Content uploaded by Xinfeng Liang
Author content
All content in this area was uploaded by Xinfeng Liang on Jul 07, 2023
Content may be subject to copyright.
1
1
Near-Inertial Oscillations in the Deep Gulf of Mexico
2
3
Yingli Zhu, Xinfeng Liang
4
School of Marine Science and Policy, University of Delaware
5
6
7
Address: 700 Pilottown Road, Lewes, DE, US, 19958
8
9
Email: Yingli Zhu, Yingli.Zhu@colorado.edu
10
Xinfeng Liang, xfliang@udel.edu
11
12
Corresponding author: Yingli Zhu, Yingli.Zhu@colorado.edu
13
14
15
16
17
18
19
20
21
22
23
24
Revised Manuscript (clean version) Click here to view linked References
2
Abstract
25
Near-inertial oscillations (NIOs) are important for maintaining turbulent mixing that affects
26
ocean circulation, biogeochemistry, and climate. The spatial and temporal variability of NIOs in
27
the deep part of the Gulf of Mexico (GoM) has rarely been reported. In this study, a collection of
28
moored current observations was used to examine the spatiotemporal variability of NIOs in the
29
GoM. In the upper layer (0-800 m), a strong seasonal variability of NIOs appears in the eastern
30
GoM with larger amplitudes in winter than in summer and is attributed to the seasonal variability
31
of surface winds and mixed layer depth. In the bottom layer (within 400 m above the bottom),
32
strong NIOs are found in the middle and eastern GoM and show weaker seasonal variability.
33
While winds still matter for the seasonal variability of NIOs in the bottom layer in the eastern
34
GoM, low-frequency flows generally play a more important role in regulating NIOs through
35
interactions with topography, particularly on the continental slope.
36
37
Keywords: Near-inertial oscillation; deep ocean; surface winds; topography; Gulf of Mexico
38
39
40
1. Introduction
41
Near-inertial oscillations (NIOs) are ubiquitous in the global ocean (e.g., Fu 1981; Alford
42
2003; Alford and Whitmont 2007; Alford et al. 2016). Owing to their high kinetic energy with
43
strong vertical shear, NIOs are critical for the energy budget of the ocean and the maintenance of
44
abyssal stratification (Munk and Wunsch 1998). They also substantially contribute to energy
45
dissipation and turbulent mixing (e.g., Jing et al. 2015; Jordi et al. 2016; Cyriac et al. 2021) that
46
affect biogeochemistry and climate (Jochum et al. 2013).
47
NIOs can be generated locally by wind forcing at the sea surface (e.g., Pollard and Millard
48
1970; Alford 2001; Alford and Whitmont 2007), wave-wave interactions (e.g., MacKinnon and
49
Winters 2005), geostrophic adjustment (e.g., van Aken et al. 2005; Jaimes and Shay 2010; Liang
50
and Wunsch 2015), interactions between geostrophic currents and topography (e.g., Nikurashin
51
and Ferrari 2010a, 2010b; Liang and Thurnherr 2012; Brearley et al. 2013; Sheen et al. 2013;
52
Sun et al. 2016) and interactions between barotropic tide and topography (Hendershott, 1973).
53
3
The surface wind forcing is suggested to be responsible for the seasonal variability and the large-
54
scale spatial pattern of NIOs around the world ocean (e.g., Alford and Whitmont 2007). The
55
temporal and spatial variability of NIOs in the upper ocean particularly depends on the wind
56
forcing. Nevertheless, much wind-induced near-inertial energy at the sea surface is lost to
57
turbulent mixing in the upper ocean (upper 200 m) and a relatively small amount of near-inertial
58
energy input into the mixed layer can penetrate into the deep ocean (below 800-1000 m; Furuichi
59
et al. 2008; Zhai et al. 2009; Alford et al. 2012; Cuypers et al. 2013).
60
In the Gulf of Mexico (GoM), previous studies show that high-frequency currents in the deep
61
water of the GoM are dominated by NIOs (e.g., Hamilton et al. 2003; Donohue et al. 2006;
62
Hamilton et al. 2014; Mariano et al. 2016). Also, NIOs are found important for transporting
63
dissolved oxygen and hydrocarbons in the benthic ocean (Martens et al. 2016) and affecting the
64
offshore structures used by the oil/gas industry in the GoM (Spencer et al. 2016). The near-
65
inertial oceanic response to wind forcing has been observed from shelf to offshore in the GoM
66
(e.g., Chen et al. 1996; Hamilton et al. 2000; DiMarco et al. 2000; Hamilton et al. 2003;
67
Donohue et al. 2006; Jarosz et al. 2007; Gough et al. 2016). For example, sea breeze can lead to
68
large near-inertial ocean response due to the coincidence of the period of sea breeze and the
69
inertial period of the ocean (Zhang et al., 2010). In particular, intense NIOs have been observed
70
as hurricanes traveled across the GoM (Brooks 1983; Shay et al. 1989, 1990, 1992, 1998; Shay
71
and Uhlhorn 2008; Jaimes and Shay 2009, 2010; Pallàs-Sanz et al. 2016). Hurricane-generated
72
NIOs in the GoM can propagate vertically down to the seafloor, but the magnitude varies with
73
location (e.g., Hamilton et al. 2000). In addition to wind forcing, eddy-eddy interactions
74
(Hamilton et al. 2000; Hamilton et al. 2003) and the passage of eddy-related fronts (Sheinbaum
75
et al. 2007) could induce strong NIOs in the GoM.
76
Low-frequency currents can modulate or generate NIOs in the deep ocean. The energy
77
transfer from low-frequency flows to NIOs in the subthermocline has been reported in the GoM
78
(Jing et al. 2018). In addition, with strong deep-ocean currents, intense internal waves including
79
NIOs could be generated through the interaction between bottom flows and topography, which
80
have been found in many regions around the global ocean such as the East Pacific, the Southern
81
Ocean, and the South China Sea (Liang and Thurnherr 2012; Brearley et al. 2013; Sheen et al.
82
4
2013; Sun et al. 2016). Similarly, NIOs generated at the bottom level might exist in the GoM
83
(Polzin et al., 2021), which will be investigated in this study.
84
Although many intermittent NIO events have been reported, the spatial and temporal
85
variability of near-inertial currents in the deep part of the Gulf of Mexico (GoM) has rarely been
86
reported. In this study, a collection of moored current observations was used to present the
87
spatiotemporal variability of NIOs in the deep part of the GoM. The relationships between the
88
NIO amplitude and surface winds, topography, and low-frequency flows were also examined.
89
The paper is organized as follows: Data and methods are described in Section 2. The spatial and
90
temporal variability of NIOs in the GoM and forcing mechanisms including surface wind stress,
91
topography, and low-frequency flows are investigated in Section 3. Conclusions and discussion
92
are presented in Section 4.
93
2. Data and methods
94
2.1. Data
95
In this study, oceanic current records collected from nine research programs that were carried
96
out after 1993 were used to examine NIOs. The current records are archived in the Gulf of
97
Mexico Research Initiative Information and Data Cooperative (GRIIDC) and the National
98
Oceanic and Atmospheric Administration (NOAA) National Centers for Environmental
99
Information (NCEI). Current measurements were recorded with either Acoustic Doppler Current
100
Profilers (ADCPs) or single-point recording current meters (RCMs). Most of the collected data
101
were quality controlled in a similar way as described in Zhu and Liang (2020). Specifically,
102
velocities larger than eight standard deviations of each record were removed. Records that had a
103
maximum velocity below 1000 m larger than 1.2 m/s were removed, because the large deep-
104
ocean velocity (>1.2 m/s) is not consistent with surrounding observations and is likely due to
105
instrument errors. Records with completeness less than 90% or with a length shorter than 100
106
days were not selected. Time steps of the moored velocity data range from 0.5 to 4.5 hours. To
107
focus on NIOs, we only chose the moored velocity records with time steps no longer than 3
108
hours and further interpolated them into hourly data.
109
Table 1 shows the observation period, region, and the number of the selected moorings from the
110
nine selected programs. A total of 85 moorings deployed in water deeper than 1000 m were
111
selected. The moorings were deployed over the northern and the eastern GoM during various
112
5
periods between 1997 and 2013 (Fig. 1). The temporal coverage of the current observations is
113
not uniform, ranging from March 1997 to April 1999 (Program 1), August 1999 to September
114
2001 (Program 2), February 2000 to October 2004 (Program 3), May 2000 to July 2007
115
(Program 4), February 2003 to April 2004 (Program 5), March 2004 to July 2005 (Program 6),
116
January 2005 to January 2006 (Program 7), April 2009 to November 2011 (Program 8), July
117
2012 to July 2013 (Program 9). Note that although moorings have been deployed over the
118
southwestern GoM (Tenreiro et al. 2018), those data were not accessible at the time of this study.
119
Figure 2 shows the vertical distribution of the mooring observations. Vertically high-resolution
120
measurements are mainly in the upper 500 m at 43 moorings but are rarely found below 1000 m.
121
Prior to our analyses, tidal currents estimated with the UTide Matlab Package (Codiga 2011)
122
were removed from the current records.
123
124
Fig. 1. Locations of the collected moorings with different colors representing different
125
research program IDs listed in Table 1. The mooring used in Fig. 3, Fig. 5 and Fig. 10 is labeled
126
with the green triangle. The thin black contours indicate 1000, 2000, and 3000 m isobaths. The
127
thick gray contour marks the temporal mean SSH for 1993-2019 at 0.5 m to represent the mean
128
location of the Loop Current (LC).
129
130
Program ID
Program name (Reference)
Region
Period
Mooring
number
6
1
DeSoto Canyon Eddy Intrusion Studya
(Hamilton et al. 2000)
Northeastern GoM
March 1997 to
April 1999
3
2
Deepwater Observations in the Northern
GoM from In Situ Current Meters and
PIESa (Hamilton et al. 2003)
Sigsbee Escarpment
August 1999 to
September 2001
5
3
Physical Oceanographic Time Series Data
from Louisiana State Universitya (McKone
et al. 2007)
Northern GoM
February 2000 to
October 2004
4
4
Observation of the Deepwater
Manifestation of the Loop Current and
Loop Current Rings in the Eastern GoMa
(Welsh et al. 2009)
Eastern GoM
May 2000 to
July 2007
5
5
Exploratory Study of Deepwater Currents in
the Northern GoMa (Donohue et al. 2006)
Northern GoM
February 2003 to
April 2004
19
6
Northwest GoM Studya (Donohue et al.
2008)
Northwestern GoM
March 2004 to
July 2005
13
7
Study of Deepwater Currents in the Eastern
Gulf of Mexicob (Cox et al. 2010)
Eastern GoM
January 2005 to
January 2006
8
8
Dynamics of the Loop Currentb (Hamilton
et al. 2014)
Eastern GoM
April 2009 to
November 2011
23
9
GoM Integrated Spill Response
Consortiumc (DiMarco 2014)
Northern GoM
July 2012 to
July 2013
5
Table 1. Mooring arrays from nine selected research programs. aFunded by the U.S.
131
Department of the Interior, Minerals Management Service. bFunded by the U.S. Department of
132
the Interior, Bureau of Ocean Energy Management. cFunded by the Gulf of Mexico Research
133
Initiative.
134
135
7
136
Fig. 2. Vertical distributions of velocity measurements from the selected moorings. Different
137
colors represent the research program IDs listed in Table 1. The colors are the same as those used
138
in Fig. 1. The black solid line represents the seafloor. The solid and dashed gray lines indicate
139
the lower boundary of the upper layer (800 m below surface) and the upper boundary of the
140
bottom layer (400 m above seafloor), respectively.
141
142
To mark the Loop Current (LC) and Loop Current Eddy (LCE), we used satellite observed
143
sea surface height (SSH) provided by the Copernicus Marine Environment Monitoring Service
144
(CMEMS). This SSH dataset from 1993 to 2019 has a daily time interval and a spatial resolution
145
of 0.25° (Pujol et al. 2016). In addition, the GoM bathymetry data distributed by GRIIDC
146
(Panagiotis 2014) was used to calculate topography steepness that controls the dynamics of
147
interaction between low-frequency currents and topography.
148
A slab mixed-layer model (see details in Section 2.4) was used to investigate the role of
149
surface winds in the NIO variability. For this purpose, wind stress and mixed layer depth (MLD)
150
were required. In this study, wind stress was derived from the gridded surface wind of CCMP
151
(Cross-Calibrated Multi-Platform) Version 2.0 (Atlas et al. 2011) with the bulk parameterization
152
method (Large and Yeager, 2004). This wind product produced by RSS (Remote Sensing
153
Systems) has a spatial resolution of 0.25° and a temporal interval of 6 hours from 1997 to 2013.
154
The climatology of MLD was obtained from the Word Ocean Atlas 2018 (WOA18). Note that
155
the climatology variability of MLD cannot capture the variability on higher-resolution temporal
156
8
and spatial scales. For example, MLD could have more variability in regions with rich eddies,
157
which can induce different NIO variability obtained in the slab mixed-layer model (see details in
158
Section 2.4).
159
The WOA18 temperature (Locarnini et al. 2018) and salinity (Zweng et al. 2018) were used
160
to calculate the annual mean buoyancy frequency for Wentzel-Kramers-Brillouin (WKB) scaling
161
of the NIO amplitude (see details in Section 2.3).
162
2.2. Near-inertial velocities
163
Our focus in this study is the inertial variability near the local Coriolis frequency, f. Rotary
164
spectrum analysis has been widely used to examine the energy distribution of clockwise (CW)
165
and counterclockwise (CCW) current vector fields in the frequency domain (e.g., Mooers 1973).
166
We can identify NIOs in current fields as the components that rotate clockwise in the Northern
167
Hemisphere.
168
Here, we present a sample analysis using one current record obtained at 3217 m from one
169
mooring in the eastern GoM (marked with the green triangle in Fig. 1). Figure 3a shows the
170
detided velocities, which display variability on multiple time scales. The rotary spectra were
171
computed using multitaper spectral analysis with the 31 lowest-order Slepian tapers (Lilly 2019).
172
Figure 3d shows the rotary spectra of the velocities, displaying a larger CW component than
173
CCW component near the local Coriolis frequency. The Eulerian inertial frequency, the observed
174
frequency with the CW spectral peak at the local spatial point, was found within the frequency
175
band from 0.7f to 1.3f. Given that spectra follow a chi-square distribution with 62 degrees of
176
freedom arising from using 31 Slepian tapers, the near-inertial spectral peak is significant within
177
the 99% confidence levels. By applying the same analysis to all the current records, 2843 of the
178
total 2865 selected current records show significant NIOs.
179
180
9
181
Fig. 3. (a) Detided observed zonal (blue) and meridional (red) velocities (cm/s) from 19 April
182
2003 to 11 June 2004. (b) Observed zonal near-inertial velocities (cm/s) obtained with the
183
complex demodulation at 3217 m from one mooring in the eastern GoM (marked with the green
184
triangle in Fig. 1). (c) Same as (b) but for the meridional near-inertial velocities. (d) Rotary
185
spectra of the velocities shown in (a) at frequencies above 0.25 cycle/day. CW and CCW power
186
spectrum density (PSD) is denoted by the red and blue lines, respectively. The local Coriolis
187
frequency, f, is marked with the black solid line. The green asterisk indicates the maximum CW
188
PSD within the frequency band from 0.7f to 1.3f marked with two black dash lines. (e) Rotary
189
spectra with the near-inertial velocities removed. The 99% confidence levels of the spectra are
190
denoted by the vertical red bars.
191
10
192
NIOs represented by the CW rotating current component were further isolated by applying
193
the complex demodulation (Mooers 1973; Shay and Elsberry 1987) at the observed Eulerian
194
inertial frequency with a significant CW spectral peak. The velocity used in the complex
195
demodulation is the full velocity with the tide components removed. It has an hourly interval and
196
is represented as
, where and is the zonal and meridional velocity
197
component, respectively. First, the current vector time series were multiplied by the function
198
, where is the observed Eulerian frequency, t is time and , yielding
199
. Second,
was low-pass filtered with a half-amplitude response at 5 days. Choosing
200
a cutoff period of 5 days in the low-pass filter can isolate most of the NIO components
201
represented by the higher energy than the background spectrum around the inertial frequency
202
(Fig. 3d-e).
203
Finally, the amplitude () and phase () of the CW rotating component were obtained from
204
the low-pass filtered velocity vector
with periods larger than 5 days. The NIO velocity vector
205
was reconstructed from the time series of amplitude and phase with
. The
206
zonal and meridional velocity component of reconstructed NIOs, and , is the real
207
and imaginary part of
, respectively. Figure 3b-c shows the near-inertial velocities isolated
208
from the detided sampled current data (Fig. 3a) from 19 April 2003 to 11 June 2004 at 3217 m
209
from one mooring in the eastern GoM. Wave packets represent the low-frequency variability of
210
the NIO amplitude. We also calculated and presented the rotary spectra of velocities after
211
removing the near-inertial velocities (Fig. 3e). The CW spectra component near the inertial
212
frequency was mostly removed, confirming that NIOs were successfully isolated with the
213
complex demodulation.
214
2.3. WKB scaling
215
Variations in stratification can affect the NIO amplitude and the vertical wavenumber when
216
they propagate vertically (Leaman and Sanford 1975). The influence of stratification on the
217
change of NIO amplitude can be delineated through the WKB scaling. In this study, each current
218
record was WKB scaled using climatological buoyancy frequency calculated with the WOA18
219
data as:
220
11
221
where N is the buoyancy frequency and N0 is the vertical mean of N. One example of the scaling
222
factor,
, at one mooring (marked with the green triangle in Fig. 1) in the eastern GoM is
223
shown in Fig. 4. The scaling factor estimated from the climatological buoyancy frequency
224
generally decreases with depth from 2.4 to 0.4. We also compared WKB scaling factors
225
estimated from the WOA18 data and the available ship CTD data. They generally show similar
226
estimates in our study region but with relatively large differences in the upper 200 m (20±25% of
227
the WOA18 estimate; see details in Appendix A). It should be noted that different estimates of
228
the scaling factor may cause differences in the WKB-scaled near-inertial velocities.
229
230
Fig. 4. Square root of the normalized buoyancy frequency obtained from WOA18 at one
231
mooring (marked with the green triangle in Fig. 1) in the eastern GoM.
232
233
2.4. Slab model of NIOs driven by winds
234
A slab model (Pollard and Millard 1970; D’Asaro 1985) has been used to predict the inertial
235
oceanic response to wind forcing in the northeastern GoM (Gough et al. 2016). The same model
236
was used in this study. Following the notation of D’Asaro (1985), the model equation of inertial
237
oscillations is written in a complex form:
238
12
239
where is the inertial velocity in the mixed layer with a depth of ,
240
is the wind stress, and is a complex variable composed of a decay parameter, , and
241
the local Coriolis frequency, . The damping parameter is a parameterization of energy transfer
242
from the mixed layer to the deeper ocean and was set at , which was used in Gough et al.
243
(2016). MLD was obtained from the WOA18 data. was obtained iteratively with D’Asaro’s
244
solution (D’Asaro 1985):
245
246
where and are inertial velocities at times and , respectively, and is the
247
change in over the time interval .
248
3. Results
249
3.1. Seasonal and spatial variability
250
The amplitudes of NIOs obtained from the moored current data were used to examine the
251
temporal and spatial variability of NIOs. Here we first show results from one sample mooring in
252
the eastern GoM (marked with the green triangle in Fig. 1). The monthly NIO amplitude for May
253
2003 to May 2004 averaged within depth bins (200-m interval) from 0 to 3600 m is presented in
254
Fig. 5a. NIO amplitude is larger in the upper 800 m than in deeper layers, and is larger in January,
255
February, November, and December than other months (Fig. 5a). The WKB-scaled NIO
256
amplitude shows comparable and even larger values in deep layers below 2000 m than in the
257
upper layers, likely due to the relatively weak stratification in deep layers (Fig. 5b). In particular,
258
NIO amplitude in the bottom level is larger than the other levels when the WKB scaling is
259
considered.
260
13
261
Fig. 5. (a) Monthly observed NIO amplitude (cm/s) at one mooring in the eastern GoM
262
(marked with the green triangle in Fig. 1) in vertical bins from 0 to 3600 meter below surface
263
(MBS) from January to December. (b) Same as (a) but for WKB-scaled NIO amplitude (cm/s).
264
The monthly-depth NIO amplitude was estimated based on the NIO amplitude for May 2003 to
265
May 2004.
266
267
To examine the mean characteristics of the seasonal and vertical variability of the NIO in
268
the GoM, the monthly WKB-scaled NIO amplitude at all selected moorings was averaged within
269
depth bins (200-m interval) both from the surface to bottom and from the bottom to surface. NIO
270
amplitude from the surface (bottom) provides reasonable ways to examine seasonal and vertical
271
patterns of NIOs in the upper (bottom) layer.
272
273
14
274
Fig. 6. (a) WKB-scaled observed NIO amplitude (cm/s) averaged over all moorings in
275
month-depth bins from the surface to 800 meter below surface (MBS). (b) WKB-scaled observed
276
NIO amplitude (cm/s) averaged over all moorings in month-depth bins from the bottom to 400
277
meter above bottom (MAB). (c) Standard deviation of the monthly WKB-scaled observed NIO
278
amplitude (cm/s) over all moorings in the month-depth bins from the surface to 800 MBS. (d)
279
Standard deviation of the monthly WKB-scaled observed NIO amplitude (cm/s) over all
280
moorings in month-depth bins from the bottom to 400 MAB. (e) Observation number (Nobs) in
281
each month-depth bin from the surface to 800 MBS. (f) Observation number (Nobs) in each
282
month-depth bin from the bottom to 400 MAB.
283
284
Because moorings were deployed in water with a depth ranging from 1035 to 3580 m, the
285
topography could have different influences on the NIOs observed in different moorings. A larger
286
portion of NIOs from shallower moorings were likely affected by topographic effects than those
287
from deep moorings. Therefore, the role of topography in affecting NIOs varies with depth and
288
cannot be easily distinguished from the role of surface forcing, especially in deep layers.
289
Therefore, we did not discuss the mean WKB-scaled NIO amplitude in these deep layers but
290
15
focused on the variability of NIOs in the upper 800 m (Fig. 6a), where the influence of
291
topography is likely small compared to deep layers. Generally, the amplitude of the WKB-scaled
292
NIO in the upper 800 m is large in winter (December, January, and February) and small in
293
summer (June, July, and August). The largest values appear at around 800 meter below surface
294
(MBS; Fig. 6a).
295
To examine the possible impacts of topography and bottom current, we also presented the
296
mean WKB-scaled NIO amplitude at vertical levels viewed above from the seafloor (Fig. 6b).
297
Because the NIOs close to the seafloor are less likely affected by the surface forcing, we focused
298
on NIOs in the bottom layer (0-400 meter above bottom (MAB)). The amplitude of NIOs in the
299
bottom layer shows relatively weak seasonal variability. In addition, we examined the
300
seasonality of NIO amplitude in the bottom layer versus water depth at each mooring (not
301
shown), however, no apparent relationship between the water depth and seasonality of bottom-
302
layer NIO amplitude is found.
303
The standard deviation of WKB-scaled NIO amplitude across the moorings in the GoM was
304
also estimated within the vertical bins and from January to December (Fig. 6c-d). The standard
305
deviation shows seasonal variability in the upper layer (Fig. 6c), indicating that the seasonal
306
variability of NIOs varies spatially in the GoM. The standard deviation in the bottom layer is
307
also large (Fig. 6d), suggesting that NIO amplitude in the bottom layer varies significantly
308
between different moorings. As a reference, the numbers of the daily NIO amplitude
309
observations in each gridded depth bin were also presented (Fig. 6e, f). It should be noted that
310
the number of total observations in each depth bin was much more in the upper 800 m than the
311
bottom 400 m (Fig. 6e).
312
To demonstrate the spatial variability of NIO in the GoM more clearly, WKB-scaled NIO
313
amplitude was binned into regular spatial grids and grouped into two vertical layers as shown in
314
Fig. 6, the upper layer (0-800 MBS) and the bottom layer (0-400 MAB). The mean and standard
315
deviation of NIO amplitude were also calculated within each spatial bin for the two layers over
316
the summer and winter time. Here, we focused on the summer and winter time when the weakest
317
and strongest NIO amplitude was observed (Fig. 6).
318
319
16
320
Fig. 7. (a-b) Mean WKB-scaled observed NIO amplitude (cm/s) in the upper layer (0-800
321
MBS) during summer (June, July, and August) and winter times (December, January, and
322
February), respectively. (c-d) Standard deviations of WKB-scaled observed NIO amplitude (cm/s)
323
in the upper layer during summer and winter times, respectively. (e-f) Observation number in
324
each spatial bin (Nobs) in the upper layer during summer and winter times, respectively.
325
Contours indicate 1000 and 3000 m isobaths.
326
327
In the upper layer, the mean NIO amplitude is large in the northeast GoM close to the De
328
Soto Canyon and the western GoM slope in summer, while they are larger in the eastern GoM
329
and the northwestern GoM in winter than in summer (Fig. 7a-c). Hamilton et al. (2014) also
330
found the seasonal variability of NIO amplitude in the LC region with larger amplitude in winter,
331
which is consistent with our results. As we show in the next section, the seasonality of NIO in
332
the upper layer is largely related to surface winds. In addition, the standard deviation of NIO
333
amplitude was also calculated using the available data in each season and each spatial bin in the
334
17
upper layer. The standard deviation of NIO amplitude represents the variability on multiple
335
temporal scales spanning from days to years and shows similar spatial patterns as the temporal
336
mean values (Fig. 7c-d). The temporal variability of NIOs in the summer season is larger in the
337
northeast corner of the GoM and on the middle continental slope of the western GoM, while the
338
temporal variability in the winter season is large in the deep eastern GoM and northwestern GoM.
339
Note that the number of observations used for the calculation is also not spatially uniform (Fig.
340
7e-f).
341
342
Fig. 8. (a-b) Mean WKB-scaled observed NIO amplitude (cm/s) in the bottom layer (0-400
343
MAB) during summer (June, July, and August) and winter times (December, January, and
344
February), respectively. (c-d) Standard deviations of WKB-scaled observed NIO amplitude (cm/s)
345
in the bottom layer during summer and winter times, respectively. (e-f) Observation number in
346
each spatial bin (Nobs) in the bottom layer during summer and winter times, respectively.
347
Contours indicate 1000 and 3000 m isobaths.
348
349
18
In the bottom layer, the mean amplitude of NIOs is large on the Sigsbee Escarpment, the
350
Mississippi Fan, and east of Campeche Bank in both summer and winter (Fig. 8a-b). Therefore,
351
the seasonal variability in the bottom layer is weak on the eastern GoM slope, indicating a
352
relatively small influence of surface winds on the seasonal timescale. Nevertheless, NIOs have a
353
larger amplitude in winter than in summer in the bottom layer west of the Florida Escarpment
354
and on the northwestern GoM slope. The clear winter enhancement of NIOs has been observed
355
in the Northern Hemisphere for latitudes 25°-45°N at all depths (Alford and Whitmont 2007). In
356
addition, the standard deviation of NIO amplitude in the bottom layer was calculated using the
357
available data in each season and each spatial bin, showing similar spatial patterns as the
358
temporal mean values (Fig. 8c-d). The temporal variability of NIO amplitude in the bottom layer
359
is also large on the continental slope. Note that only a few observations were used in some
360
spatial grids in the bottom layer (Fig. 8e-f).
361
3.2. Forcing mechanisms of NIOs
362
3.2.1. Wind forcing
363
Surface winds are one of the most important energy sources for NIOs and could be
364
responsible for the temporal and spatial patterns of NIOs revealed in the last section. Here, the
365
influence of wind forcing on the NIO variability in the GoM was examined with the slab model
366
described in Section 2.4. The amplitude of the simulated NIOs was low-pass filtered with a half-
367
amplitude response at 5 days which is the same as that used in filtering observed NIO amplitude.
368
369
19
370
Fig. 9. Average observed NIO amplitude (cm/s) in the upper 100 m in summer (a) and winter
371
(b). Simulated NIO amplitude (cm/s) with the slab model at the mooring locations and
372
observation times in summer (c) and winter (d). The slab model was applied in the surface mixed
373
layer. Contours indicate 1000 and 3000 m isobaths.
374
375
Figure 9 shows the observed NIO amplitude in the upper 100 m and the simulated NIO
376
amplitude obtained from the slab mixed-layer model. The slab model predicts the seasonal and
377
spatial variability of NIOs in the GoM reasonably well (Fig. 9c-d). Both the slab model and
378
observations in the upper 100 m show that NIO amplitude is larger in winter than in summer in
379
the eastern GoM. Moreover, in summer the NIO amplitude in the eastern GoM is smaller than on
380
the northern and western continental slope, while in winter the largest NIO amplitude is located
381
in the eastern GoM. The similar spatial pattern of observed and simulated NIO amplitude
382
indicates that over many regions of the GoM the seasonal and spatial variability of NIO in the
383
upper layer is associated with wind forcing. Additionally, in this study we used the
384
climatological MLD in the slab model. Sensitivity tests of using a constant MLD and MLD with
385
different resolutions in the slab model (Appendix B) suggest that the seasonal variability of MLD
386
is important for the seasonal and spatial pattern of NIOs as well. In addition, extreme weather
387
events such as hurricanes also have influence on the seasonal pattern of NIO amplitude
388
(Appendix B), showing larger amplitude in the northeastern and southeastern GoM in summer. It
389
20
should also be noted that multiple causes such as uncertainties in the wind analysis product and
390
uncertainties of the MLD estimate could contribute to the differences between the simulated and
391
the observed NIOs in the GoM (Appendix B).
392
393
Fig. 10. Time-depth section of NIO zonal velocities observed by the mooring in the eastern
394
GoM (marked with the green triangle in Fig. 1) from 22 January 2006 to 3 February 2006.
395
396
The comparison of the observed bottom NIO amplitude (Fig. 8a-b) and simulated mixed-
397
layer NIO amplitude (Fig. 9c-d) also suggests that the seasonal variability of NIOs in the bottom
398
layer in the eastern GoM west of the Florida Escarpment could be related to surface winds. It is
399
clear that both the slab model simulations for the mixed-layer NIOs and observations in the
400
bottom layer show larger NIO amplitude in winter than in summer in the eastern GoM west of
401
the Florida Escarpment (Fig. 8a-b, 9c-d). For example, strong NIOs were observed in winter of
402
2006 in the eastern GoM, and the zonal velocity of NIOs shows upward phase propagation and
403
downward energy propagation (Fig. 10), indicating that NIOs generated at the surface can
404
propagate downward to the bottom ocean. Although part of wind-induced near-inertial energy is
405
lost to turbulent mixing within the top 200 m (Zhai et al. 2009), some wind-induced near-inertial
406
energy at the sea surface can be drained locally into the deep ocean (e.g., Zhai et al. 2007). In
407
addition, inertial current energy tends to be trapped in negative vorticities such as anticyclonic
408
eddies, known as 'eddy trapping' (e.g., Kunze 1985; Zhai et al. 2007) and in the high-velocity
409
region when NIOs propagate against the background flow due to Doppler-shifting by the mean
410
flow (e.g., Kunze 1985; Kunze 1986; Kunze et al. 1995; Jing et al. 2015). The deep eastern GoM
411
located west of the Florida Escarpment is frequently under the influence of the LC that is a high-
412
velocity current and generates negative relative vorticity within the LC region (mean position of
413
21
the LC is marked in Fig. 1). Therefore, the wind-induced NIOs west of Florida Escarpment were
414
likely drained into the bottom ocean.
415
3.2.2. Influence of topography and low-frequency currents
416
The above analyses indicate that surface winds can largely account for the seasonal and
417
spatial patterns of NIO in the upper layer over many regions of the GoM and in the bottom layer
418
west of the Florida Escarpment. However, they cannot explain most of the temporal and spatial
419
patterns of the NIO amplitude in the bottom layer, particularly on continental slopes.
420
Topographic features and background low-frequency currents could be the controlling
421
mechanism for NIOs in the bottom layer.
422
The influence of low-frequency currents on NIOs was first examined with correlations
423
between the low-frequency current speed and NIO amplitude. Note that the length of current
424
records varies with location, which spans from 100 days to several years. Correlations calculated
425
from records with different lengths cannot be directly compared due to the influence of different
426
record lengths. Therefore, we first calculated the correlations over 90-day moving segments that
427
are available for all the selected mooring observations. Second, we averaged the moving
428
correlations to reduce noise. The length of the moving segment was selected as 90 days because
429
we attempted to reduce the influence of observation lengths on the correlation values and NIOs
430
were likely modulated on the mesoscale scale (e.g., Brearley et al. 2013; Liang and Thurnherr
431
2012).
432
Figure 11 shows the 90-day moving correlations that were averaged and mapped to regular
433
grids in the upper layer (0-800 MBS) and the bottom layer (0-400 MAB). The correlation is
434
small in most regions of the upper layer (Fig. 11a) and is substantially increased in the bottom
435
layer, particularly on the continental slope such as the Sigsbee Escarpment, east of the Campeche
436
Bank, and the northwestern GoM slope (Fig. 11b). The distribution of high correlations indicates
437
that low-frequency flows are important for the bottom layer NIOs on the continental slope.
438
Correlations between the low-frequency current speed and NIO amplitude in the bottom 200 m
439
over their entire observational period were also calculated with mooring records longer than 300
440
days (Fig. 11c). The relatively large and significant correlations are mostly located on the
441
continental slope, especially above the Sigsbee Escarpment, confirming the significant role of
442
low-frequency currents in NIO amplitude with the long records. Additionally, low-frequency
443
22
currents in the bottom do not show strong seasonal variability (not shown). The observed weak
444
seasonal variability of NIOs on the continental slope (Fig. 8a-b) is therefore consistent with the
445
expectation of low-frequency currents controlling the NIO variation. Nevertheless, correlations
446
in the deep eastern GoM (west of the Florida Escarpment) are small, confirming that the seasonal
447
variability of NIOs in the deep eastern GoM is not related to the low-frequency flows but to wind
448
forcings as we discussed in Section 3.2.1.
449
450
Fig. 11. Average moving correlations between the observed low-frequency current speed and
451
NIO amplitude (a) in the upper 800 m (0-800 MBS), and (b) in the bottom 400 m (0-400 MAB).
452
(c) Correlations between the observed low-frequency current speed and NIO amplitude at the
453
bottom level between 200 m above the bottom and the seafloor (0-200 MAB) at the selected
454
mooring with observation length longer than 300 days. Those significant at the 5% significance
455
level are enclosed by black circles. Contours indicate 1000 and 3000 m isobaths.
456
457
To further examine the relationship between the low-frequency currents and NIOs, we
458
selected and examined one mooring on the northern continental slope during a period when it
459
was directly affected by the LC and the LCE. As indicated by SSH, the LC penetrated northward
460
to the mooring location at the end of February 2001 and stayed on the slope in March (Fig. 12a).
461
The LC subsequently shed one LCE in April and the LCE affected the mooring in April and May
462
2001 (Fig. 12b).
463
464
23
465
Fig. 12. Sea surface height (SSH, colored) on (a) 2 March 2001 and on (b) 3 May 2001. One
466
mooring on the northern continental slope is marked with the magenta dot. The LC and the LCE
467
follow the high SSH gradients in a clockwise direction and are marked at 0.5-m SSH with the
468
thick gray line. Thin black contours represent 1000, 2000, and 3000 m isobaths. Current
469
measurements at the mooring are shown in Figs. 13-14.
470
471
472
24
Fig. 13. (a) NIO amplitude (cm/s) obtained from the slab model applied in the surface mixed
473
layer at the mooring on the northern continental slope (marked with the magenta dot in Fig. 12)
474
from 5 September 2000 to 31 August 2001. (b) Time-depth section of observed WKB-scaled
475
NIO amplitude (cm/s) at the mooring. The blue rectangle marks the time-depth section shown in
476
Fig. 14. (c) Time-depth section of observed low-frequency current speed (period>5 days) at the
477
mooring. (d) Observed low-frequency current speed (red) and WKB-scaled NIO amplitude (blue)
478
at 1979 m at the mooring. The greed dots mark the period when the steepness parameter is in the
479
range of 0.3<S<0.7. The vertical black lines mark the starting and ending dates when the
480
mooring was directly affected by the LC and the subsequently shed LCE.
481
482
483
Fig. 14. Time-depth section of NIO zonal velocities observed by the mooring in the northern
484
GoM in the upper 330 m from 14 March 2001 to 24 March 2001 (the section is marked by the
485
blue rectangle in Fig. 13b).
486
487
Figure 13 presents the NIO amplitude and low-frequency current speed at the mooring on the
488
continental slope (marked with the magenta dot in Fig. 12) from 5 September 2000 to 31 August
489
2001. The slab model shows varying NIO amplitude in the mixed layer (Fig. 13a). Consistent
490
with the results shown in section 3.2.1, NIOs generated at the surface show upward phase
491
propagation, indicating downward energy propagation in the upper layer (Fig. 14). However,
492
strong bottom NIOs were observed over periods (April-August 2001) when surface forcing was
493
not always the strongest (Fig. 13a-b), confirming forcings other than surface winds are important
494
for the variability of bottom NIOs on the continental slope.
495
25
As the LC penetrated close to the mooring site after 20 February 2001 (Fig. 12a), the current
496
speed increased to over 50 cm/s in the upper 300 m and decreased with depth (Fig. 13c). When
497
the mooring was affected by the LC and the LCE, relatively strong currents appeared in the deep
498
and bottom layers. Correlations between the current speed at the upper 800 m and the bottom
499
level (at 1979 m) are about 0.37 (above 5% significance level). Note that the observation level of
500
1979 is 22 MAB and could be more easily influenced by the topography. Given the significant
501
correlation between upper-layer and bottom current speeds when the LC and the LCE moved to
502
the mooring sites, the LC and the LCE can induce a strong current response at the bottom, which
503
has been observed in previous studies (e.g., Hamilton 2009; Zhu and Liang 2020). Although the
504
LC and the LCE did not extend from the surface to the bottom, they squeezed the lower layer
505
and induced negative relative vorticity to conserve the potential vorticity (Donohue et al. 2008;
506
Tenreiro et al. 2018). This is how the surface and bottom currents are linked.
507
We also found that the strong low-frequency currents under the LC and the LCE correspond
508
to enhanced NIOs (Fig. 13b-c). The correlation between the low-frequency current speed and
509
NIO amplitude is 0.36 at 1979 m (significant at the 5% significance level) and is larger than that
510
at 1400 and 1800 m. One possible explanation is that NIOs were generated from lee waves that
511
were excited by low-frequency flows impinging on the steep topography (e.g., Nikurashin and
512
Ferrari 2010a, 2010b). The characteristic of lee waves generated through flow-topography
513
interaction is a function of the steepness parameter (e.g., Nikurashin and Ferrari 2010a, 2010b),
514
, where is the bottom stratification, h is the characteristic topographic height, and
515
is the bottom low-frequency geostrophic velocity. For S<0.3, radiated waves are characterized by
516
stationary lee waves and NIOs grow slowly. For 0.3<S<0.7, NIOs grow rapidly due to resonant
517
feedback between the low-frequency flows and internal waves, increasing the amount of radiated
518
energy and making the wave field time-dependent. For S>0.7, some flows are blocked by
519
topography and the energy flux saturates (e.g., Nikurashin and Ferrari 2010a, 2010b).
520
Here we calculated the steepness parameter at this mooring with
, where the
521
characteristic topographic height h was estimated from the topography data. Because geostrophic
522
flows radiate stationary lee waves over topography with horizontal scales
where is
523
the horizontal wavenumber, the topographic features in the radiative wavenumber range are the
524
26
most important for the estimation of h. At the mooring (Fig. 12), the average bottom low-
525
frequency velocity, , was about during the observation period and was increased to
526
about when the LC or the LCE passed the mooring between 20 February 2001 and 13
527
May 2001 (Fig. 13d). Given and the climatology of bottom stratification
528
, the radiative wavelength range of lee waves is between 300 and 18800 m.
529
Because the horizontal resolution of the topography data is 1000 m (Panagiotis 2014), we
530
focused on the topography features on the horizontal scale between 1000 and 20000 m by high-
531
pass filtering the topography data with a cutoff scale of 20000 m. We calculated the root mean
532
square of the high-pass filtered topography that is less than 30 km away from the mooring to
533
represent the characteristic topography height. The characteristic topography height is about 55
534
m and was used to calculate S. When the LC and the LCE moved close to the mooring and
535
induced strong low-frequency bottom currents between 20 February 2001 and 13 May 2001, S is
536
between 0.3 and 0.7 during most of the time. NIOs can thus be generated rapidly through the
537
interaction between low-frequency flows and the topography (Fig. 13d). Low-frequency flows
538
thus can regulate NIOs in the bottom layer on the continental slope.
539
4. Conclusions and discussion
540
In this study, we present the spatial and temporal variability of NIOs in the deep GoM by
541
examining a collection of moored current observations. In the upper layer (0-800 MBS), NIOs
542
have a strong seasonal variability in the eastern and northwestern GoM with the largest
543
amplitude in winter and the smallest amplitude in summer. These temporal and spatial patterns
544
are largely determined by the distributions of wind stress and MLD. In the bottom layer (0-400
545
MAB), strong NIOs appear mostly over the middle and eastern GoM and show a much weaker
546
seasonal variability. Although wind-generated NIOs can propagate down to the ocean bottom
547
and affect bottom NIOs in the deep eastern GoM (west of the Florida Escarpment), low-
548
frequency flows are generally more important in regulating NIOs on the continental slope,
549
especially on the Sigsbee Escarpment. In particular, when the LC and the LCE reach the
550
continental slope and induce strong bottom low-frequency flows, NIOs in the bottom layer grow
551
rapidly through the interaction between bottom low-frequency flows and topographic features
552
over the continental slope.
553
27
One interesting finding of this study is that NIOs in the bottom layer were enhanced on the
554
continental slope due to the interaction between bottom low-frequency currents and continental
555
topography. The bottom low-frequency currents in that region have also been related to the LC
556
and the LCE that dominate the upper-layer circulation in the GoM (e.g., Liu et al., 2016;
557
Weisberg and Liu, 2017) through the mechanisms such as baroclinic instability and potential
558
vorticity conversation of the deep layer (e.g., Hamilton 2009; Donohue et al. 2016; Tenreiro et al.
559
2018; Yang et al., 2020; Zhu and Liang 2020). Therefore, the movement of the LC and the LCE
560
over the continental slope should be important for the NIO generation in the deep GoM.
561
Although it is difficult to regularly measure currents in the deep ocean, it is much easier to
562
estimate surface currents from regular satellite SSH observations. Based on the connection
563
between the LC and the LCE and low-frequency bottom currents as well as the connection
564
between low-frequency bottom currents and NIOs, monitoring the strong surface currents related
565
to the LC and the LCE may help us predict the times and places of NIO generation in the bottom
566
GoM. Furthermore, the spatial and temporal variability of NIOs in the deep part of the GoM
567
revealed in the observations provide a context for understanding the distribution of energy
568
dissipation and turbulent mixing in the deep ocean that can affect biogeochemistry.
569
Even though the current study presents the seasonal and spatial variability of NIOs in the
570
GoM, it is largely based on limited observations at moorings in the eastern and northern GoM.
571
Longer records and mooring observations in the southern GoM are needed in future studies to
572
give a complete picture of NIO distribution in the GoM. In addition, although the slab model
573
suggests that surface winds can largely account for the temporal and spatial variability of NIOs
574
in many regions of the upper GoM, the slab model underestimates the intensity of NIOs due to
575
the relatively low resolution of the wind analysis product and uncertainties of model parameters
576
such as MLD (Appendix B). We would have a better prediction of NIO variability, at least in the
577
upper layer, if high-resolution wind products (Fig. B3 in Appendix B) and more accurate
578
estimates of model parameters were available. Moreover, it is difficult to clearly isolate the
579
topographic effects from the wind forcing in observations, but numerical experiments with high-
580
resolution models may help future studies to isolate the individual mechanisms. High-resolution
581
models can also provide more variables such as the temperature, salinity and velocity that can be
582
used to estimate energy transfer and momentum flux between the LC/LCE and NIOs at deep
583
ocean, which will be used in our future works.
584
28
Abbreviations
585
GoM: Gulf of Mexico
586
NIO: Near-inertial oscillation
587
MLD: Mixed layer depth
588
WKB: Wentzel-Kramers-Brillouin
589
MAB: Meter above bottom
590
MBS: Meter below surface
591
LC: Loop Current
592
LCE: Loop Current eddy
593
594
Acknowledgments
595
We thank three anonymous reviewers for their constructive comments and suggestions. The
596
work was supported in part by the Gulf of Mexico Research Initiative through Grant G-231804
597
and the National Aeronautics and Space Administration through Grant 80NSSC20K0757.
598
599
Data Availability Statement
600
SSH data were produced and provided by CMEMS. The SSH fields used in this study are
601
available at https://dx.doi.org/10.7266/n7-ya5j-0t90 (Zhu 2020a). Historical mooring data in the
602
GoM were collected from GRIIDC and NOAA/NCEI. The subsurface current data from the
603
moorings are freely accessible at https://dx.doi.org/10.7266/n7-7g3d-9q79 (Zhu 2020b). CCMP
604
Version-2.0 vector wind analyses were produced by Remote Sensing Systems. Data are available
605
at www.remss.com (Atlas et al. 2011). Wind data from NDBC buoy 42003 were obtained from
606
https://www.ncei.noaa.gov/thredds-
607
ocean/catalog/ndbc/cmanwx/2010/01/catalog.html?dataset=ndbc/cmanwx/2010/01/42003_20100
608
1.nc. The monthly climatology of MLD, temperature (Locarnini et al. 2018) and salinity (Zweng
609
et al. 2018) was provided by WOA18 at https://www.nodc.noaa.gov/cgi-
610
29
bin/OC5/woa18/woa18.pl. The bathymetry data were downloaded from GRIIDC at
611
https://data.gulfresearchinitiative.org/data/R1.x138.080:0014 (Panagiotis 2014).
612
613
APPENDIX A
614
Validation of WKB scaling factor
615
To validate the scaling factor calculated with the WOA18 data, we also averaged the
616
conductivity, temperature, and depth (CTD) observations that were located less than 1 degree
617
from each mooring and calculated the independent scaling factor. The CTD observations were
618
collected from ship cruises during six research programs (Program 1, 4, 5, 6, 7 and 8 as listed in
619
Table 1) and their locations are shown in Fig. A1. Their comparison at one selected mooring
620
(marked with the green triangle in Fig. 1) shows that the WOA18 result agrees well with that
621
estimated from ship CTD observations between 200 and 2500 m (Fig. A2a). The difference
622
between the two estimates of scaling factor is relatively large in the upper 200 m and in the
623
bottom ocean (>2500 m), which is about 25% and 40% of the climatological estimate at the
624
surface and bottom level, respectively. Note that CTD observations are snapshots at particular
625
times, therefore, they are different from the annual climatology.
626
627
Fig. A1. Locations of ship CTD observations (asterisks) used for comparison of WKB
628
scaling factor. CTD observations used in Fig. A2a at one selected mooring (marked with the
629
green triangle in Fig. 1) are marked with the green asterisks. Contours indicate 1000, 2000, and
630
3000 m isobaths.
631
30
632
633
Fig. A2. (a) Square root of normalized buoyancy frequency obtained from WOA18 (red) and
634
ship CTD observations (blue) at one mooring (marked with the green triangle in Fig. 1) in the
635
eastern GoM. (b) Mean (blue line) and standard deviation (STD, blue shading) of differences of
636
the square root of normalized buoyancy frequency obtained from WOA18 and ship CTD
637
observations (locations are indicated in Fig. A1).
638
639
Additionally, the mean and standard deviation of the differences in scaling factor were
640
calculated over moorings with available ship CTD observations (Fig. A2b). The mean difference
641
is smaller than 0 in the upper 200 m, the absolute value of which is smaller than 20% of the
642
scaling factor estimated from WOA18. The standard deviation of the differences is mostly
643
smaller than 0.4 in the upper 500 m and smaller than 0.23 below 500 m, which is mostly less
644
than 25% of scaling factors estimated from WOA18. In summary, the WOA18 data and ship
645
CTD data yield similar estimates of WKB scaling factor in our study region. Since the ship CTD
646
observations were obtained only at limited locations and in many cases were not available in the
647
deep ocean, we used the WOA18 data in this study.
648
649
APPENDIX B
650
Factors influencing NIO Amplitude in the Slab Model
651
31
To examine factors influencing NIO amplitude in the slab mixed-layer model, we simulated
652
the wind-driven NIOs across the whole GoM using the wind data at all grid points for 1997 to
653
2013. Note that the seasonal climatology of MLD with a horizontal resolution of 0.25° obtained
654
from WOA18 data was used in the slab model. The wind-driven NIOs have larger amplitude in
655
the Bay of Campeche and the middle eastern GoM in winter than in summer, but they have
656
larger amplitude in the northwestern and middle GoM in summer than in winter (Fig. B1a-b).
657
NIOs are generally more energetic on the slope in the western GoM than those in the open and
658
deep ocean.
659
660
32
Fig. B1. Simulated NIO amplitude (cm/s) with the slab model in water deeper than 800 m
661
obtained from 1997 to 2013 in summer (a) and winter (b) with 0.25° seasonal climatology of
662
mixed layer depth (MLD) used in the slab model. (c-d) Same as (a-b) but for NIO amplitude
663
obtained with 1° seasonal climatology of MLD in the slab model. (e-f) Same as (a-b) but for NIO
664
amplitude obtained with MLD fixed at 25 m in the slab model. Buoy 42003 is marked with the
665
magenta dot in (a). Contours in (a-d) indicate MLD (m) in summer and winter, respectively.
666
667
It is noted that the slab model underestimates the NIO amplitude by more than 50%
668
compared to the observations (Fig. 9). Since MLD is a key parameter in the slab model,
669
uncertainties of simulated NIOs could arise from the changing MLD (Plueddemann and Farrar
670
2006). If smaller MLD were used, the amplitude of NIOs would be larger than that shown in Fig.
671
9c-d and would be closer to the observation (Fig. 9a-b). To test the sensitivity of NIO amplitude
672
to the seasonal variability of MLD, low-resolution (1°) climatology of MLD and a constant MLD
673
value of 25 m were also used in the slab model, respectively. When the low-resolution
674
climatology of MLD was used, the spatial pattern of simulated NIO amplitude is smoother than
675
that obtained with the 0.25° climatology of MLD (Fig. B1c-d). When a constant MLD value was
676
used in the slab model, NIOs are more intense in winter than in summer over most of GoM due
677
to stronger surface winds (Fig. B1e-f) but show some differences in the spatial patterns of NIO
678
amplitude compared to those when the seasonal MLD was considered (Fig. B1a-d). Therefore,
679
the seasonal variability of both surface winds and MLD accounts for the seasonal and spatial
680
pattern of NIOs.
681
33
682
Fig. B2. Simulated NIO amplitude (cm/s) with the slab mixed-layer model in water deeper
683
than 800 m with MLD fixed at 25 m during the extreme weather conditions (wind stress
684
magnitude > 95% threshold) from 1997 to 2013 in summer (a) and winter (b). (c-d) Same as (a-b)
685
but for the normal weather conditions (wind stress magnitude < 95% threshold).
686
We also examined the seasonality of NIO amplitude during extreme weather conditions
687
defined as times when wind stress magnitude is larger than 95% threshold (Fig. B2). Wind-
688
induced NIOs during extreme weather events in summer have larger amplitude in the
689
northeastern and southwestern GoM, which is different from the spatial pattern of NIO amplitude
690
during normal weather conditions when wind stress magnitude is smaller than 95% threshold
691
(Fig. B2a and B2c). However, the amplitude of wind-induced NIOs during extreme weather
692
conditions in winter has a similar spatial pattern as that during the normal weather conditions
693
(Fig. B2b and B2d).
694
695
34
696
Fig. B3. Simulated NIO amplitude (cm/s) with the slab model driven by the CCMP wind
697
analysis product (blue), the 1-hour buoy winds (red), and the 6-hour buoy winds (green) at the
698
buoy location (marked with the magenta dot in Fig. B1a) in January 2010.
699
700
The simulated NIOs could also be affected by finite sampling intervals of the wind product
701
(Hsu 2018; Jiang et al. 2005; Niwa and Hibiya 1999; Rimac et al. 2013). To determine the
702
influence of temporal resolution of wind product, the hourly winds in January 2010 obtained
703
from the National Data Buoy Center (NDBC) buoy 42003 located in the eastern GoM (marked
704
with the magenta dot in Fig. B1a) were used in the slab model. It should be noted that surface
705
winds were observed at 5 m above the sea surface and were adjusted to 10 m for wind stress
706
calculation with the power-law method (Hsu et al. 1994). Figure B3 shows the low-frequency
707
variability of NIO amplitude derived from the CCMP analysis product and the buoy observations
708
at the buoy point in January 2010. The CCMP product captures the major variability of winds at
709
that point and gives rise to NIO variability consistent with that obtained with the buoy wind
710
observations. However, compared to the CCMP product, the buoy observed winds with 6-hour
711
and 1-hour intervals increase NIO amplitude by 18% and 36%, respectively. The underestimated
712
NIO amplitude in the slab model could thus be related to the relatively low resolution and
713
uncertainties in wind analysis product and uncertainties of the MLD estimate.
714
In addition, note that mooring observations sometimes do not reach the sea surface, and
715
might not fully resolve the MLD especially in regions where the MLD is shallow, which may
716
also cause differences between observed and simulated NIOs in the mixed layer.
717
718
35
REFERENCES
719
Alford, M. H., 2001: Internal swell generation: The spatial distribution of energy flux from the
720
wind to mixed layer near-inertial motions. Journal of Physical Oceanography, 31, 2359-2368,
721
https://doi.org/10.1175/1520-0485(2001)031<2359:isgtsd>2.0.co;2
722
——, 2003: Redistribution of energy available for ocean mixing by long-range propagation of
723
internal waves. Nature, 423, 159-162, https://doi.org/10.1038/nature01628
724
Alford, M. H., and M. Whitmont, 2007: Seasonal and spatial variability of near-inertial kinetic
725
energy from historical moored velocity records. Journal of Physical
726
Oceanography, 37, 2022-2037, https://doi.org/10.1175/jpo3106.1
727
Alford, M. H., M. F. Cronin, and J. M. Klymak, 2012: Annual cycle and depth penetration of
728
wind-generated near-inertial internal waves at ocean station papa in the northeast
729
Pacific. Journal of Physical Oceanography, 42, 889-909, https://doi.org/10.1175/jpo-d-11-
730
092.1
731
Alford, M. H., J. A. MacKinnon, H. L. Simmons, and J. D. Nash, 2016: Near-inertial internal
732
gravity waves in the ocean. Annual Review of Marine Science, 8, 95-123,
733
https://doi.org/10.1146/annurev-marine-010814-015746
734
Atlas, R., R. N. Hoffman, J. Ardizzone, S. M. Leidner, J. C. Jusem, D. K. Smith, and D. Gombos,
735
2011: A cross-calibrated, multiplatform ocean surface wind velocity product for
736
meteorological and oceanographic applications. Bulletin of the American Meteorological
737
Society, 92, 157-+, https://doi.org/10.1175/2010bams2946.1
738
Brearley, J. A., K. L. Sheen, A. C. N. Garabato, D. A. Smeed, and S. Waterman, 2013: Eddy-
739
induced modulation of turbulent dissipation over rough topography in the Southern
740
Ocean. Journal of Physical Oceanography, 43, 2288-2308, https://doi.org/10.1175/jpo-d-12-
741
0222.1
742
Brooks, D. A., 1983: The wake of Hurricane Allen in the western Gulf of Mexico. Journal of
743
Physical Oceanography, 13, 117-129, https://doi.org/10.1175/1520-
744
0485(1983)013<0117:twohai>2.0.co;2
745
36
Chen, C. S., R. O. Reid, and W. D. Nowlin, 1996: Near-inertial oscillations over the Texas
746
Louisiana shelf. Journal of Geophysical Research-Oceans, 101, 3509-3524,
747
https://doi.org/10.1029/95jc03395
748
Codiga, D.L., 2011: Unified tidal analysis and prediction using the UTide Matlab functions.
749
Technical Report 2011-01, 59pp, ftp://www.po.gso.uri.edu/
750
pub/downloads/codiga/pubs/2011Codiga-UTide-Report.pdf
751
Cox, J., C. Coomes, S. DiMarco, K. Donohue, G. Z. Forristall, P. Hamilton, R. R. Leben, and
752
D. R. Watts, 2010: Study of deepwater currents in the eastern Gulf of Mexico. OCS Study
753
BOEMRE 2010-041, 473 pp.
754
Cuypers, Y., X. Le Vaillant, P. Bouruet-Aubertot, J. Vialard, and M. J. McPhaden, 2013:
755
Tropical storm-induced near-inertial internal waves during the Cirene experiment: Energy
756
fluxes and impact on vertical mixing. Journal of Geophysical Research-Oceans, 118, 358-
757
380, https://doi.org/10.1029/2012jc007881
758
Cyriac, A., H.E. Phillips, N.L. Bindoff, H. Mao, and M. Feng, 2021: Observational Estimates of
759
Turbulent Mixing in the Southeast Indian Ocean, Journal of Physical Oceanography, 51,
760
2103-2128, https://doi.org/10.1175/JPO-D-20-0036.1
761
D’Asaro, E. A., 1985: The energy flux from the wind to near-inertial motions in the surface
762
mixed layer. Journal of Physical Oceanography, 15, 1043-1059,
763
https://doi.org/10.1175/1520-0485(1985)015<1043:tefftw>2.0.co;2
764
de Boyer Montégut, C., G. Madec, A. S. Fischer, A. Lazar, and D. Iudicone, 2004: Mixed layer
765
depth over the global ocean: An examination of profile data and a profile-based
766
climatology. Journal of Geophysical Research-Oceans, 109,
767
https://doi.org/10.1029/2004jc002378
768
DiMarco, S., 2014: Ocean current, temperature and salinity time series, G04 moorings,
769
DeepWater Horizon oil spill region, July 2012 to July 2013. Gulf of Mexico Research
770
Initiative Information and Data Cooperative (GRIIDC), Harte Research Institute, Texas
771
A&M University–Corpus Christi, https://doi.org/10.7266/N7930R44
772
37
DiMarco, S. F., M. K. Howard, and R. O. Reid, 2000: Seasonal variation of wind-driven diurnal
773
current cycling on the Texas-Louisiana Continental Shelf. Geophysical Research
774
Letters, 27, 1017-1020, https://doi.org/10.1029/1999gl010491
775
Donohue, K., P. Hamilton, K. Leaman, R. Leben, M. Prater, D. R. Watts, and E. Waddell, 2006:
776
Exploratory study of deepwater currents in the Gulf of Mexico. OCS Study MMS 2006-074,
777
430 pp.
778
Donohue, K., P. Hamilton, R. Leben, R. Watts, and E. Waddell, 2008: Survey of deepwater
779
currents in the northwestern Gulf of Mexico. OCS Study MMS 2008-031. 375 pp.
780
Donohue, K. A., D. R. Watts, P. Hamilton, R. Leben, and M. Kennelly, 2016: Loop Current
781
Eddy formation and baroclinic instability. Dynamics of Atmospheres and Oceans, 76, 195-
782
216, https://doi.org/10.1016/j.dynatmoce.2016.01.004
783
Fu, L.L., 1981: Observations and models of inertial waves in the deep ocean, Rev.
784
Geophys., 19, 141–170, https://doi.org/10.1029/RG019i001p00141.
785
Furuichi, N., T. Hibiya, and Y. Niwa, 2008: Model-predicted distribution of wind-induced
786
internal wave energy in the world's oceans. Journal of Geophysical Research-Oceans, 113,
787
https://doi.org/10.1029/2008jc004768
788
Gough, M. K., A. Reniers, J. H. MacMahan, and S. D. Howden, 2016: Resonant near-surface
789
inertial oscillations in the northeastern Gulf of Mexico. Journal of Geophysical Research-
790
Oceans, 121, 2163-2182, https://doi.org/10.1002/2015jc011372
791
Hamilton, P., 2009: Topographic Rossby waves in the Gulf of Mexico. Progress in
792
Oceanography, 82, 1-31, https://doi.org/10.1016/j.pocean.2009.04.019
793
Hamilton, P., T. J. Berger, J. J. Singer, E. Waddell, J. H. Churchill, R. R. Leben, T. N. Lee, and
794
W. Sturges, 2000: Desoto Canyon eddy intrusion study. OCS Study MMS 2000-080, 275 pp.
795
Hamilton, P., K. Donohue, C. Hall, R. R. Leben, H. Quian, J. Sheinbaum, and D. R. Watts, 2014:
796
Observations and dynamics of the Loop Current. OCS Study BOEM 2015-006. 417 pp.
797
Hamilton, P., J. J. Singer, E. Waddell, and K. Donohue, 2003: Deepwater observations in the
798
northern Gulf of Mexico from in-situ current meters and PIES. OCS Study MMS 2003-049.
799
95 pp.
800
38
Hendershott, M. C., 1973: Inertial oscillations of tidal period, Prog. Oceanogr., 6, 1– 27.
801
Hsu, C-Y., 2018: Wind-driven near-inertial waves and their impact in the Gulf of Mexico. Texas
802
A & M University, 90 pp, http://hdl.handle.net/1969.1/173605.
803
Hsu, S. A., E. A. Meindl, and D. B. Gilhousen, 1994: Determining the power-law wind-profile
804
exponent under near-neutral stability conditions at sea. Journal of Applied
805
Meteorology, 33, 757-765, https://doi.org/10.1175/1520-
806
0450(1994)033<0757:dtplwp>2.0.co;2
807
Jaimes, B., and L. K. Shay, 2009: Mixed layer cooling in mesoscale oceanic eddies during
808
Hurricanes Katrina and Rita. Monthly Weather Review, 137, 4188-4207,
809
https://doi.org/10.1175/2009mwr2849.1
810
——, 2010: Near-inertial wave wake of Hurricanes Katrina and Rita over mesoscale oceanic
811
eddies. Journal of Physical Oceanography, 40, 1320-1337,
812
https://doi.org/10.1175/2010jpo4309.1
813
Jarosz, E., Z. R. Hallock, and W. J. Teague, 2007: Near-inertial currents in the DeSoto Canyon
814
region. Continental Shelf Research, 27, 2407-2426, https://doi.org/10.1016/j.csr.2007.06.014
815
Jiang, J., Y. Y. Lu, and W. Perrie, 2005: Estimating the energy flux from the wind to ocean
816
inertial motions: The sensitivity to surface wind fields. Geophysical Research Letters, 32,
817
https://doi.org/10.1029/2005gl023289
818
Jing, Z., P. Chang, S. F. DiMarco, and L. X. Wu, 2015: Role of Near-Inertial Internal Waves in
819
Subthermocline Diapycnal Mixing in the Northern Gulf of Mexico. Journal of Physical
820
Oceanography, 45, 3137-3154, https://doi.org/10.1175/jpo-d-14-0227.1
821
——, 2018: Observed Energy Exchange between Low-Frequency Flows and Internal Waves in
822
the Gulf of Mexico. Journal of Physical Oceanography, 48, 995-1008,
823
https://doi.org/10.1175/jpo-d-17-0263.1
824
Jochum, M., and Coauthors, 2013: The Impact of Oceanic Near-Inertial Waves on
825
Climate. Journal of Climate, 26, 2833-2844, https://doi.org/10.1175/jcli-d-12-00181.1
826
39
Jordi, A., D. P. Wang, and P. Hamilton, 2016: Near-inertial motions in the DeSoto Canyon
827
during Hurricane Georges. Ocean Modelling, 105, 34-43,
828
https://doi.org/10.1016/j.ocemod.2016.07.005
829
Kunze, E., 1985: Near-inertial wave propagation in geostrophic shear. Journal of Physical
830
Oceanography, 15, 544-565, https://doi.org/10.1175/1520-
831
0485(1985)015<0544:niwpig>2.0.co;2
832
——, 1986: The mean and near-inertial velocity fields in a warm core ring. Journal of Physical
833
Oceanography, 16, 1444-1461, https://doi.org/10.1175/1520-
834
0485(1986)016<1444:tmaniv>2.0.co;2
835
——, 1995: The energy balance in a warm-core ring’s near-inertial critical layer. Journal of
836
Physical Oceanography, 25, 942-957, https://doi.org/10.1175/1520-
837
0485(1995)025<0942:tebiaw>2.0.co;2
838
Large, W. G., & S., Yeager, 2004: Diurnal to decadal global forcing for ocean and sea-ice
839
models: The data sets and flux climatologies (No. NCAR/TN-460+STR). University
840
Corporation for Atmospheric Research. doi:10.5065/D6KK98Q6
841
Leaman, K. D., and T. B. Sanford, 1975: Vertical energy propagation of inertial waves: A vector
842
spectral analysis of velocity profiles. Journal of Geophysical Research, 80, 1975-1978,
843
https://doi.org/10.1029/JC080i015p01975
844
Liang, X., and A. M. Thurnherr, 2012: Eddy-modulated internal waves and mixing on a
845
midocean ridge. Journal of Physical Oceanography, 42, 1242-1248,
846
https://doi.org/10.1175/jpo-d-11-0126.1
847
Liang, X., and C. Wunsch, 2015: Note on the redistribution and dissipation of tidal energy over
848
mid-ocean ridges. Tellus A, 67, 27385, https://doi.org/10.3402/tellusa.v67.27385
849
Lilly, J. M., 2019: jLab: A data analysis package for Matlab, v.1.6.6,
850
http://www.jmlilly.net/jmlsoft.html.
851
Liu, Y., R. H., Weisberg, S., Vignudelli, and G. T., Mitchum, 2016: Patterns of the loop current
852
system and regions of sea surface height variability in the eastern Gulf of Mexico revealed by
853
40
the self-organizing maps, J. Geophys. Res. Oceans, 121, 2347– 2366,
854
doi:10.1002/2015JC011493.
855
Locarnini, R. A., and Coauthors, 2018: World Ocean Atlas 2018, Volume 1: Temperature. A.
856
Mishonov Technical Ed.; NOAA Atlas NESDIS 81, 52 pp.
857
MacKinnon, J. A., and K. B. Winters, 2005: Subtropical catastrophe: Significant loss of low-
858
mode tidal energy at 28.9 degrees, Geophysical Research Letters, 32,
859
https://doi.org/10.1029/2005gl023376
860
Mariano, A. J., and Coauthors, 2016: Statistical properties of the surface velocity field in the
861
northern Gulf of Mexico sampled by GLAD drifters. Journal of Geophysical Research-
862
Oceans, 121, 5193-5216, https://doi.org/10.1002/2015jc011569
863
Martens, C. S., H. P. Mendlovitz, H. Seim, L. Lapham, and M. D'Emidio, 2016: Sustained in situ
864
measurements of dissolved oxygen, methane and water transport processes in the benthic
865
boundary layer at MC118, northern Gulf of Mexico. Deep-Sea Research Part II-Topical
866
Studies in Oceanography, 129, 41-52, https://doi.org/10.1016/j.dsr2.2015.11.012
867
McKone, K., N. D. Walker, and E. Weeks, 2007: Full-water column currents near the Sigsbee
868
Escarpment (91-92ºW. Longitude) and relationships with the Loop Current and associated
869
warm and cold-core eddies. OCS Study MMS 2007-056, 107 pp.
870
Mooers, C. N. K., 1973: A technique for the cross spectrum analysis of pairs of complex-valued
871
time series, with emphasis on properties of polarized components and rotational invariants,
872
Deep Sea Res. Oceanogr. Abstr., 20, 1129–1141.
873
Munk, W., and C. Wunsch, 1998: Abyssal recipes II: energetics of tidal and wind mixing. Deep-
874
Sea Research Part I-Oceanographic Research Papers, 45, 1977-2010.
875
https://doi.org/10.1016/s0967-0637(98)00070-3
876
Nikurashin, M., and R. Ferrari, 2010a: Radiation and dissipation of internal waves generated by
877
geostrophic motions impinging on small-scale topography: Application to the Southern
878
Ocean. Journal of Physical Oceanography, 40, 2025-2042,
879
https://doi.org/10.1175/2010jpo4315.1
880
41
——, 2010b: Radiation and dissipation of internal waves generated by geostrophic motions
881
impinging on small-scale topography: Theory. Journal of Physical Oceanography, 40, 1055-
882
1074, https://doi.org/10.1175/2009jpo4199.1
883
Niwa, Y., and T. Hibiya, 1999: Response of the deep ocean internal wave field to traveling
884
midlatitude storms as observed in long-term current measurements (vol 104, pg 10981,
885
1999). Journal of Geophysical Research-Oceans, 104, 20857-20857,
886
https://doi.org/10.1029/1999jc900211
887
Pallàs-Sanz, E., J. Candela, J. Sheinbaum, and J. Ochoa, 2016: Mooring observations of the near-
888
inertial wave wake of Hurricane Ida (2009). Dynamics of Atmospheres and Oceans, 76, 325-
889
344, https://doi.org/10.1016/j.dynatmoce.2016.05.003
890
Panagiotis, V., 2014: Gulf of Mexico high-resolution (0.01°×0.01°) bathymetric grid - version
891
2.0, February 2013. Gulf of Mexico Research Initiative Information and Data Cooperative
892
(GRIIDC), Harte Research Institute, Texas A&M University Corpus Christi,
893
https://doi.org/10.7266/N7X63JZ5
894
Plueddemann, A. J., and J. T. Farrar, 2006: Observations and models of the energy flux from the
895
wind to mixed-layer inertial currents. Deep-Sea Research Part II-Topical Studies in
896
Oceanography, 53, 5-30, https://doi.org/10.1016/j.dsr2.2005.10.017
897
Pollard, R. T., and R. C. Millard, 1970: Comparison between observed and simulated wind-
898
generated inertial oscillations. Deep-Sea Research, 17, 813, https://doi.org/10.1016/0011-
899
7471(70)90043-4
900
Polzin, K. L., B., Wang, Z., Wang, F., Thwaites, and A. J. Williams, III., 2021: Moored Flux and
901
Dissipation Estimates from the Northern Deepwater Gulf of Mexico. Fluids, 6(7), 237.
902
https://doi.org/10.3390/fluids6070237
903
Pujol, M.-I., Y.F. augre, G. Taburet, S. Dupuy, C. Pelloquin, M. Ablain, and N. Picot, 2016:
904
DUACS DT2014: the new multi-mission altimeter data set reprocessed over 20 years. Ocean
905
Science, 12, 1067–1090. https://doi.org/10.5194/os-12-1067-2016
906
42
Rimac, A., J. S. von Storch, C. Eden, and H. Haak, 2013: The influence of high-resolution wind
907
stress field on the power input to near-inertial motions in the ocean. Geophysical Research
908
Letters, 40, 4882-4886, https://doi.org/10.1002/grl.50929
909
Shay, L. K., and R. L. Elsberry, 1987: Near-inertial ocean current response to Hurricane
910
Frederic. Journal of Physical Oceanography, 17, 1249-1269, https://doi.org/10.1175/1520-
911
0485(1987)017<1249:niocrt>2.0.co;2
912
Shay, L. K., and E. W. Uhlhorn, 2008: Loop current response to Hurricanes Isidore and
913
Lili. Monthly Weather Review, 136, 3248-3274, https://doi.org/10.1175/2007mwr2169.1
914
Shay, L. K., R. L. Elsberry, and P. G. Black, 1989: Vertical structure of the ocean current
915
response to a hurricane. Journal of Physical Oceanography, 19, 649-669,
916
https://doi.org/10.1175/1520-0485(1989)019<0649:vsotoc>2.0.co;2
917
Shay, L. K., S. W. Chang, and R. L. Elsberry, 1990: Free surface effects on the near-inertial
918
ocean current response to a hurricane. Journal of Physical Oceanography, 20, 1405-1424,
919
https://doi.org/10.1175/1520-0485(1990)020<1405:fseotn>2.0.co;2
920
Shay, L. K., A. J. Mariano, S. D. Jacob, and E. H. Ryan, 1998: Mean and near-inertial ocean
921
current response to Hurricane Gilbert. Journal of Physical Oceanography, 28, 858-889,
922
https://doi.org/10.1175/1520-0485(1998)028<0858:manioc>2.0.co;2
923
Shay, L. K., P. G. Black, A. J. Mariano, J. D. Hawkins, and R. L. Elsberry, 1992: Upper ocean
924
response to Hurricane Gilbert. Journal of Geophysical Research-Oceans, 97, 20227-20248,
925
https://doi.org/10.1029/92jc01586
926
Sheen, K. L., and Coauthors, 2013: Rates and mechanisms of turbulent dissipation and mixing in
927
the Southern Ocean: Results from the Diapycnal and Isopycnal Mixing Experiment in the
928
Southern Ocean (DIMES). Journal of Geophysical Research-Oceans, 118, 2774-2792,
929
https://doi.org/10.1002/jgrc.20217
930
Sheinbaum, J., A. Badan, J. Ochoa, J. Candela, D. Rivas, and J. I., González, 2007: Full water
931
column current observations in the central Gulf of Mexico. OCS Study MMS 2007-022. 58
932
pp.
933
43
Spencer, L. J., S. F. DiMarco, Z. K. Wang, J. J. Kuehl, and D. A. Brooks, 2016: Asymmetric
934
oceanic response to a hurricane: Deep water observations during Hurricane Isaac. Journal of
935
Geophysical Research-Oceans, 121, 7619-7649, https://doi.org/10.1002/2015jc011560
936
Sun, H., Q. X. Yang, W. Zhao, X. F. Liang, and J. W. Tian, 2016: Temporal variability of
937
diapycnal mixing in the northern South China Sea. Journal of Geophysical Research-
938
Oceans, 121, 8840-8848, https://doi.org/10.1002/2016jc012044
939
Tenreiro, M., J. Candela, E. P. Sanz, J. Sheinbaum, and J. Ochoa, 2018: Near-surface and deep
940
circulation coupling in the western Gulf of Mexico. Journal of Physical Oceanography,
941
48(1), 145–161, https://doi.org/10.1175/JPO-D-17-0018.1
942
van Aken, H. M., L. R. M. Maas, and H. van Haren, 2005: Observations of inertial wave events
943
near the continental slope off Goban Spur. Journal of Physical Oceanography, 35, 1329-
944
1340, https://doi.org/10.1175/jpo2769.1
945
Weisberg, R. H., & Y., Liu, 2017: On the Loop Current penetration into the Gulf of
946
Mexico. Journal of Geophysical Research: Oceans, 122, 9679–
947
9694. https://doi.org/10.1002/2017JC013330
948
Welsh, S. E., M. Inoue, L. J. Jr. Rouse, and E. Weeks, 2009: Observation of the deepwater
949
manifestation of the Loop Current and Loop Current rings in the Eastern Gulf of Mexico.
950
OCS Study MMS 2009-050, 110 pp.
951
Yang, Y., R. H. Weisberg, Y. Liu, and X. San Liang, 2020: Instabilities and Multiscale
952
Interactions Underlying the Loop Current Eddy Shedding in the Gulf of Mexico. J. Phys.
953
Oceanogr., 50, 1289–1317, https://doi.org/10.1175/JPO-D-19-0202.1.
954
Zweng, M. M., and Coauthors, 2018: World Ocean Atlas 2018, Volume 2: Salinity. A. Mishonov
955
Technical Ed.; NOAA Atlas NESDIS 82, 50 pp.
956
Zhai, X., R. J. Greatbatch, and C. Eden, 2007: Spreading of near-inertial energy in a 1/12 degrees
957
model of the North Atlantic Ocean. Geophysical Research Letters, 34,
958
https://doi.org/10.1029/2007gl029895
959
44
Zhai, X., R. J. Greatbatch, C. Eden, and T. Hibiya, 2009: On the loss of wind-induced near-
960
inertial energy to turbulent mixing in the upper ocean. Journal of Physical Oceanography, 39,
961
3040-3045, https://doi.org/10.1175/2009jpo4259.1
962
Zhang, X., D. C. Smith, S. F. DiMarco, and R. D. Hetland, 2010: A Numerical Study of Sea-
963
Breeze-Driven Ocean Poincare Wave Propagation and Mixing near the Critical Latitude. J.
964
Phys. Oceanogr., 40, 48–66, https://doi.org/10.1175/2009JPO4216.1.
965
Zhu, Y., 2020a: Daily-mean satellite altimetry measurements from the Gulf of Mexico region,
966
1993-01-01 to 2017-05-17. Gulf of Mexico Research Initiative Information and Data
967
Cooperative (GRIIDC), Harte Research Institute, Texas A&M University Corpus Christi,
968
https://doi.org/10.7266/n7-ya5j-0t90
969
Zhu, Y., 2020b: Selected historical mooring data in the Gulf of Mexico from 1997-03-20 to
970
2014-06-27. Gulf of Mexico Research Initiative Information and Data Cooperative
971
(GRIIDC), Harte Research Institute, Texas A&M University Corpus Christi,
972
https://doi.org/10.7266/n7-7g3d-9q79
973
Zhu, Y., and X. Liang, 2020: Coupling of the surface and near-bottom currents in the Gulf of
974
Mexico. Journal of Geophysical Research-Oceans, 125,
975
https://doi.org/10.1029/2020jc016488
976