OTC 20826

The Gulf of Mexico Current Simulation - A 32-Year Run of 1/12o HYCOM

W. Wang and O. E. Esenkov, ExxonMobil Upstream Research Company

Copyright 2010, Offshore Technology Conference

This paper was prepared for presentation at the 2010 Offshore Technology Conference held in Houston, Texas, USA, 3–6 May 2010.

This paper was selected for presentation by an OTC program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Offshore Technology Conference and are subject to correction by the author(s). The material does not necessarily reflect any position of the Offshore Technology Conference, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Offshore Technology Conference is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of OTC copyright.

Abstract

Drilling and production operations in the deepwater regions of the Gulf of Mexico can be impacted by strong currents from
the Loop Current and its associated Eddies (LCE). The position of the Loop Current is variable as is eddy propagation across
the Gulf. Consequently the specification of design currents for locations in the deepwater Gulf requires extensive knowledge
of the LCE system.

An improved 1/12o HYbrid Coordinate Ocean Model (HYCOM) of the Gulf of Mexico was run for a 32-year period to
simulate ocean currents in the Gulf by implementing a "Hybrid Boundary Condition" that combines original and high-inflow
open boundary conditions. The improved model results show a realistic Loop Current penetration, eddy path variability and
other eddy statistics. Eddy current extremes compare well with the best existing database in the U. S. lease area. The
currents from the 32-year model run broaden our knowledge base of the LCE system and provide a database of current
conditions for design of structures in the Gulf of Mexico. Bottom currents near the Sigsbee Escarpment were also examined.
The results show that characteristics of these bottom current events match those of observed currents, but the magnitudes of
the modeled current events are lower than the observations which is most likely due to a lower vertical model resolution near
the bottom.

1. Introduction

In the deepwater region of the Gulf of Mexico, LCE is a dominant feature which produces high currents that can adversely
impact exploration and production activities. At times, the Loop Current flows almost directly from the Yucatan Strait to the
Florida Strait. Over time the Loop Current grows and extends further into the Gulf, and then instabilities in the Loop lead to
a "pinching off" of eddy as shown in Figure 1. The detached eddy will slowly propagate west until it finally dissipates. The
process may take over a year.

The high currents associated with the LCE can be as high as 2-3 m/s at the surface and they can penetrate into the water
column to several hundreds meters below the water surface. LCE currents can be greater than hurricane generated currents as
shown in Figure 2 which presents current vertical profiles generated by LCE, winter storms, and hurricanes (API, 2009). The
high currents increase loads significantly on offshore structures, including floating hulls, mooring lines, and risers.
2 OTC 20826

Figure 1. Contours of Sea Surface Height in the Gulf of Mexico to show the Loop Current and the formation of its eddy.

0
Depth Below Surface (m)

-200

-400

-600
Loop Eddy
-800 Hurricane
Winter Storm
-1000
0 0.5 1 1.5 2 2.5 3
Current Speed (m/s)
Figure 2. Vertical current profiles for LCE, hurricane, and winter storm recommended by API.

The positions as well as strengths of the Loop and its eddies vary, therefore extensive knowledge of the Loop/eddy system is
needed to specify design currents for the deepwater Gulf. In the past several years, ExxonMobil Upstream Research
Company conducted a current research program for the Gulf of Mexico aimed at developing its current modeling capability,
so that model results can be used for improving current criteria.
OTC 20826 3

HYCOM (Wallcraft, 2003) combines the advantages of different types of coordinates used in several current models. Its
hybrid coordinates, as shown in Figure 3, include z-level coordinates in the mixed layer or unstratified sea, isopycnal
(ρ) coordinates in the open, stratified ocean, then reverting smoothly to terrain-following (σ) coordinates in shallow coastal
regions. The use of these hybrid coordinates extends the geographic range of model applicability of the HYCOM model.
We chose HYCOM for our research because of its flexible vertical coordinate nature, accurate formulation of physical
processes, and the promising results obtained in previous applications in other areas over the world.

Figure 3. Schematic showing HYCOM hybrid vertical coordinate .

Our validation of the original 1/12o HYCOM model results indicated that the model generally performed well except in terms
of the Loop Current penetration into the Gulf and the current speed, i.e., the Loop Current and its eddies did not penetrate far
enough north into the Gulf and maximum current speed was underpredicted. Through a systematic sensitivity analysis we
identified a mechanism that controls the Loop Current growth and eddy separation. We then implemented a "Hybrid
Boundary Condition" to simulate this mechanism that significantly improved the model performance. The Hybrid Boundary
Condition consisted of a combination of the open boundary conditions of an original base case obtained from the North
Atlantic model run (Srinivasan 2004) and a case with high constant inflow through the Yucatan Strait. The modified
HYCOM with 1/12o spatial resolution was run in-house for a 32 year simulation forced by NOAA WaveWatch3 hindcast
surface winds (Tolman, 2002).

The 32-year model results showed a realistic Loop Current penetration and represented eddy climatology much better than
the original results. Current extremes compared well with the best existing database in the U. S. lease area which was
derived from the Gulf Eddy Model (GEM), a measurement-based parametric model, developed by the Climatology And
Simulation of Eddies (CASE) joint industry project.

The bottom currents near the Sigsbee Escarpment were also examined. The general characteristics of these bottom current
events matched the occurrences of observed currents, but the magnitudes of the current events were lower than the
observations probably due to a lower vertical model resolution near the bottom.

The outline of the paper is as follows. Section 2 discusses the comparison of original HYCOM model results to available
data. Section 3 describes the improvements of the HYCOM performance. Section 4 presents the comparison of improved
model results to available data, and section 5 gives a summary and conclusions.
4 OTC 20826

2. Evaluation of 1/3o and 1/12o Original HYCOM Results

We first evaluated two preliminary HYCOM simulations of the Gulf of Mexico circulation to determine how well the model
can replicate the Loop and eddy environment of the Gulf. The first simulation was a 16-year long 1/3o HYCOM simulation
performed by the HYCOM group (Srinivasan 2004). This simulation contained 193 monthly snapshots archived over the
period from January 1980 to December 1995. The second simulation was a 5-year run of the 1/12o resolution model. The
evaluation presented here focused on the climatology of the LCE. The basis for comparison of the model results was the
measurement-based climatology which was primarily derived from the GEM model, archived LCE charts, and satellite data.
The following aspects of the LCE climate were examined:

1. The percent occurrence of LCE water by location;

2. The generation of eddies and the resulting eddy paths;
3. The translation speed of eddies;
4. The 95% non-exceedance surface-layer current;
5. The extreme surface-layer current.

In order to evaluate the LCE climate, the first step was to identify eddy events from the HYCOM model results. We used
two criteria to determine eddy events and their locations. The first criterion, which is traditionally used, is based on the in-
situ water temperature data, in particular, locations where the 15oC isotherm exceeds depths greater than 200m (Cochrane,
1972; Copper, 1990). The second criterion was a sea surface height (SSH) anomaly derived from satellite altimeter data
which is on the order of 10's of centimeters (Berger, 1996). For the 1/3o model results, it was found, through a trial-and-error
process, that if a dual criterion of (1) the 19oC isotherm crossing a depth of 200 m and (2) the sea surface height anomaly of
10 cm, were used, the LCE positions could be identified quite reliably. Since the 19oC isotherm criterion was used rather
than the 15oC isotherm, this indicated a problem with the temperature structure of the 1/3o model results, i.e., the model
temperatures were too warm in this depth range. For the 1/12o model results, a 15o C isotherm at the 200m water depth and a
9 cm anomaly were used, significant improvement regarding the temperature structure.

From our results, we learned that 1/3o is not high enough to capture all of the eddy dynamics. Hence it is not surprising that a
number of deficiencies were identified in the 1/3o model results, including (1) elevated temperature of the Gulf water at the
200m depth, (2) inadequate penetration of the LCE north into the Gulf , (3) less scatter in the eddy tracks than the observed
eddies, (4) low LCE current speeds, (5) high LCE drift speeds across the Gulf and, (6) increased eddy shedding frequency -
26 eddies in the 16-year model period versus 20 eddies in the 16-year GEM database

Most of the areas of deficiency noted above were improved when the resolution was increased from 1/3o to 1/12o as indicated
below and shown in Table 1 and Figure 4,

• the warm water bias was less severe,

• the eddy shedding rate was close to the GEM climatology
• the westward drift rate of the eddies was near the GEM climatology
• the peak LCE speeds were significantly higher
• the eddy tracks showed more scatter.

However, the model predictions did not show the LCE penetrating far enough north. In fact, none of the eddies generated in
the 5-year run crossed 27.5oN whereas the GEM results indicated that eddies regularly cross that threshold and occasionally
even cross 28oN.

1/3o HYCOM 1/12o HYCOM GEM Data

Shedding Rate (No./Year) 1.6 1.5 1.4
Drift Rate (n.m./day) 2.2 1.7 1.7
Max. Surface Current Speed (cm/s) 160 210 270
o o
Table 1. Comparison of original HYCOM 1/3 and 1/12 model results to GEM data
OTC 20826 5

Figure 4. Comparison of eddy paths of two original model runs to GEM eddy paths.

3. Efforts to Improve HYCOM Performance

As discussed in the previous section, the comparison of original HYCOM model results to GEM data shows that the original
model is unable to match data for eddy path and current speed. We know that eddy path is an important aspect of eddy
climatology for engineering application because if the LCE positions are not correct then the model results are not of
sufficient quality for defining site specific current conditions. However, the original HYCOM formulation did not focus on
the eddy path. Therefore, we needed to improve the model to ensure that the model results include realistic Loop penetration,
eddy separation, eddy propagation, and current speed.

In 1972 Reid (Reid, 1972) from Texas A&M University presented a simplified model for Loop Current penetration based on
the conservation of potential vorticity. In his model the Loop Current penetration is a function of the inflow angle and inflow
velocity at the Yucatan Strait. A schematic of the model is shown in Figure 5, and the equation for the Loop penetration is
given as follows,
1/ 2
⎧⎛ 2V ⎞ ⎫
b = ⎨⎜⎜ c ⎟⎟(1 − cosθ o )⎬
⎩⎝ β ⎠ ⎭
where b is the Loop Current penetration in meters, θ0 is the inflow angle in degrees counter clockwise from North, Vc is the
velocity in ms-1 , and β (m-1s-1) is the variation of Coriolis parameter with latitude.

Oey (Oey, 2003 and 2004) assumed a fixed inflow angle and derived a similar relation in terms of the relative vorticity and
inflow velocity. Therefore, the dominant parameters governing the LCE flow were expected to be the inflow velocity and
angle. To confirm these findings, we set up a test matrix to examine the parameter space. The schematic of the test matrix is
shown in Figure 6, and the test cases are given in Table 2.
6 OTC 20826

b
Vc θ0

Figure 5. The schematic of Reid model.

Penetration (km)

High Angle
Low Angle
High
Low
Inflow
Inflow
Figure 6. The schematic of nominal test matrix.

Inflow Inflow Volume (V)

Angle
(θ0) High (Sv*) Low (Sv)
Large
(~120o) 35 22
Small
(~60o) 33
6 3
* 1 Sv=10 m /s

Table 2. Test matrix to confirm that the inflow velocity and angle are the dominant parameters for LCE penetration. Please note
since high inflow velocity is typically associated with high inflow volume, the inflow volumes are provided in this table for
calculation convenience.
OTC 20826 7

We found that a large inflow angle was associated with a cold eddy located immediately south of the Yucatan Strait and a
small angle was associated with a warm eddy in the same area. The test results showed that a high inflow with a small inflow
angle case did not generate deep penetration or even eddies, but a high inflow with a large inflow angle case created
excessively deep penetration. Figures 7 and 8 show the differences of the Loop positions between the two extreme cases of
inflow angle (same inflow volume) on the same model day with the same initial condition.

Vc

Figure 7. The SSH map for the High Inflow/Small Angle case on model Day 99. This case did not generate deep penetration.
8 OTC 20826

Vc

Figure 8. The SSH map for the High Inflow/Large Angle case on model Day 99. This case created excessively deep penetration.

Since the observed behavior of historical eddies was between the two identified cases, we proposed a combination of the two
open boundary conditions, called a "hybrid boundary condition", to obtain a solution close to the measurement-based eddy
climatology. To do so we used a linear combination of the base case which was the case with the original boundary
condition and the "high inflow/large angle" boundary conditions, i.e.,

(hybrid B. C.) = α (high inflow/large angle)+ (1-α) (base case)

We tested four cases with the hybrid B. C. by running 5 or 10 year simulations:

Case 1. α=1.00 high inflow/large angle

Case 2. α=0.75 hybrid 1
Case 3. α=0.50 hybrid 2
Case 4. α=0.00 base case

The eddy paths for the four cases are shown in Figure 9. The benchmark used to evaluate the model results is 16-years of
GEM eddy paths which are also included in the figure as green curves. It is seen from the plot that the Case 3 with α=0.5 is
much closer to the GEM eddy paths with realistic penetration and scatter variation. The other three cases had the eddy tracks
located either on the northern side (Case 1 and 2) or on the southern side (Case 4) with less scatter. Comparisons of other
eddy statistics of the Case 3 hybrid boundary condition, including eddy shedding rate, eddy westward drft speed and current
speed, also showed encouraging results as given in Table 3.
OTC 20826 9

Case 3 Model Result GEM Data

Shedding Rate (No./Year) 1.5 1.4
Drift Rate (n.m./day) 1.8 1.7
Max. Surface Current Speed (cm/s) 290 270

Table 3. Comparison of Case 3 (α=0.50) model results to GEM data

Figure 9. Eddy paths for four test cases and GEM 16-year eddy tracks used as a benchmark to evaluate the test cases.

4. Results of the Improved 32-Year Model Run

The open boundary condition used in the production run was test Case 3, i.e., a combination of 50% of the base case and 50%
of the high inflow/large angle case. NOAA's WaveWatch3 (WW3) hindcast winds (Tolman, 2002) were used as wind
forcing. The model array size was 258x175 with 20 vertical layers, and it was run for a 32-year period with daily output
archived. In this section, results of the production run are discussed.

4.1 Eddy Formation Rate

In the model assessment, the 15oC isotherm at a 200 m depth and the sea surface height anomaly of 9 cm were used to define
eddy spatial extent. During the 32-year simulation 49 eddies were generated. The mean annual eddy formation rate was
therefore about 1.5 per year. The period of the GEM database that we used to evaluate the production run is from 1986 to
2005. During this 20-year period there were 28 eddies found, so the mean annual eddy formation rate derived from GEM
10 OTC 20826

was 1.4 per year. Therefore the annual mean rate derived from the model results was comparable with the one estimated
from the GEM database.

4.2 Eddy Path and Translation Speed

Figure 10 shows eddy tracks estimated from both the model output and GEM. It is clear that the model eddy tracks had
realistic penetration distance and path variability. However, we notice from the plot that almost all eddies in HYCOM were
formed west of 88oW, but the GEM eddies originated as far east as 85.5oW. The reason for this was that the developers of
GEM started to track some eddies before they detached from the Loop Current. We started to track eddies only after they
clearly separated from the Loop in the HYCOM data. Eddy translation speeds for the production run and GEM are plotted in
Figure 11. The median translation speed estimated from the model results was about 1.8 n.m./day while from the GEM data
it was about 1.7 n.m./day. The ranges of the translation speeds for both datasets were almost the same, from 0.5 n.m./day to
4.2 n.m./day.

Figure 10. Eddy paths for the production run and for the GEM database.
OTC 20826 11

o o
Figure 11. Histogram of eddy westward translation speed between 90 W and 92 W for production run and GEM. The red line is the
median translation speed.

4.3 Eddy Occurrence

Figure 12 shows our best estimate of measurement-based eddy occurrence in percentage of time from all data sources. The
data used included satellite thermal imagery, sea surface height charts, LCE charts and the GEM database. The model eddy
occurrence shown in Figure 13 was compared with Figure 12. Figure 14 shows the differences between the two occurrence
estimates. The model had lower eddy occurrence near the steep slope region (1000m-2000m water depths) on the eastern
side of the region. This was likely because the movement of the eddies in the model followed the bathymetry more closely
than the measurement-based results. On the southern side, near 26oN, the model showed higher eddy occurrence rates than
the measurements.
12 OTC 20826

Figure 12. Eddy occurrence (percentage of time) estimated from all data sources with our best adjustment including satellite data
and GEM database.

Figure 13. Eddy occurrence (percentage of time) estimated from the HYCOM model production results.
OTC 20826 13

Figure 14. The difference (HYCOM - best estimate) between the two eddy occurrence estimates (percentage of time).

4.4 Surface Current Speeds

The 95% non-exceedence surface current speeds estimated from the improved HYCOM model production run and GEM are
shown in Figures 15 and 16, respectively. The spatial distribution patterns as well as the speeds were similar in the two
datasets; Figure 17 shows the difference. As for the eddy occurrence, the maximum differences in surface current speed
occurred in the steep slope region near the eastern side due to several possible reasons. Since currents change rapidly with
the steep slope, the model resolution may not be fine enough to respond to the rapid changes. On the other hand, the contours
for measurement-based current speeds were estimated in a conservative way for this region. The comparisons for other
extremes, from the 1-year through the 200-year return values, are not included here. However, the comparisons for all of the
return values showed results similar to that for the 95% non-exceedence comparison. Table 4 lists the differences of the
maximum return values within the GEM area. All the differences except the 1-year return value were smaller than 10%,
which indicates that the model results were comparable to the GEM database.

HYCOM GEM Difference (%)

(cm/s) (cm/s) (HYCOM-GEM)
95% nx 147 133 9%
1-yr 212 159 25%
5-yr 217 200 8%
10-yr 228 218 4%
20-yr 237 232 2%
50-yr 249 260 -4%
100-yr 258 269 -4%
200-yr 266 279 -5%

Table 4. The difference between the maximum return values of the surface current speed estimated from HYCOM and GEM for the
GEM area.
14 OTC 20826

Figure 15. 95% non-exceedence surface current speeds (cm/s) for GEM.

Figure 16. 95% non-exceedence surface current speeds (cm/s) for HYCOM.
OTC 20826 15

Figure 17. The difference (HYCOM-GEM) of 95% non-exceedence surface current speeds (cm/s) between HYCOM and GEM.

4.5 Surface Currents for the Whole Gulf of Mexico

The discussion in the previous section indicates that the model results show realistic eddy characteristics and surface current
extremes. In Figure 18 the eddy occurrence is presented for the whole Gulf for water depths greater than 200m. Exceedence
curves of surface current speeds derived from the production run were compared with the GEM results. An example of the
comparison at a representative deep water site is presented in Figure 19. The comparison indicates that the shape of the
exceedence distribution curve is similar to the GEM result. Therefore, the model current speeds can be calibrated in a
straightforward way by scaling the speeds, but the scaling factor may vary by region.

Since the model results are the only existing data source for operational currents, including eddies, in the whole Gulf, they
are very valuable for providing design conditions for those areas, such as Mexican waters. Another advantage of the model
results is that they provide a continuous time series of the currents, which in itself is useful for operational criteria or
structural fatigue assessments.
16 OTC 20826

Figure 18. The occurrence of the Loop Current and its eddies for the whole Gulf of Mexico for water depth greater than 200m
estimated from the 32-year model output.

Figure 19. Comparison of the surface current exceedence values derived from the 32-year model output (blue dots) and the GEM
o o
result (red dots) for a deep water site, 89 W, 26 N.

4.6 Bottom Currents in the Sigsbee Escarpment Area

Numerous studies (Hamilton, 1990, 2001, 2007; McKone, 2007) have showed the presence of high currents (~ 100cm/s) near
the bottom along the Sigsbee Escarpment likely caused by bottom topographic waves (BTW). The high bottom currents have
a direct impact on offshore operations for the oil and gas industry, such as pipeline design and installation as well as piling
operations.
OTC 20826 17

We examined the statistics of the bottom current events caused by the bottom topographic waves in the Sigsbee Escarpment
area from the 32-year model results. The main features found include:

• event occurrence rate was about 1 event per year

• event duration was about two days,
• the water depth of the events was within several hundred meters from the bottom,
• BTW propagated from the east to west along the Escarpment.

Figure 20 shows a BTW event that appeared in the 32-year model run. During the event the BTW caused relatively strong
bottom currents at several locations along the Escarpment. From the time series of this event, we saw that the event started in
the middle of the Gulf at about 88oW and 26oN where the upper-layer signature of the Loop Current was located, and the
strong current patches propagated from there to the west. The bottom current characteristics estimated from the model output
matched the observations (Hamilton, 2001, 2007; McKone, 2007). However, the magnitudes of the model current events
were lower than the observed values. For the event shown in Figure 20, the maximum bottom current speed was 35 cm/s, but
the measured speed could be about 100 cm/s. One possible reason for the model results being low is the lower vertical model
resolution near the bottom. Since strong bottom currents could be an issue for pipeline design and installation operations, it
is necessary to improve model performance to meet the requirement for this type of application.

Sigsbee
Escarpment

Figure 20. A bottom topographic wave event appeared during the 32-year model run in which relatively strong currents occurred at several places
along the Escarpment.

Conclusions

Results of the original 1/3o and 1/12o HYCOM models of the Gulf of Mexico circulation were evaluated against measurement
based data. The evaluation showed that the original HYCOM is capable of generating general current features in the Gulf of
Mexico except in terms of the Loop Current penetration into the Gulf and current speeds, i.e., the Loop Current and its eddies
did not penetrate far enough north in the Gulf and current speeds were low. Through a systematic sensitivity analysis we
identified a physical mechanism which controls the Loop Current evolution and eddy separation. We then implemented a
"Hybrid Boundary Condition" to simulate this physical mechanism which led to significant improvement the model
18 OTC 20826

performance. The Hybrid Boundary Condition consists of a combination of the open boundary conditions of the original
case and a case with a high inflow and high angle through the Yucatan Strait. The case we used in the long-term run
combined 50% the original case and 50% the high inflow and high angle case. A long-term current simulation using the
improved 1/12o HYCOM model with the Hybrid Boundary Condition was performed for a 32-year period.

The improved 32-year model results have been compared with the available data, and the comparison indicated that the
model results showed a realistic Loop Current climatology, including the Loop penetration, eddy path variability, eddy
occurrence and eddy shedding rate. The model results also produced realistic surface current extremes, and they can be used
for improving surface current design conditions for the deep water Gulf region.

The bottom currents near the Sigsbee Escarpment were also examined. The examination showed that the characteristics of
the bottom current events matched the observations, but the magnitudes of the current events predicted by the model were
lower than the observations. One possible reason was the lower vertical resolution near the bottom.

Exploration and production operations are moving to deeper and deeper water areas in the Gulf of Mexico, and the impact of
the Loop Current and its eddies on these operations in the area can be significant. The specification of design currents
requires extensive knowledge of the Loop/eddy system due to the high variability of the Loop Current and its eddies. The
currents from the improved 32-year model run broaden our knowledge base of the Loop/eddy system and can be used to
provide improved criteria for design of structures for the Gulf of Mexico.

Acknowledgments

We thank A. J. Wallcraft for providing timely help at the beginning of the work when we set up the model in our computing
environment, M. J. Santala for his contribution to the work and E. P. Berek for valuable discussion and comments on this
paper. We are also grateful to CASE JIP for the GEM database used in this study.

References

1. American Petroleum Institute (2009), "RP2MET: Recommended Practice for Derivation of Metocean Design and Operating
Conditions", Draft under API review
2. Berger, T. J., Hamilton, P., Singer, R. R., Leben, R. R., Born, G. H. and Fox, C. A. (1996) "Louisiana/Texas Shelf Physical
Oceanography Program: Eddy Circulation Study, Final Synthesis Report", OCS Study MMS 96-0052. U.S. Dept. of the interior,
Minerals Management Services, Gulf of Mexico OCS Region, New Orleans. LA.
3. Cochrane, J. O. (1972) "Separation of an Anticyclone and Subsequent Developments in the Loop Current", in Contribution on the
Physical Oceanography of the Gulf of Mexico, Texas A&M Univ. Oceanography Study, Vol. 2, edited by L. R. A. Capurro and J. L.
Reid, 91-106, Gulf Publishing, Houston
4. Cooper, C. and Forristall, G. Z. and Joyce, T. M. (1990), "Velocity and Hydrographic Structure of Two Gulf of Mexico Warm-Core
Rings", J. Geophys. Res., 95, C2, 1663-1679
5. Hamilton, P. (1990), "Deep Currents in the Gulf of Mexico", J. Phys. Oceanogr., 20, 1087-1104
6. Hamilton, P. and A. Lugo-Fernandez (2001), "Observation of High Speed Deep Currents in the Northern Gulf of Mexico", Geophys.
Res. Lett., 20, 2867-2870
7. Hamilton, P. (2007) "Deep-Current Variability near the Sigsbee Escarpment in the Gulf of Mexico", J. Phys. Oceanogr., 37,708-726
8. McKone, K., N. D. Walker, and E. Weeks (2007), "Full-Water Column Currents near the Sigsbee Escarpment (91-92ºW. Longitude)
and Relationships with the Loop Current and Associated Warm and Cold-Core Eddies", U.S. Dept. of the Interior, Minerals
Management Service, Gulf of Mexico OCS Region, New Orleans, LA. OCS Study MMS 2007-056. 107 pp.
9. Oey, L.-Y., Lee, H.-C., Schmitz Jr., W. J. (2003), "Effects of Winds and Caribbean Eddies on the Frequency of Loop Current Eddy
Shedding: A Numerical Model Study", J. Geophys. Res. Vol. 108, No. C10.
10. Oey, L.-Y. (2004), "Vorticity Flux through the Yucatan Channel and Loop Current Variability in the Gulf of Mexico", J. Geophys.
Res., Vol. 109, C10004.
11. Reid, R. O. (1972), "A Simple Dynamic Model of the Loop Current", Contributions on the Physical Oceanography of the Gulf of
Mexico, L. R. A. Capurro and J. L. Reid, Eds., Vol. 2, Gulf Publishing, 157-159.
12. Srinivasan, A. (2004), personal communication, Rosenstiel School of Marine and Atmospheric Science, University of Miami.
13. Tolman, H. L. (2002), "User Manual and System Documentation of Wave Watch-III Version 2.22", NOAA/NWS/NCEP/NMAB
Technical Note 222, 133pp.
14. Wallcraft, A., Carroll, S. N., Kelly, K. A., Rushing, K. V. (2003), "Hybrid Coordinate Ocean Model (HYCOM) User's Guide, Version
2.1", http://www.hycom.org/attachments/063_hycom_users_guide.pdf