convergence of the EOF decomposition modes, only the first two dominant EOF modes are
analyzed for detecting elliptic regions with negative OWN values. The first mode dominates the overall OWN variations, since
it accounts for nearly 56% of the total variance down to 15 m depth, and ~47% for deeper depths. The second mode contributes
240 nearly 23 % to the total variance for the near surface layers and the remaining modes, which each contains less than 5% of the
total variance, are not presented here.
Three closed trajectories that encircle regions with negative 𝐸OWN values can be distinguished in the first mode at different
depth layers (Figure 3a). It should be noted that these regions can indicate the presence of three large-scale gyres in the Grand
Lac, if 𝐸OWN × 𝑃OWN < 0 (the time series of 𝑃OWN are discussed in the following sections). The location of the center of each
245 gyre in different depth layers is calculated by averaging the coordinates of the gyre centers (based on criteria 1-3 as discussed
above). The closed trajectories weaken with depth down to 20 m and disappear at 30-m depth due to the presence of the
thermocline, which is situated ~15-m depth.
A region bounded by closed trajectories with negative OW N can also be considered as an indicator for areas where pelagic
upwelling or downwelling are more likely to occur. Details of the spatial OW N pattern (e.g., Figure 4c, d) confirm the presence
250 of negative OWN regions in the center of the large-scale gyres. The rotation sign of the three-gyre pattern after the Bise event
is given in Figures S5, S6, and S7 for different months. The Anticyclonic Gyre (clockwise rotating) in the west and two
Cyclonic Gyres (counterclockwise rotating) in the center and in the east are hereinafter referred to as AG, CG1, and CG2,
respectively (Figure 3a). The second EOF mode reveals the simultaneous existence of smaller eddies, often with shorter
lifetimes than basin-scale gyres (Figure 3b). More details on these eddies are presented below.
11
https://doi.org/10.5194/gmd-2022-104
255 Figure 3. EOF analysis of the MITgcm output for Lake Geneva for the month of September 2018: (a) First and (b) second
modes of EOF of the normalized OW N parameter are shown for different depth layers. The first mode is dominated by three
large-scale gyres (circular zones of negative OWN values), whereas in the second mode, (sub)mesoscale eddies also appear.
Colorbars give the range of the EOF.
4.1.2 Detection of the outer boundary of a gyre or an eddy
260 The OWN analysis shows that gyre centers have strongly negative OWN values and that positive OWN areas surround the gyres
(Figure 4b, d). The transition zone between the two zones (│OW N│ ≤ 1), which should indicate the outer edge of the gyre, is
wide in certain parts of the circumference and does not allow determining the outer edge of the gyre (Figure 4b, d). SAR
images may be used to confirm/define this edge (more details in S1). The boundaries between the two Cyclonic Gyres in the
12
https://doi.org/10.5194/gmd-2022-104
eastern part of lake, i.e., CG1 and CG2, can be determined from SAR data obtained from Sentinel-1 on 21 July and 12 October
265 2018 (Figure 4a, c), where two (elliptical) gyres are evident. The minor/major axes of CG1 and CG2 are approximately 6.5/12.9
and 9.8/15.9 km, respectively. Smaller eddies, marked by strong negative OW N values in their center surround the large-scale
CG1 and CG2 gyres.
Figure 4. Remote sensing evidence (SAR images, Sentinel-1) of two Cyclonic Gyres (CG) in the eastern part of Lake Geneva:
(a) CG1 for 21 July 2018 and c) CG2 for 12 October 2018. Red dashed lines indicate the dimensions of the gyres. Panels (b)
270 and (d): Corresponding modeled surface velocity fields (small black arrows show sense of rotation) and OW N parameter values
shown by colors. Strong negative OWN parameter values (blue) indicate the location of core zones of large-scale gyres and
mesoscale eddies. The colorbars give the range of the OW N values.
4.1.3 Detecting large-scale gyres: Link between environmental forcing and gyre signature
To find a link between the pattern of external forcing and the computed spatial pattern of the OW N parameter, EOF analyses
275 of the total wind stress and the net upward heat flux extracted from the COSMO atmospheric data for September 2018 were
13
https://doi.org/10.5194/gmd-2022-104
carried out. The first spatial mode and the corresponding principal component time series of the total wind stress
(1) (1) (1) (1)
(𝑐𝑤 (𝑥), 𝜑𝑤 (𝑡)) and the upward heat flux (𝑐𝐻𝐹 (𝑡), 𝜑𝐻𝐹 (𝑡)) are presented in Figures 5a, b, respectively. The first mode of the
external forces dominates the variation of the forces, since it constitutes ~98% and 99% of the total variance related to the total
wind stress and heat flux, respectively. The total variance differences between the first spatial modes and the principal
280 component time series obtained from the solution with two dominant modes and the solution with all possible modes is
negligible (O(10-10-10-14)) compared to the calculated absolute values (O(10 -3)) (Figure S2).
The EPFL Buchillon field station (Figure 1b) data show that a strong Bise event started at ~00:00 on 24 September
2018 and ended at ~11:00 on 26 September 2018. The mean wind speed ±1 standard deviation was 4.21 ± 1.88 m s -1 with wind
gusts of 8.40 ± 3.37 m s-1. The mean wind direction was 61 ± 13° (Figure 5c, d). Figure 5e suggests that there is a link between
(1)
285 the strong Bise event and the first EOF mode of OWN indicated by the principal component (𝑃OW𝑁 (𝑡)) time series. To determine
this link, lagged cross-correlations were computed between the principal component time series of the dominant EOF modes
(1) (1)
of atmospheric forcing, i.e., the total wind stress (𝜑𝑤 (𝑡)) and the net upward heat flux, (𝜑𝐻𝐹 (𝑡)), and the two dominant EOF
(1)
modes of OWN (Figure 6). The lagged cross-correlation between the upward heat flux and 𝑃OW𝑁 (𝑡) does not exceed 0.3 in the
(1)
depth layers influenced by the gyres (i.e., 2-20 m; Figure 3). On the other hand, the cross correlation between 𝜑𝑤 (𝑡) and
(1)
290 𝑃OW𝑁 reaches values > 0.8 in the same depth layers and a positive peak with a lag of 1-1.5 days for the near surface layers (1-
(1) (1)
20 m; Figure 6a-f). A positive peak means that 𝜑𝑤 (𝑡) and 𝑃OW𝑁 (𝑡) have the same sign. These results imply that the three-
gyre pattern in the first mode (Figure. 3a) is predominantly controlled by the spatial and temporal variations of wind stress,
(1) (1) (1)
i.e., 𝑐𝑤 (𝑥) and 𝜑𝑤 (𝑡). Furthermore, the time series of the first mode, 𝑃OW𝑁 (𝑡), in different depth layers suggest that the
three-gyre pattern can persist for several days (nearly 5 days) after the wind peak (Figure 5e).
295 Thereafter (29 September 2018), the large-scale gyres break down into smaller gyres/eddies, and the second mode of
(2)
OWN dominates (Figure 5f). The cross-correlation between wind stress, upward heat flux and 𝑃OW𝑁 (𝑡) is not greater than 0.35
(2)
in the depth layers influenced by the gyres (Figure 6g-l). For wind stress, 𝐸OW𝑁 is excited with a lag of ~ 5.5 days, with the
same sign, and for upward heat flux with a lag of more than 4 days, again, with the same sign. Moreover, negative peaks in
(2) (1) (2)
the cross-correlation between 𝐸OW𝑁 and 𝜑𝐻𝐹 (𝑡) show that 𝐸OW𝑁 can be excited with a lag of 1-2 days, due to differential
(2)
300 heating and cooling during day-night cycles (Figure 6g-l). However, 𝐸OW𝑁 is only weakly excited by wind stress or upward
heat flux. In addition to the effects of environmental forcing presented above, the basin shape and bathymetry also change the
gyre pattern over time.
14
https://doi.org/10.5194/gmd-2022-104
(1) (1)
Figure 5. For September 2018: (a) First spatial mode (𝑐𝑤 (𝑋)) and (b) principal component time series (𝜑𝑤 (𝑡)) of the total
(1) (1)
wind stress. (c) First spatial mode 𝑐𝐻𝐹 (𝑋) and (d) principal component time series (𝜑𝐻𝐹 (𝑡)) of the net upward heat flux. (e)
305 Mean wind speed and wind gusts at the Buchillon station (see Fig. 1 for location) for the period indicated by the red dashed-
lined box in (b) and (d), and the principal component time series of the OW N parameter associated with the first spatial mode
(1)
(𝑃OW𝑁 (𝑡)) in different depth layers. (f) Wind direction at the Buchillon station for the period indicated by the red dashed-lined
15
https://doi.org/10.5194/gmd-2022-104
box in (b) and (d),, and the principal component time series of the OW N parameter associated with the second spatial mode
(2)
(𝑃OW𝑁 (𝑡)) at different depth layers. Colorbars in (a) and (c) show the range of the parameters. The colors in the legends of the
310 two lowest panels correspond to the different depths. In (e) and (f), negative (𝐸OWN 𝑃OWN ) indicates the negative regions of
OWN parameter (the elliptic regions) and the presence of gyres.
(1)
Figure 6. (a)-(f) Lagged cross-correlation of the principal component time series of the total wind stress (𝜑𝑤 (𝑡)) and the net
(1)
upward heat flux (𝜑𝐻𝐹 (𝑡)) with the principal component time series of the OW N parameter associated with the first spatial
16
https://doi.org/10.5194/gmd-2022-104
(1)
315 mode (𝑃OW𝑁 (𝑡)) in different depth layers (depths are indicated). (g)-(l) Lagged cross-correlation of the principal component
(1) (1)
time series of the total wind stress (𝜑𝑤 (𝑡)) and the net upward heat flux (𝜑𝐻𝐹 (𝑡)) with the principal component time series
(2)
of OWN the parameter associated with the second spatial mode (𝑃OW𝑁 (𝑡)) in different depth layers.
4.2 Detecting large-scale gyres: Transect field campaigns
4.2.1 A cyclonic gyre at the center of the Grand Lac basin
320 To validate the proposed procedure for detecting the location and time of basin-scale gyres after a strong wind event, a field
measurement campaign was planned for September 2019 and focused on capturing the largest basin-scale gyre, CG1, in the
central part of the Grand Lac basin (Figures 3, 4). Buchillon field data recorded a strong Bise event that started on 17 September
2019 at ~20:00 and ended on 20 September 2019 at ~13:00 (mean wind speed ±1 standard deviation was 3.65 ± 0.86 m s-1,
with gusts of 7.69 ± 1.78 m s -1 and mean wind direction of 64 ± 11°). Furthermore, the spatial pattern of total wind stress
325 computed from the forecasted COSMO atmospheric data (Figure S3) also revealed the same spatial pattern as observed for the
September 2018 Bise event discussed above. For September 2018, it was shown that the three-gyre system pattern is highly
correlated with strong Bise events (Figures 5 and 6). This pattern forms and persists for several days after the wind event has
ceased. Therefore, based on these results, a field campaign was designed to capture the boundary and temporal variation of
CG1 from 20 to 22 September 2019.
330 Four transects, T2, T2L, T2R, and T2H (see Figure 1b; coordinates are given in Table S1 in SI) were chosen to
investigate the 3D structure of CG1 and its evolution. A few hours after the strong Bise event ended on 20 September,
measurements at 10 points with 1-km spacing along transect T2, from South to North, were taken. The measured velocity field
confirmed the existence of a cyclonic circulation in the center of the Grand Lac (Figures 7a, b).
In the proposed procedure, the center of each gyre is calculated by averaging the coordinates of the centers of the minimum
335 OWN zones at different depth layers. To take into account the uncertainty in selecting the location of each transect, the standard
deviation of coordinates of minimum OWN zones from the average transect was also calculated. It was ~ 1.8 km for September
2018. Therefore, measurements along two additional transects, T2R and T2L, were taken to cover the low velocity core zones
(almost zero) surrounding the center. The velocity field for each transect is given in Figure 7c. The 3D velocity field shows a
cyclonic gyre. The gyre velocity field penetrated down to ~15 m, limited by the strong thermal stratification in September (see
340 Figure S1 for temperature profile at SHL2 near the center of CG1). The maximum horizontal water velocity reaches ~35 cm
s-1 in the near-surface layer. The center of the cyclonic gyre can clearly be seen in all transects by the strong decrease in
horizontal velocity.
In order to capture the gyre boundary and to determine the complete CG1 velocity field, 12 points with 1.5-km spacing were
measured along transect T2H (see Figure 1b for location), from West to East on 22 September. For example, point 1 at the
345 western end of the transect was clearly outside the gyre field because both the velocity magnitude and direction changed.
Transect 2 (T2), from South to North was repeated. The field results (Figure 7d) confirmed the boundaries of CG1, which were
17
https://doi.org/10.5194/gmd-2022-104
identified in the SAR image (see Figure 4a). Furthermore, the velocity field shows that the nearshore (South and North) currents
are much stronger than currents at the East-West boundary of CG1, indicating that the nearshore bathymetry deforms/confines
the gyre flow field.
350 An analysis similar to that of September 2018 discussed above was carried out for September 2019 and the results showed
good agreement (Figure S3). Comparisons between the modeled and observed velocity fields for the September 2019 campaign
are also given in Figure S5, again confirming the close resemblance between the results of field observations and the numerical
modeling.
355 Figure 7. Measured current velocity field for sampling points along the selected transects (T) shown in Figure 1b for (a), (b)
20 September (c) 21 September (d)-(f) 22 September 2019. The coordinates of sampling points in T2 are the same in (a)-(d)
18
https://doi.org/10.5194/gmd-2022-104
and (f). The colorbars indicate horizontal velocity in cm s-1. Positive velocities are pointing eastward for transect T2, and
northward for transect T2H. Positive velocities are pointing eastward for transect T2, and northward for transect T2H.
4.2.2 Observations of two- and three-gyre patterns
360 In most previous numerical studies on Lake Geneva, two basin-scale gyres (a dipole), located in the center of the deep Grand
Lac basin were considered to be the main basin-scale circulation (Le Thi et al., 2012; Razmi et al., 2017). To confirm the
existence of a dipole in Lake Geneva, field measurements were conducted from 23 to 25 October 2019 and from 24 to 26
November 2019 after strong Bise wind events occurred (Figure S4). However, the primary objective of the field campaign on
25 October and 25 November 2019 was to determine whether three large-scale gyres predicted by the EOF analysis (Figures
365 3, 4) exist in Lake Geneva.
The average wind direction during the Bise event in October 2019 was ~69° with a duration of 40 h, and was ~ 47° with a
duration of 94 h in November 2019 (Figure S4). For the October 2019 campaign, two transects, T2 (western part) and T3
(eastern part) (Figure 1b), which consisted of 10 measurement points with 1-km spacing, were selected following the proposed
procedure, but with different transect coordinates (e.g., T2) for each month. To confirm the existence of the third gyre in the
370 eastern part of lake, CG2, two transects, T1 and T1H (see Figure 1b), each with eight points were selected. The distance
between the T1H measurement points was 1.5 km in order to capture the CG2 boundary observed in the SAR image (Figure
4). In the 3D velocity field recorded during the October campaign at different transects (Figure 8), a dipole consisting of an
anticyclonic gyre, AG, at transect T3, and a cyclonic gyre, CG1, at T2 is clearly evident. The maximum velocity at CG1 (22
cm s-1) is stronger than at AG (15 cm s-1), with the velocity field of AG penetrating into deeper layers. The 3D velocity field
375 observed at T1 and T1H confirms the existence of the third cyclonic gyre (CG2) in the eastern part of Lake Geneva (Figure
8d). The CG2 velocity field penetrated down to nearly 20 m and the maximum horizontal water velocity reaches ~17 cm s -1
near the surface layer. The depth of the CG2 gyre is similar to that of CG1 and the magnitude of its velocity is comparable to
AG.
To further confirm the existence of the three gyre pattern in Lake Geneva during weakly stratified months, measurements
380 along the three transects, T1, T2, and T3 (Figure 1b), were carried out in November 2019, based on the EOF analysis for
November 2018. During November, a different location for T3 was chosen because, compared to the October campaign, a
different spatial pattern was observed in the EOF results (see Table S1). The transects T1 (eastern part), T2 (central part), and
T3 (western part) were sampled on the same day, with 9, 10 and 8 points, respectively, each with a 1-km spacing. The measured
3D velocity fields reveal two cyclonic circulations at T1 and T2 and one anticyclonic circulation at T3 (Figure 8g). In contrast
385 to the October campaign, the magnitude and depth of the velocity field at CG2 and CG1 are comparable. Comparisons between
the modeled and observed velocity fields for October and November 2019 campaigns show good agreement (Figures S6 and
S7 in SI).
19
https://doi.org/10.5194/gmd-2022-104
Figure 8. Field campaigns of October and November 2019: (a) the current velocity field (arrows) of gyres AG and CG1 and
390 (b, c) the horizontal velocity, u, for sampling points along the selected transects (T) for the 24 October 2019 campaign. (d) The
current velocity field of CG2 and (e,f) the horizontal components of velocity field, u and v, for sampling points along the
selected transects for the 25 October 2019 campaign. (g) The current velocity field of AG, CG1 and CG2 and (h-j) the
horizontal velocity, u, for sampling points along the selected transects for the November 2019 campaign. The colorbars
indicates horizontal velocity in cm s-1. Positive velocities are pointing eastward for transect T1, T2 and T3, and northward for
395 transect T1H.
4.3 Detecting small eddies
Small Eddies (SE) have not yet been investigated in Lake Geneva. In the present study, smaller-sized coherent structures with
negative OWN values were detected in the vicinity of the basin-scale gyres or at coastal headlands (Figure 3). Two dominant
and frequently occurring patterns of such eddies were selected in the EOF analysis for different months (Figures 9 and 10).
400 The focus here was on eddies with a lifetime of several days, comparable to the lifetime of basin-scale gyres, i.e., SE1 and
SE2 (see Figure 9c for location). In the SAR images, small eddies appear as radar-dark filaments wound into spirals. A close
20
https://doi.org/10.5194/gmd-2022-104
similarity between the patterns observed in the SAR data and numerical results is observed (cf. Figures 9 and 10). According
to the numerical results, the rotation of SE1 is cyclonic and its diameter is of O(5) km during October 2019.
Such small-scale patterns cannot be adequately captured by velocity measurements at fixed points. Therefore, the ADCP was
405 mounted on a small, boat-towed catamaran. Continuous vertical current profiles were measured along the preselected transects.
SE1 was frequently observed by the proposed procedure during October 2018 and 2019 (not shown). A field campaign was
conducted in October 2020 to investigate the vertical structure of SE1. Two transects, T4 and T5, were chosen based on EOF
results and SAR images for the years 2018 and 2019. The horizontal lengths of T4 and T5 were ~8.5 and 3.5 km, respectively.
The spatial resolution of the measured current profiles depends on the boat’s speed (1.5-2 km/h). Velocity data were averaged
410 every minute, which resulted in a 25-33 m spatial resolution along each transect. The numerical results and measured horizontal
velocities at T4 and T5 are given in Figures 9e-h. A cyclonic circulation formed in the southwestern part of lake, and an
anticyclonic circulation in the northwestern part (Figure 9d). There is a close match between the measured and modeled
velocity fields (see Figure 8e, h). Field data indicate that the velocity field of SE1 extended to depths between 35-40 m,
comparable to the depth of the anticyclonic gyre (AG) in this part of the lake measured during the October 2019 campaign.
415 Flow separation in the vicinity of headlands or baroclinicity due to favorable upwelling in the Petit Lac can also lead
to the formation of a submesoscale eddy, such as SE2 (Figure 10). The dimensions of SE2 are generally smaller than those of
SE1. However, SE2 is more frequently observed during certain months of year. For example, the signature of SE2 can be
clearly observed in SAR images taken on 19 August 2018 and 7 November 2018 (Figure 10). The high-resolution numerical
results confirm the presence of cold cyclonic eddies with the same dimensions for the same period (Figures 10b,d). The
420 dimension of SE2 is of O(4 km). A field campaign on 3 September 2020 confirmed the existence of SE2. The simulated and
measured velocity fields indicate a cyclonic circulation in the northwestern part of Grand Lac basin of Lake Geneva (Figures
10e, f). Field data also show that the SE2 velocity field can reach 20-m depth, which is 5-m deeper than the depth of the
cyclonic circulation during the September 2019 campaign. Generally, such small eddies have a transient nature. However,
eddies SE1 and SE2 are trapped between larger-scale gyres (Figure S6) or between boundaries and gyres, and remain at fixed
425 locations for several days.
21
https://doi.org/10.5194/gmd-2022-104
Figure 9. Evidence for the existence of submesoscale eddy (SE1). (a) Remote sensing evidence (SAR images, Sentinel-1) of
SE1 for 27 October 2019. (b) Corresponding modeled surface velocity fields (small black arrows show sense of rotation) and
OW parameter values shown by colors. (c) The first modes of EOF of the normalized OW N parameter for October 2019 with
430 large gyres AG1, CG1 and CG2 and small eddies SE1 and SE2. (d) Modeled surface velocity fields for 19 October 2020 and
22
https://doi.org/10.5194/gmd-2022-104
selected transect locations (T4 and T5). (e) Measured horizontal current velocity profile along T4. (f) Corresponding
modeled horizontal velocity along T4. (g) Measured horizontal current velocity profile along T5. Red dashed-lined rectangle
in (c) indicates the location of zoom panels a), b) and d). Colorbars give the range of the parameters in the panels. Positive
horizontal velocity in cm s-1 is pointing eastward.
435
Figure 10. Evidence for the existence of submesoscale eddy (SE2). (a) Remote sensing evidence (SAR image, Sentinel-1) of
SE2 for 19 August 2018. (b) Corresponding modeled surface velocity fields (small black arrows show sense of rotation) and
temperature parameter values shown by colors. (c) SAR image of SE2 for 7 November 2018. (d) Corresponding modeled
surface velocity fields and temperature for 7 November 2018. (e) The modeled surface velocity and selected transect (T6) for
23
https://doi.org/10.5194/gmd-2022-104
440 field campaign on 3 September 2020. (f) Measured horizontal current velocity along T6. For location of these panels, see
Figure 9c. Colorbars give the range of the parameters in the panels. Positive horizontal velocity in cm s -1 is pointing
eastward.
5 Discussion
5.1 Procedure
445 Large scale gyres and eddies are important transport processes in large lakes allowing the rapid spreading of materials from
the nearshore zone into the lake interior. The present study proposed a procedure that allows locating gyres and small eddies
in a large lake and determining the details of their patterns. Its application to Lake Geneva has shown that a unique relationship
exists between Bise wind forcing events and the resultant three-gyre pattern which is predictable and stable in space and time
for a certain period after the forcing has ceased. The subsequent confirmation of the results by transect field measurements
450 whose location and timing was based on the predicted pattern has demonstrated the feasibility and the robustness of the
procedure. As a result, strategies for such field campaigns can be designed ensuring a high success rate. Without the
information provided by this procedure, it is almost impossible to carry out such field campaigns, since the EOF results
highlighted the complex nature of circulation patterns that exist in Lake Geneva.
5.2 Gyre pattern: Two or three gyres?
455 Numerical studies have shown that large-scale gyres contribute significantly to the transport of water masses and potential
pollutants from the nearshore zone into the interior of the lake (Cimatoribus et al., 2019; Reiss et al., 2020). Particle tracking
used in these studies indicated a rapid spreading of these water masses over large areas of the Grand Lac basin within a short
time, once they were caught in gyres. Previous studies on Lake Geneva suggested that two basin-scale gyres located at the
center of the deep Grand Lac basin drive the main basin-scale circulation (Le Thi et al., 2012; Razmi et al., 2017) and can
460 have an impact on lake bio-chemical-physical interactions (Cotte and Vennemann, 2020). However, no field measurements
were carried out to confirm the existence of a two-gyre pattern observed in these numerical simulation studies. In the present
study, transect field measurements provided detailed evidence of these two gyres.
Furthermore, transect field measurements confirmed that a third gyre located in the eastern part of Grand Lac basin is as
well developed as the other two gyres. Previously, modeling had suggested that this third gyre may be important for the rapid
465 spreading of water masses and potential pollutants brought into the lake by the Rhône River, because its plume directly fed
into this gyre (Lemmin, 2016). Generally, the signature of the cyclonic gyres, CG1 and CG2, in the center and eastern part of
the lake, is more pronounced than that of the anticyclonic gyre, AG, in the west. The western part of lake is directly exposed
to strong wind energy input, whereas the eastern parts are sheltered from strong wind (Rahaghi et al., 2019; Figure 5a).
24
https://doi.org/10.5194/gmd-2022-104
Based on the EOF analysis in the proposed procedure, it was possible in the present study to extend previous individual gyre
470 investigations into a statistically significant gyre pattern. Thus, the link of this pattern to forcing was made evident and
indicated that the location of gyre centers and the duration of the gyre pattern only changed slightly between events.
5.3 Forcing: Importance of spatial heterogeneity
The dominant role of wind as the primary force in gyre formation, the impact of stratification and Coriolis forcing, and the
importance of surface heating and cooling in driving or enhancing gyre flows have recently been highlighted (Hogg and
475 Gayen, 2020). The spatial pattern of the net heat flux suggests that the heterogeneity between heating and cooling in the
eastern part of the lake is more significant than in the western part (Figure 5b). This agrees with findings of spatial heat flux
variability by Rahaghi et al., (2019). The variability may impact on the strength of gyre flows in the eastern part of Lake
Geneva. Further research is needed in order to quantify the role of differential surface buoyancy fluxes on the strength of
cyclonic gyres in the eastern part of lake, particularly during summertime when diurnal heating and cooling are stronger.
480 As shown in Figure 8, the depth influenced by the gyre field (AG) in the western part of the lake is greater than the depths
influenced by CG1 and CG2 located in the center and eastern part of the lake for the November and October campaigns. The
gyre velocity is constrained by the thermocline depth, as previously discussed. The thermocline depth can also be affected by
the spatial heterogeneity of atmospheric forcing and gyre motions. Forcing by wind stress, the primary source of energy for
mixing the water column, is more pronounced in the western part (Figures 5a; S3, and S4 in SI). Consequently, a deeper
485 mixed layer would be expected in the western part of the lake. The lower velocity of gyre flow in the western part of the lake
can be attributed to the fact that the wind energy can penetrate into deeper layers due to a deeper thermocline, whereas it is
confined by a shallower thermocline in the eastern part of lake. As a result, the maximum velocity field of CG1 and CG2 is
generally greater than AG.
5.4 Small eddies
490 The spatial pattern in the OW N computed for October 2019 is slightly different than that of September 2019, i.e., a strong
smaller-scale cyclonic circulation (SE1) exists near gyre AG in the western part of the lake (Figure 9). Small eddies, with
approximately the size of the Rossby radius, are predicted by the numerical modeling and are important in the mixing and
transport of lake water masses. They were most often observed shedding from headlands, trapped between basin-scale gyres,
as well as being ejected from bays (Razmi et al., 2017). Although it is beyond the scope of this study to investigate the origin
495 of small eddies, it was demonstrated that the proposed procedure can detect and locate them correctly, as was confirmed by
transect field measurements. Small eddies can play an important role with respect to sea and ocean turbulence, stratification
and primary production (Mahadevan, 2016). Generally, such small eddies are expected to be transient in nature. However,
eddies SE1 and SE2 were trapped between large-scale gyres or between boundaries and gyres, hardly lost any strength and
remained active at fixed locations for several days, comparable to the lifetime of large-scale gyres.
25
https://doi.org/10.5194/gmd-2022-104
500 The results of this study suggest that by applying the proposed procedure, it is possible to develop strategies for detailed
transect field studies that can be extended to bio-geo-chemical parameters which may have ecological importance.
6 Summary and conclusions
In order to advance the understanding of water mass movement dynamics in large lakes, it is essential to determine the
contribution of basin-scale gyres and (sub)mesoscale eddies. In particular, it is important to know to what extent the flow field
505 generated by gyres and eddies observed during a single forcing event is “typical” and thus key to the longterm development
of the lake flow field. Unfortunately, at present, a procedure that would allow identifying and tracking basin-scale and
mesoscale water mass movements in a wider lake database does not exist. We therefore developed a procedure combining
high-resolution 3D numerical simulations, the Okubo-Weiss (OW) parameter and EOF analysis that can provide direct
evidence of the existence of cyclonic (counterclockwise) and anticyclonic (clockwise) rotating basin-scale gyres and meso-
510 scale eddies in large lakes. Its feasibility and robustness was assessed and confirmed by transect field measurements taken in
Lake Geneva.
The gyre flow field is characterized by a coherent pattern of the normalized Okubo-Weiss parameter, OWN, in different layers
of the lake, as was detected by the EOF analysis. The results showed a clear link between strong large-scale wind events and
the computed spatial patterns in the first mode of the EOF analysis. The procedure allows detecting the location of gyre centers
515 where almost zero horizontal velocity zones indicate the occurrence of pelagic upwelling or downwelling which has a great
impact on the biological-chemical-physical development of large lake ecological systems (Troitskaya et al., 2015).
Furthermore, field observations confirmed for the first time that three persistent gyres, two cyclonic gyres and one anticyclonic
gyre, formed after strong Bise wind events, as was predicted by the proposed procedure. According to the EOF analysis, the
horizontal gyre motion is mainly responsive to the wind stress, whereas the depth of the gyre flow field mainly depends on
520 thermocline depth.
Field observations during October and November 2019 demonstrated that the depth of gyre penetration is greater in the western
(anticyclonic gyre) than in the eastern part of the lake (cyclonic gyres). The spatial inhomogeneities in external forcing can
lead to significant spatial variability of the gyre velocity field in the western and eastern parts of Lake Geneva.
The proposed procedure can also detect (sub)mesoscale eddies, if the resolution of the numerical modeling grid is high enough.
525 These eddies were predominantly cyclonic, and their diameters were of O(4-5 km). They may occur due to flow separation in
the vicinity of headlands and embayments. These eddies are transient, but if trapped between larger-scale gyres, they can
remain at a fixed location for several days, and can have a lifetime comparable to basin-scale gyres. Their patterns were
confirmed by field campaigns designed following the proposed procedure.
This study highlighted the significance of 3D processes in large lakes, thus indicating that 1D concepts cannot adequately
530 describe the dynamics in such lakes. Further research is required to investigate gyre/eddy-formation mechanisms and the role
that eddies and large basin-scale gyres play in the interaction of biological-chemical-physical processes. They may rapidly
spread materials entering the lake in the nearshore zone into the whole lake, and thus affect the long-term ecological system
26
https://doi.org/10.5194/gmd-2022-104
development of large lakes. Although the feasibility and robustness of the proposed procedure was assessed in Lake Geneva,
it can be applied to any large lake with a comparable data base, since it is based on universally valid concepts. It is a powerful
535 tool for designing strategies for detailed field studies and will allow new types of field measurements that can contribute to
advancing the understanding of large lake dynamics.
Acknowledgments
This research was supported by the Swiss National Science Foundation (SNSF Grant 178866). The spatiotemporal
meteorological data were provided by the Federal Office of Meteorology and Climatology in Switzerland (MeteoSwiss). We
540 also extend our appreciation to the Commission Internationale pour la Protection des Eaux du Léman (CIPEL) for in situ
temperature measurements. Water temperature profiles were collected at the CIPEL SHL2 station for 2018-2019 by the Eco-
Informatics ORE INRA Team at the French National Institute for Agricultural Research (SOERE OLA-IS, INRA Thonon-les-
Bains, France).
Data availability
545 The SAR images used in this study are based on Sentinel-1 raw data, which are made available by the ESA and can freely be
downloaded from the ESA's Sentinel data hub (https://scihub.copernicus.eu/). The three-dimensional model used in this study
is based on the MIT General Circulation Model (MITgcm, http://mitgcm.org/, https://doi.org/10.1029/96JC02775), which is
publicly available. The in situ data and numerical configurations supporting the findings of this study are available online at
https://github.com/mahmoodziabar/Data-for-GMD-paper.git.
550 Author contributions
SMHZ planned the field campaign; SMHZ, FS, and MF performed the measurements; SMHZ analyzed the data; SMHZ
wrote the manuscript draft; DAB and UL reviewed and edited the manuscript.
Competing interests
The authors declare that they have no conflict of interest.
27
https://doi.org/10.5194/gmd-2022-104
555 References
Akitomo, K., Kurogi, M., and Kumagai, M.: Numerical study of a thermally induced gyre system in Lake Biwa, Limnology,
5(2), 103-114, https://doi.org/10.1007/s10201-004-0122-9, 2004.
Bai, X., Wang, J., Schwab, D. J., Yang, Y., Luo, L., Leshkevich, G. A., and Liu, S.: Modeling 1993–2008 climatology of
seasonal general circulation and thermal structure in the Great Lakes using FVCOM, Ocean Model., 65, 40-63.
560 https://doi.org/10.1016/j.ocemod.2013.02.003, 2013.
Baracchini, T., Wüest, A., and Bouffard, D.: Meteolakes: an operational online three-dimensional forecasting platform for lake
hydrodynamics, Water Res., 172, 115529, https://doi.org/10.1016/j.watres.2020.115529, 2020.
Baracchini, T., Chu, P. Y., Šukys, J., Lieberherr, G., Wunderle, S., Wüest, A., and Bouffard, D.: Data assimilation of in situ
and satellite remote sensing data to 3D hydrodynamic lake models: a case study using Delft3D-FLOW v4.03 and OpenDA
565 v2.4, Geosci. Model Dev., 13, 1267–1284, https://doi.org/10.5194/gmd-13-1267-2020, 2020.
Beletsky, D., Hawley, N., and Rao, Y. R.: Modeling summer circulation and thermal structure of Lake Erie, J. Geophys. Res.
Oceans., 118(11), 6238-6252, https://doi.org/10.1002/2013JC008854, 2013.
Beletsky, D., Saylor, J. H., and Schwab, D. J.: Mean circulation in the Great Lakes. J. Great Lakes Res., 25(1), 78-93,
https://doi.org/10.1016/S0380-1330(99)70718-5, 1999.
570 Beletsky, D., and Schwab, D.: Climatological circulation in Lake Michigan. Geophys. Res. Lett., 35(21), L21604,
https://doi.org/10.1029/2008GL035773, 2008.
Bennett, J. R.: On the dynamics of wind-driven lake currents, J. Phys. Oceanogr., 4(3), 400-414, https://doi.org/10.1175/1520-
0485(1974)004<0400:OTDOWD>2.0.CO;2, 1974.
Bennington, V., McKinley, G. A., Kimura, N., and Wu, C. H.: General circulation of Lake Superior: Mean, variability, and
575 trends from 1979 to 2006. J. Geophys. Res.: Oceans, 115(C12), C12015. https://doi.org/10.1029/2010JC006261, 2010.
Birchfield, G. E.: Horizontal transport in a rotating basin of parabolic depth profile, J. Geophys. Res., 72(24), 6155-6163,
https://doi.org/10.1029/JZ072i024p06155, 1967.
Chang, Y. L., and Oey, L. Y.: Analysis of STCC eddies using the Okubo–Weiss parameter on model and satellite data, Ocean
Dyn., 64(2), 259-271, https://doi.org/10.1007/s10236-013-0680-7, 2014.
580 Cimatoribus, A. A., Lemmin, U., Bouffard, D., and Barry, D. A.: Nonlinear dynamics of the nearshore boundary layer of a
large lake (Lake Geneva), J. Geophys. Res.: Oceans, 123(2), 1016-1031, https://doi.org/10.1002/2017JC013531, 2018.
Cimatoribus, A. A., Lemmin, U., and Barry, D. A.: Tracking Lagrangian transport in Lake Geneva: A 3D numerical modeling
investigation. Limnol. Oceanogr., 64(3), 1252-1269. https://doi.org/10.1002/lno.11111, 2019.
CIPEL: Rapports sur les études et recherches entreprises dans le bassin lémanique, Campagne 2018, Commission
585 internationale pour la protection des eaux du Léman (CIPEL), Nyon, Switzerland. https://www.cipel.org/wp-
content/uploads/2018/04/RapportScientifique_camp_2016_VF.pdf, 2019.
28
https://doi.org/10.5194/gmd-2022-104
Corman, J.R., McIntyre, P.B., Kuboja, B., Mbemba, W., Fink, D., Wheeler, C.W., Gans, C., Michel, E., and Flecker, A. S.:
Upwelling couples chemical and biological dynamics across the littoral and pelagic zones of Lake Tanganyika, East Africa,
Limnol. Oceanogr., 55(1), 214-224, https://doi.org/10.4319/lo.2010.55.1.0214, 2010.
590 Cotte, G., and Vennemann, T. W.: Mixing of Rhône River water in Lake Geneva: Seasonal tracing using stable isotope
composition of water, J. Great Lakes Res., 46(4), 839-849, https://doi.org/10.1016/j.jglr.2020.05.015, 2020.
Csanady, G. T.: Wind-induced barotropic motions in long lakes, J. Phys. Oceanogr., 3(4), 429-438,
https://doi.org/10.1175/1520-0485(1973)003<0429:WIBMIL>2.0.CO;2, 1973.
Csanady, G. T.: Hydrodynamics of large lakes, Annu. Rev. Fluid Mech., 7(1), 357-386,
595 https://doi.org/10.1146/annurev.fl.07.010175.002041, 1975.
Cushman-Roisin, B., and Beckers, J. M.: Introduction to geophysical fluid dynamics: physical and numerical aspects,
Academic press, 2011.
DiGiacomo, P. M., and Holt, B.: Satellite observations of small coastal ocean eddies in the Southern California Bight, J.
Geophys. Res.: Oceans., 106(C10), 22521-22543, https://doi.org/10.1029/2000JC000728, 2001.
600 Dokken, S. T., and Wahl, T.: Observations of spiral eddies along the Norwegian Coast in ERS SAR images,
http://hdl.handle.net/20.500.12242/1449, 1996.
Donelan, M. A., and Pierson Jr, W. J.: Radar scattering and equilibrium ranges in wind‐generated waves with application to
scatterometry, J. Geophys. Res.: Oceans, 92(C5), 4971-5029, https://doi.org/10.1029/JC092iC05p04971, 1987.
Elhmaïdi, D., Provenzale, A., and Babiano, A.: Elementary topology of two-dimensional turbulence from a Lagrangian
605 viewpoint and single-particle dispersion, J. Fluid Mech., 257, 533-558, https://doi.org/10.1017/S0022112093003192, 1993.
Gao, Y., Guan, C., Sun, J., and Xie, L.: A wind speed retrieval model for Sentinel-1A EW mode cross-polarization images,
Remote Sens., 11(2), 153, https://doi.org/10.3390/rs11020153, 2019.
Henson, S. A., and Thomas, A. C.: A census of oceanic anticyclonic eddies in the Gulf of Alaska. Deep Sea Res., Part I., 55(2),
163-176, https://doi.org/10.1016/j.dsr.2007.11.005, 2008.
610 Hogg, A. M., and Gayen, B.: Ocean gyres driven by surface buoyancy forcing, Geophys. Res. Lett., 47(16), e2020GL088539,
https://doi.org/10.1029/2020GL088539, 2020.
Hui, Y., Farnham, D. J., Atkinson, J. F., Zhu, Z., and Feng, Y.: Circulation in Lake Ontario: numerical and physical model
analysis, J. Hydraul. Eng., 147(8), 05021004, https://doi.org/10.1061/(ASCE)HY.1943-7900.0001908, 2021.
Ishikawa, K., Kumagai, M., Vincent, W. F., Tsujimura, S., and Nakahara, H.: Transport and accumulation of bloom-forming
615 cyanobacteria in a large, mid-latitude lake: the gyre-Microcystis hypothesis, Limnology, 3(2), 87-96,
https://doi.org/10.1007/s102010200010, 2002.
Isern-Fontanet, J., Font, J., García-Ladona, E., Emelianov, M., Millot, C., and Taupier-Letage, I.: Spatial structure of
anticyclonic eddies in the Algerian basin (Mediterranean Sea) analyzed using the Okubo–Weiss parameter, Deep Sea Res.,
Part II, 51(25-26), 3009-3028, https://doi.org/10.1016/j.dsr2.2004.09.013, 2004.
29
https://doi.org/10.5194/gmd-2022-104
620 Isern-Fontanet, J., García-Ladona, E., and Font, J.: Vortices of the Mediterranean Sea: An altimetric perspective. J. Phys.
Oceanogr., 36(1), 87-103, https://doi.org/10.1175/JPO2826.1, 2006.
Ji, Z. G., and Jin, K. R.: Gyres and seiches in a large and shallow lake, J. Great Lakes Res., 32(4), 764-775,
https://doi.org/10.3394/0380-1330(2006)32[764:GASIAL]2.0.CO;2, 2006.
Johannessen, J. A., Kudryavtsev, V., Akimov, D., Eldevik, T., Winther, N., and Chapron, B.: On radar imaging of current
625 features: 2. Mesoscale eddy and current front detection, J. Geophys. Res.: Oceans, 110(C7), C07017,
https://doi.org/10.1029/2004JC002802, 2005.
Johannessen, J. A., Shuchman, R. A., Digranes, G., Lyzenga, D. R., Wackerman, C., Johannessen, O. M., and Vachon, P. W.:
Coastal ocean fronts and eddies imaged with ERS 1 synthetic aperture radar. J. Geophys. Res.: Oceans, 101(C3), 6651-6667,
https://doi.org/10.1029/95JC02962, 1996.
630 Karimova, S.: Spiral eddies in the Baltic, Black and Caspian seas as seen by satellite radar data, Adv. Space Res., 50(8), 1107-
1124, https://doi.org/10.1016/j.asr.2011.10.027, 2012.
Laval, B. E., Imberger, J., and Findikakis, A. N.: Dynamics of a large tropical lake: Lake Maracaibo, Aquat. Sci., 67(3), 337-
349, https://doi.org/10.1007/s00027-005-0778-1, 2005.
Le Thi, A. D., De Pascalis, F., Umgiesser, G., and Wildi, W.: Structure thermique et courantologie du Léman (Thermal
635 structure and circulation patterns of Lake Geneva), Arch. Sci., 65, 65-80, https://archive-ouverte.unige.ch/unige:27717, 2012.
Lemmin, U., Mortimer, C. H., and Bäuerle, E.: Internal seiche dynamics in Lake Geneva, Limnol. Oceanogr., 50(1), 207-216,
https://doi.org/10.4319/lo.2005.50.1.0207, 2005.
Lemmin, U.: Insights into the dynamics of the deep hypolimnion of Lake Geneva as revealed by long‐term temperature,
oxygen, and current measurements, Limnol. Oceanogr., 65(9), 2092-2107, https://doi.org/10.1002/lno.11441, 2020.
640 Lemmin, U. Mouvements des masses d’eau (Water mass movement), In: Lemmin U. (ed.), Dans les abysses du Léman
(Descent into the abyss of Lake Geneva), Presses Polytechniques et Universitaires Romandes (PPUR), Switzerland. ISBN:
978-2-88915-105-9, 2016.
Lemmin, U., and D’Adamo, N.: Summertime winds and direct cyclonic circulation: Observations from Lake Geneva, Ann.
Geophys, 14(11), 1207-1220, https://doi.org/10.1007/s00585-996-1207-z, 1996.
645 Lin, S., Boegman, L., Shan, S., and Mulligan, R.: An automatic lake-model application using near-real-time data forcing:
development of an operational forecast workflow (COASTLINES) for Lake Erie, Geosci. Model Dev., 15, 1331–1353,
https://doi.org/10.5194/gmd-15-1331-2022, 2022.
Liu, F., Zhou, H., and Wen, B.: DEDNet: Offshore eddy detection and location with HF radar by deep learning, Sensors, 21(1),
126, https://doi.org/10.3390/s21010126, 2020.
650 McKinney, P., Holt, B., and Matsumoto, K.: Small eddies observed in Lake Superior using SAR and sea surface temperature
imagery, J. Great Lakes Res., 38(4), 786-797. https://doi.org/10.1016/j.jglr.2012.09.023, 2012.
McWilliams, J. C.: The emergence of isolated coherent vortices in turbulent flow, J. Fluid Mech., 146, 21-43,
https://doi.org/10.1017/S0022112084001750, 1984.
30
https://doi.org/10.5194/gmd-2022-104
Mahadevan, A.: The impact of submesoscale physics on primary productivity of plankton, Ann. Rev. Mar. Sci., 8, 161-184,
655 https://doi.org/10.1146/annurev-marine-010814-015912, 2016.
Mao, M., and Xia, M.: Monthly and episodic dynamics of summer circulation in Lake Michigan, J. Geophys. Res.: Oceans.,
125(6), e2019JC015932, https://doi.org/10.1029/2019JC015932, 2020.
Marmorino, G. O., Holt, B., Molemaker, M. J., DiGiacomo, P. M., and Sletten, M. A.: Airborne synthetic aperture radar
observations of “spiral eddy” slick patterns in the Southern California Bight, J. Geophys. Res.: Oceans., 115(C5), C05010,
660 https://doi.org/10.1029/2009JC005863, 2010.
Marshall, J., Adcroft, A., Hill, C., Perelman, L., and Heisey, C.: A finite‐volume, incompressible Navier Stokes model for
studies of the ocean on parallel computers, J. Geophys. Res.: Oceans., 102(C3), 5753-5766,
https://doi.org/10.1029/96JC02775, 1997.
Moran, M. S., Hymer, D. C., Qi, J., and Kerr, Y.: Comparison of ERS-2 SAR and Landsat TM imagery for monitoring
665 agricultural crop and soil conditions, Remote Sens. Environ., 79(2-3), 243-252, https://doi.org/10.1016/S0034-
4257(01)00276-0, 2002.
Nõges, T., and Kangro, K.: Primary production of phytoplankton in a strongly stratified temperate lake, Hydrobiologia, 547(1),
105-122, https://doi.org/10.1007/1-4020-4363-5_10, 2005.
Okubo, A.: Horizontal dispersion of floatable particles in the vicinity of velocity singularities such as convergences, Deep-Sea
670 Res. Oceanogr. Abstr., 17(3), 445-454, https://doi.org/10.1016/0011-7471(70)90059-8, 1970
Ostrovsky, I., and Sukenik, A.: Spatial heterogeneity of biogeochemical parameters in a subtropical lake, In Monitoring and
Modelling Lakes and Coastal Environments, Springer, Dordrecht, https://doi.org/10.1007/978-1-4020-6646-7_6, 2008.
Pasquero, C., Provenzale, A., and Babiano, A.: Parameterization of dispersion in two-dimensional turbulence, J. Fluid Mech.,
439, 279-303, https://doi.org/10.1017/S0022112001004499, 2001.
675 Perroud, M., Goyette, S., Martynov, A., Beniston, M., and Annevillec, O.: Simulation of multiannual thermal profiles in deep
Lake Geneva: A comparison of one‐dimensional lake models, Limnol. Oceanogr., 54(5), 1574-1594,
https://doi.org/10.4319/lo.2009.54.5.1574, 2009.
Pickett, R.L., and Rao, D.B.: One-and two-gyre circulations in homogeneous lakes. International Field Year for the Great
Lakes Bulletin 19, 45-49, U.S. Dept. of Commerce, National Oceanic and Atmospheric Administration, Rockville, Maryland.
680 https://play.google.com/books/reader?id=AtZ918jAK10C&hl=en_GB&pg=GBS.PA45, 1977.
Qazi, W. A., Emery, W. J., and Fox-Kemper, B.: Computing ocean surface currents over the coastal California current system
using 30-min-lag sequential SAR images, IEEE Trans. Geosci. Electron, 52(12), 7559-7580,
https://doi.org/10.1109/TGRS.2014.2314117, 2014.
Rahaghi, A. I., Lemmin, U., Cimatoribus, A., Bouffard, D., Riffler, M., Wunderle, S., and Barry, D. A.: Improving surface
685 heat flux estimation for a large lake through model optimization and two‐point calibration: The case of Lake Geneva, Limnol.
Oceanogr.: Methods, 16(9), 576-593, https://doi.org/10.1002/lom3.10267, 2018.
31
https://doi.org/10.5194/gmd-2022-104
Rahaghi, A. I., Lemmin, U., Cimatoribus, A. A., and Barry, D. A.: The importance of systematic spatial variability in the
surface heat flux of a large lake: A multiannual analysis for Lake Geneva, Water Resour. Res., 55(12), 10248-10267,
https://doi.org/10.1029/2019WR024954, 2019.
690 Ralph, E. A.: Scales and structures of large lake eddies, Geophys. Res. Lett., 29(24), 30-1.
https://doi.org/10.1029/2001GL014654, 2002.
Rao, D. B., and Murty, T. S.: Calculation of the steady state wind-driven circulations in Lake Ontario, Arch. Meteorol.
Geophys.Bioklimatol., A., 19(2), 195-210, https://doi.org/10.1007/BF02249005, 1970.
Razmi, A. M., Barry, D. A., Bakhtyar, R., Le Dantec, N., Dastgheib, A., Lemmin, U., and Wüest, A.: Current variability in a
695 wide and open lacustrine embayment in Lake Geneva (Switzerland), J. Great Lakes Res., 39(3), 455-465,
https://doi.org/10.1016/j.jglr.2013.06.011, 2013.
Razmi, A. M., Lemmin, U., Bouffard, D., Wüest, A., Uittenbogaard, R. E., and Barry, D. A.: Gyre formation in open and deep
lacustrine embayments: the example of Lake Geneva, Switzerland, Environ. Fluid Mech., 17(3), 415-428.
https://doi.org/10.1007/s10652-016-9494-8, 2017.
700 Reiss, R. S., Lemmin, U., Cimatoribus, A. A., and Barry, D. A.: Wintertime coastal upwelling in Lake Geneva: An efficient
transport process for deepwater renewal in a large, deep lake, J. Geophys. Res.: Oceans, 125(8), e2020JC016095,
https://doi.org/10.1029/2020JC016095, 2020.
Schwab, D. J., and Beletsky, D.: Relative effects of wind stress curl, topography, and stratification on large‐scale circulation
in Lake Michigan, J. Geophys. Res.: Oceans, 108(C2), 3044, https://doi.org/10.1029/2001JC001066, 2003.
705 Shilo, E., Ashkenazy, Y., Rimmer, A., Assouline, S., Katsafados, P., and Mahrer, Y.: Effect of wind variability on topographic
waves: Lake Kinneret case, J. Geophys. Res.: Oceans, 112(C12), https://doi.org/10.1029/2007JC004336, 2007.
Simons, T. J.: Circulation models of lakes and inland seas, Can. Bull. Fish. Aquat. Sci., 203, 146, 1980.
Souza, J. M. A. C., de Boyer Montégut, C., and Le Traon, P. Y.: Comparison between three implementations of automatic
identification algorithms for the quantification and characterization of mesoscale eddies in the South Atlantic Ocean, Ocean
710 Sci., 7(3), 317-334, https://doi.org/10.5194/os-7-317-2011, 2011.
Steissberg, T. E., Hook, S. J., and Schladow, S. G.: Measuring surface currents in lakes with high spatial resolution thermal
infrared imagery, Geophys. Res. Lett., 32(11), L11402, https://doi.org/10.1029/2005GL022912, 2005.
Troitskaya, E., Blinov, V., Ivanov, V., Zhdanov, A., Gnatovsky, R., Sutyrina, E., and Shimaraev, M.: Cyclonic circulation and
upwelling in Lake Baikal, Aquat. Sci., 77(2), 171-182, https://doi.org/10.1007/s00027-014-0361-8, 2015.
715 Umlauf, L., and Lemmin, U.: Interbasin exchange and mixing in the hypolimnion of a large lake: The role of long internal
waves, Limnol. Oceanogr., 50(5), 1601-1611, https://doi.org/10.4319/lo.2005.50.5.1601, 2005.
Voudouri, A., Avgoustoglou, E., and Kaufmann, P.: Impacts of observational data assimilation on operational forecasts,
Perspectives on atmospheric sciences, 143-149, Springer, Cham, https://doi.org/10.1007/978-3-319-35095-0_21, 2017.
Verburg, P., Antenucci, J. P., and Hecky, R. E.: Differential cooling drives large‐scale convective circulation in Lake
720 Tanganyika, Limnol. Oceanogr., 56(3), 910-926, https://doi.org/10.4319/lo.2011.56.3.0910, 2011.
32
https://doi.org/10.5194/gmd-2022-104
Wang, B., and An, S. I.: A method for detecting season‐dependent modes of climate variability: S‐EOF analysis, Geophys.
Res. Lett., 32(15), L15710, https://doi.org/10.1029/2005GL022709, 2005.
Wang, C., Tandeo, P., Mouche, A., Stopa, J.E., Gressani, V., Longepe, N., Vandemark, D., Foster, R.C., and Chapron, B.:
Classification of the global Sentinel-1 SAR vignettes for ocean surface process studies, Remote Sens. Environ., 234, 111457.
725 https://doi.org/10.1016/j.rse.2019.111457, 2019.
Wanner, H., and Furger, M.: The Bise—Climatology of a regional wind north of the Alps, Physics Meteorol. Atmos. Phys.,
43(1), 105-115, https://doi.org/10.1007/BF01028113, 1990.
Weiss, J.: The dynamics of enstrophy transfer in two-dimensional hydrodynamics, Phys. D: Nonlinear Phenom., 48(2-3), 273-
294, https://doi.org/10.1016/0167-2789(91)90088-Q, 1991.
730 Wu, T., Qin, B., Huang, A., Sheng, Y., Feng, S., and Casenave, C.: Reconsideration of wind stress, wind waves, and turbulence
in simulating wind-driven currents of shallow lakes in the Wave and Current Coupled Model (WCCM) version 1.0, Geosci.
Model Dev., 15(2), 745-769, https://doi.org/10.5194/gmd-15-745-2022, 2022.
Xiu, P., Chai, F., Shi, L., Xue, H., and Chao, Y.: A census of eddy activities in the South China Sea during 1993–2007, J.
Geophys. Res.: Oceans., 115(C3), C03012, https://doi.org/10.1029/2009JC005657, 2010.
735 Yamaguchi, S., and Kawamura, H.: SAR-imaged spiral eddies in Mutsu Bay and their dynamic and kinematic models, J.
Oceanogr., 65(4), 525-539, https://doi.org/10.1007/s10872-009-0045-5, 2009.
Zhan, S., Beck, R. A., Hinkel, K. M., Liu, H., and Jones, B. M.: Spatio-temporal analysis of gyres in oriented lakes on the
Arctic Coastal Plain of northern Alaska based on remotely sensed images, Remote Sens., 6(10), 9170-9193,
https://doi.org/10.3390/rs6109170, 2014.
740
33

Basin-Scale Gyres and Mesoscale Eddies in Large La

Uploaded by

Copyright:

Available Formats

Basin-Scale Gyres and Mesoscale Eddies in Large La

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Basin-Scale Gyres and Mesoscale Eddies in Large La

Uploaded by

Copyright:

Available Formats

https://doi.org/10.

Basin-scale gyres and mesoscale eddies in large lakes: A novel

90 2 Materials and Methods

2.1 Study site

2.2 Hydrodynamic modeling

2.3 Transect field measurements

2.4 SAR remote sensing

155 2.5 Okubo-Weiss parameter and Empirical Orthogonal Functions

3 Procedure for detecting gyres

3.1 Detection of the core of a gyre or eddy

3.2 Detecting the outer boundary of a gyre or an eddy

4.1.1 Detection of the location of gyre centers

4.1.2 Detection of the outer boundary of a gyre or an eddy

4.2 Detecting large-scale gyres: Transect field campaigns

4.2.1 A cyclonic gyre at the center of the Grand Lac basin

4.2.2 Observations of two- and three-gyre patterns

4.3 Detecting small eddies

5.2 Gyre pattern: Two or three gyres?

5.3 Forcing: Importance of spatial heterogeneity

5.4 Small eddies

6 Summary and conclusions

550 Author contributions

The authors declare that they have no conflict of interest.

You might also like