Name: Sumalinog, Joycelyn P.

Section: 11-Samson
Strand: Stem

Analysis of Oceanography Surface Current Pattern

Compiled and Arrange Review Article

Surface Currents in operational oceanography: Key application, mechanisms, and methods

https://www.tandfonline.com/doi/full/10.1080/1755876X.2021.1903221

Surface currents are a central subject in operational oceanography with

rapidly growing observing and forecasting capabilities. We are entering an era where
observations provide a sufficient level of detail, and numerical models the predictive
skill, for applications to use surface current information to aid marine safety, value
creation and environmental monitoring. Observation techniques now provide real-time
surface current fields at the scale of kilometres (Isern-Fontanet et al. 2017), and high-
resolution ocean models are on the verge of having predictive skill on short time scales
(Jacobs et al. 2014b; Sandery and Sakov 2017; Christensen et al. 2018). In this context,
we consider short time scales, i.e. time scales characterised by the inertial period (hours
to a few days).

This work is motivated by challenges in obtaining and understanding observations and

model data of ocean currents near the surface. Differing designs of Lagrangian drifters,
as illustrated in Figure 1, show very different behaviour (Poulain et al. 2009; Morey et
al. 2018), and Eulerian current meter poorly resolve the surface. A challenge in
modelling is to resolve vertical shear in the upper few centimetres to decimetres that
has a substantial bearing on some drift applications (Figure 1). Consistent synthesis
between in situ measurements, remote sensing and modelling is needed in the analysis
of surface currents because each method provides a unique representation of surface
current dynamics, as illustrated in Figure 2. Yet there is no consensus on which depth
and time scale the term ‘surface current’ refers to (Laxague et al. 2018).

Figure 1. Schematic view of surface current regimes and types of surface drifters. In this
view, the ocean surface boundary layer is divided into the friction layer, the constant flux
layer, and the Ekman layer. The entire boundary layer is characterised by elevated
turbulent mixing, with largest and most efficient mixing in the interior and lower mixing
near the surface and at the base of the mixed layer. Surface currents are characterised
by strong wind-driven shear of velocity magnitude in the friction layer and the constant
flux layer, and a strong shear in direction throughout the Ekman layer. The direction of
the deep ocean current is arbitrary in this illustration and is in practice not related to the
wind direction. Surface drifters, from left to right: Bamboo plate, iSphere drifter, drogued
CODE-drifter, drogued SVP-drifter.

For most of these applications, usage of ocean currents from

operational forecasting has long been hampered by the chaotic nature of oceanic
flow and its variability on scales not covered by observation networks. Ocean
currents are comprised of numerous phenomena, from large-scale geostrophic
currents, to eddies and fronts at intermediate scales (Chelton et al. 2007), and
Langmuir circulation cells that approach the scale of oceanic turbulence
(McWilliams et al. 2012). Other components of ocean currents are better
distinguished by their temporal variation, for example, tides and wind-driven near
inertial waves (Alford and Gregg 2001).

A frequent use of operational ocean circulation models is the prediction of drift

trajectories (Griffa et al. 2007). For oceanic drift at or near the ocean surface, it is
critical to use the current at the target depth for the application considered, as
drift in the uppermost part of the ocean differs greatly with depth. However,
ocean models often do not resolve the near-surface gradients of wind-induced
shear (e.g. Chassignet and Verron 2006; Dagestad and Röhrs 2019), or they
lack coupling mechanisms for air-wave-sea interactions to describe strongly
forced drift in severe weather events.

When referring to surface currents, it should always be clear which part of the
ocean circulation is meant. We suggest a classification for surface currents in this
review, based on a review of surface current mechanisms and applications that
require current information. A clearer definition of surface currents was also
highlighted as a priority by the surface current working group arising from the first
meeting of the Australian Forum for Operational Oceanography in 2015. They
recommend that users and suppliers of ocean information be very clear about
what depth below the surface and averaging timescales are applicable.

Section 2 describes the most common applications, and what type of information

they require. We provide an overview of the physical mechanisms that govern
surface currents in Section 3. Section 4 reviews observation techniques and to
which application each of them is best suited. Surface current products from
models are discussed in Section 5, with a particular section considering particle
tracking models. The discussion (Section 6) includes a suggested terminology for
surface currents, and a view on the variability and predictability of surface
currents.
Other reviews that treat surface currents include Isern-Fontanet et al. (2017),
which focuses on remote sensing and data assimilation, Shutler et al. (2016) and
Le Traon et al. (2015), who treat satellite observations of the ocean surface, and
Lumpkin et al. (2017), covering the use of Lagrangian drifters to study surface
currents. Possibilities for short-term drift predictions are explored in Christensen
et al. (2018), and LaCasce (2008) and van Sebille et al. (2018) consider a range
of analysis techniques for Lagrangian data relevant for surface circulation.

The data presented in Figure 8 includes all model grid points within the
circle in Figure 6 for a time period of 48 h, which is the time scope of a
hypothetical forecast in this case. The histograms in Figure 8(a,b) highlight the
prevalence of strong northeastward and westward currents in the region, which
are the result of the coastal jet current as well as tides. The time-dependent
direction histogram in Figure 8(c) also shows when each tidal phase is to be
expected. Together with the time series of total current speed in Figure 8(d), one
can infer that the strongest currents are up to 0.6 m/s and in northeasterly
direction in the selected region. Figure 6 indicated that the extreme current
speeds should be expected in the southern part of the encircled region due to the
coastal jet. This information can be extracted from the model data despite the
existence of unpredictable mesoscale features.

Since operational applications require data on time scales of hours to days, wind-
driven currents and tides exhibit a major relevance in operational oceanography.
Since these are predictable, a useful forecast of surface currents can be issued
by separating predictable components from other parts, i.e. mesoscale features.
In addition, the low-frequency geostrophic flow is a matter of now-casting,
already provided in operational context.

Sea surface current response patterns to tropical cyclone

https://www.sciencedirect.com/science/article/pii/S0924796320300415?
fbclid=IwAR0DoWakNvK2vaBX6h2GOWLy3fhWmQeIU_bZHFJ0_k3CRd3e_TZ
m0lCqqQQ

Tropical cyclones (hereafter TCs) are strong synoptic systems. When

moving over the ocean, TCs deepen the surface ocean's mixed layer and induce sea
surface cooling and subsurface warming (the “heat pump” effect). TCs also cause
strong upwelling and cool the upper ocean (the “cold suction” effect). Consequently, sea
surface is cooled by TCs (Pun et al., 2018; Wu et al., 2018; Yue et al., 2018), and the
subsurface is generally warmed with a cold or warm anomaly that depends on the
relative importance of the mixing and upwelling at the different observation positions
(D'Asaro et al., 2007; Sun et al., 2015; Zhang et al., 2016, Zhang et al., 2018, Zhang et
al., 2019). TC-induced upper ocean thermal response may contribute to the global
ocean heat uptake and transport (Emanuel, 2001; Li et al., 2016) and influence past and
potential future climate change (Fedorov et al., 2010; Sriver et al., 2010). Similar to
thermal response, TC-induced mixing and upwelling also modulates the upper
ocean's salinity structure (Domingues et al., 2015; Zhang et al., 2016; Yan et al., 2017)
and causes surface plankton bloom and enhancement of primary production by bringing
subsurface nutrients to surface (Babin et al., 2004; Ye et al., 2013; Huang and Oey,
2015; Vidya et al., 2017).
TCs also import a large quantity of mechanical energy into the ocean. First, TCs cause
ocean surface storm surges and surface waves (Drost et al., 2017; Li et al., 2018).
Second, TC winds drive the ocean surface current (Price, 1981; Price, 1994; Chang et
al., 2014; Guan et al., 2014; Xu et al., 2019). Third, the oceanic gravitational
potential energy also increases due to the surface cooling and subsurface warming
induced by TCs (Liu et al., 2008). Studies show that during years 1984–2003, the wind
stress of tropical cyclones puts ~1.62 TW of kinetic energy into the surface waves,
~0.10 TW into the surface currents, and there was ~0.05 TW of gravitational potential
energy increase (Liu et al., 2008). The kinetic energy inputs to surface currents by a TC
can significantly influence ocean interior dynamics through the propagation of near-
inertial waves (Gill, 1984; Nilsson, 1995; Jaimes and Shay, 2010), although the kinetic
energy is normally one order smaller than the energy transformed to surface waves.
Most of the oceanic responses (i.e., temperature, salinity, plankton, current, waves,
mixing and upwelling) show a rightward bias in the Northern Hemisphere and a leftward
bias in the Southern Hemisphere (Chang et al., 2013; Prakash and Pant, 2016; Wang et
al., 2016; Wu et al., 2018, Wu et al., 2020). The key mechanism causing the oceanic
response asymmetry has been qualitatively explained by Price (1981), who stated that a
TC wind stress vector turns clockwise (anti-clockwise) with time on the right (left) side of
the track and roughly resonates with the surface ocean current in the Northern
(Southern) Hemisphere, and the asymmetry of TC wind field is not the main reason for
the ocean response asymmetry. It is the mechanism for surface current response
asymmetry but it is generally used by researchers to explain the asymmetry of other
responses such as temperature, salinity and plankton (e.g. D'Asaro et al.,
2007; Domingues et al., 2015; Sun et al., 2015; Yue et al., 2018; Zhang et al.,
2016, Zhang et al., 2018, Zhang et al., 2019) because surface current asymmetry
results in a stronger mixing by the shear instability on the right (left) side of a TC track
and the net surface rightward (leftward) advection, which leads to the rightward (left)
bias of the responses in the Northern (Southern) Hemisphere.
The natural following questions of the surface current response to a TC are: (1) what
are the possible sea surface current response patterns to a TC? (2) how do different TC
configurations and environmental conditions control the sea surface response
patterns? Geisler (1970) did a theoretical study and showed that a response within the
range of the TC wind field dominate when TC moves slowly, and an internal wake in the
lee of TC dominate when TC moves fast. Some other works also did part of the two
questions by doing a couple of simulations that changes TC translation speed
(e.g. Chang and Anthes, 1978; Samson et al., 2009), Coriolis frequency (e.g. Chang
and Anthes, 1978) or nondimensional translation speed which is the ratio of the local
inertial period to the TC residence time (e.g. Greatbatch, 1984; Price, 1994). This paper
tries to further study this topic. The following of this paper will firstly analyze the surface
current response patterns during three tropical cyclones that passed over our
buoy/mooring array in the northern South China Sea during the years 2014 and 2016
based on a combination of observations and results from a three-dimensional numerical
model simulation and a one-dimensional semi-analytical model. The one-dimensional
semi-analytical model provides a simple and easy-to-use method to estimate the
surface current response pattern to TCs when TCs and associated environmental
configurations are given. This paper will show how TC and ocean configurations
(especially nondimensional translation speed) control the ocean current response
patterns and also briefly study the distribution and variation of surface current response
extremes and asymmetry.

We deployed five observation stations during the years 2014 to 2015 (Fig.
1a) and three observation stations during the year 2016 (Fig. 1b). Each station consists
of a buoy and a mooring with the exception of Stations 3 and 8, which did not have
moorings. See Appendix A for details of the observation stations. Several TCs passed
over our observation arrays during the years 2014 to 2016. This paper needs the TC
cases with both surface wind and current observations, and TCs should also be close
enough to the observation arrays that their influence was obvious; thus, Rammasun
(2014), Kalmaegi (2014) and Sarika (2016) were chosen. See Table 1 for their basic
characteristics. We also plotted the TCs Matmo (2014), Fung-wong (2014), Aere (2016)
and Haima (2016) in Fig. 1 because they may influence the observations of Rammasun,
Kalmaegi and Sarika. Fig. 1. (a): Tracks of Rammasun (purple), Matmo (gray),
Kalmaegi (black) and Fung-wong (brown) during the year 2014, showing
the TC positions every 6 h (dots). Text boxes show the dates and sustained maximum
wind speeds at UTC 00:00 on the shown dates marked by dots. The positions of the
observation stations are denoted by big numbered red dots. (b): Same as (a) but for TC
Aere (purple), Sarika (gray) and Haima (black) during the year 2016.

This study uses the three-dimensional version of the Price-Weller-Pinkel model

(3DPWP; Price et al., 1986, Price, 1994), which has been successfully used in previous
studies to simulate and understand the physical processes involved in oceanic
responses to a TC (Sanford et al., 2007; Sanford et al., 2011; Guan et al., 2014; Zhang
et al., 2016, Zhang et al., 2018; Pun et al., 2018). The model domain spans 2400 km in
the across-track direction and 12,000 km in the along-track direction with a horizontal
resolution of 8 km. The model depth is 1660 m (1400 m) for Rammasun and Kalmaegi
(Sarika), and the vertical resolution is 10 m. The model starts from rest, and the initial
temperature and salinity were set as horizontally homogeneous, using three-day
average profiles of all of the available observation stations before the TCs' arrival,
except that the initial temperature and salinity for Rammasun is the average profiles in
July, because most of the buoys were deployed after Rammasun.
Table 1 shows the parameters of the three TCs when they traveled through our
observation array, which are also the variables used for the 3DPWP model simulation.
We assumed that the TC translation speed (U) was constant, and the TC wind field was
constructed using a maximum wind speed (Vmax) and a radius of maximum wind (rm)
based on the Sea, Lake and Overland Surge from Hurricanes (SLOSH) method
(Jelesnianski, 1966). Note that the rm of Kalmaegi is ~130 km if estimated from the
observations (see Appendix in Zhang et al., 2016), while it is ~40.13 km based on the
JTWC measurements. This discrepancy may be due to the wind field method we
choose, but may also be due to the fact that the JTWC is estimated using the data from
cloud tops, and our observations are from the buoys and monitored at the air-sea
surface. This discrepancy merits further study as it is merely studied before because the
lack of direct ocean surface wind observation during TCs. We use the same ratio
(130/40.13 ≈ 3.24) to amplify the rm from the JTWC during Rammasun and Sarika for
model simulation. The drag coefficient for wind stress was taken from Oey et al. (2006),
which is consistent with Large and Pond (1981) for low-to-moderate winds and
with Powell et al. (2003) for high winds. The Coriolis frequency for Rammasun and
Kalmaegi corresponds to the latitude of Station 3 at 18.7°N (Station 7 at 19.5°N) and an
inertial period of 37.33 h (35.85 h).

This section shows the surface distribution of the current speed, along-track
current speed and across-track current speed of the 3DPWP results (Fig. B1) and semi-
analytical model results (Fig. B2) as a supplement to Section 4.1. In the 3DPWP results,
the largest positive kinetic energy input was on the right rear part of the TCs, at 112 km,
250 km and 168 km (128 km, 128 km and 112 km) on the along-track (across-track)
direction. In the ideal result, the largest positive kinetic energy input was on the right
rear part of the TCs, at 318.6 km, 539.5 km and 330 km (70.2 km, 91 km and 78 km) in
the along-track (across-track) direction. Note that because of neglecting horizontal
advection in the semi-analytical model, the current speed pattern (Fig. B2a-c) seems
similar to an accumulated kinetic energy input into the currents (Fig. 7g-i). In the
3DPWP model result, the current speed (Fig. B1a-c) spreads more than the
accumulated kinetic energy inputs into the currents (Fig. 6g-i). Fig. B1. Surface current
speed (a-c), the corresponding along-track (d-f) and across-track (g-i) current speed (d-
f) during the three tropical cyclones that are reproduced by the 3DPWP model. The
positive values correspond to forward (rightward) in the along-track (across-track)
current. The red dot is the tropical cyclone center and the two black circles surrounding
it are 1 and 3 times of the radius of maximum wind, respectively. The solid line is the
tropical cyclone track. The dashed lines are the positions of the observation stations;
the lines in (b, e, h) represent stations 1, 4 and 2 from top to bottom. (For interpretation
of the references to colour in this figure legend, the reader is referred to the web version
of this article.)
Measurement and Analysis of Ocean Current using High Frequency(HF) Radar
Observation in the Bali Strait

https://iopscience.iop.org/article/10.1088/1755-1315/893/1/012053/meta

The Bali Strait has an important role in various important sectors such as
maritime, transportation, trade, and tourism and its role in connecting the most populous
island to the island with the best tourism in Indonesia. Besides, the Bali strait is
regarded as an international shipping route, in which this route is getting busier every
year with many ships, both locally and internationally, passing this strait. Thus, accurate
maritime information is needed, which is expected to reduce the number of ship
accidents. Ocean current information is essential for shipping safety, especially in the
strait areas where the current velocity is relatively high such as Bali Strait, which
connects the Indian Ocean to the Java Sea, has strong currents due to tidal
amplification enters a narrow strait [1]. Issuing accurate and reliable information on
ocean currents will help the transportation sector run efficiently. The Strong current in
the Bali strait also can be used as an alternative energy source, such as tidal power
plants. The total potential tidal current power plant in the Bali Strait is 1507 MW [2].
Although ocean current data in the Bali Strait is essential, observations of ocean
currents in the Bali Strait are still limited. Thus, BMKG, represented by the Maritime
Meteorology Center, has made efforts to provide ocean current observation data in Bali
by installing HF Radar since 2018. HF Radar is an instrument that uses high-frequency
(3-50 MHz) radio waves to map sea surface currents remotely. HF Radar works on the
principle of Bragg-scattering and the Doppler shift of radio waves hitting rough sea to
determine the current velocity [3]. HF Radar has many advantages, such as producing
near real-time ocean surface current data with a high spatial and temporal resolution,
can operate in all conditions, is not harmful to the environment, and low power
consumption [4], [5]. HF Radar has been used for coastal and disaster management
purposes such as search and rescue, marine pollution management, marine dredging
and construction, distribution of fish larvae, also for tsunami early warning [6]. HF Radar
has been widely used internationally, and there is even an international organization,
the Global HF Radar Network (http://global-hfradar.org/), a new recognized international
network that aims to connect world researchers interested in developing this instrument.
However, this technology is still very rarely studied and applied in Indonesia. Usually,
scientists perform numerical simulations to understand the ocean current, which of
course, still have errors and biases. Currently, BMKG has several HF Radar sites
installed for research or operational purposes, such as Seasonde CODAR HF Radar
(Coastal Ocean Dynamics Applications Radar) in the Bali Strait and Labuan Bajo
Waters, WERA (Whelan Radar) in the Sunda Strait, and Southern Yogyakarta waters.
Further research needs to be done to find out more details about hf radar capability to
map the ocean current in Indonesian waters. This research examines how the HF
Radar is used to analyze ocean currents in the Bali Strait.
 Comparison of 3-hour moving average meridional current between HF Radar
and ADCP in figure 3 shows a very high correlation of 0.813 with an RMSE of 0.22 m/s.
This result is similar to [11] that found the HF Radar RMSE and correlation in Gibraltar
Strait were 0.31 to 0.81 and 0.08 to 0.22 m/s. According to [12], the RMSE value of HF
Radar ranges from 0.06 to 0.30 m/s depending on the characteristics of HF Radars'
location. In contrast to the meridional current, the accuracy for zonal current is not very
good. The correlation and RMSE values of zonal currents are 0.235 and 0.402 m/s,
respectively. This condition presumably because the meridional current's velocity is very
strong and is only measured by two HF radars with a north-south configuration. Thus it
may not be easy to separate the meridional and zonal components. According to [13],
strong currents can distort the HF Radar antenna pattern. The accuracy will likely be
better if there are three HF pairs; one of the HF radars is placed on Bali's side. Also, the
possibility of bias can be caused by various things, such as interference of other
devices at similar frequencies. For more comprehensive testing, it needs to take a
longer time to see the HF Radar's overall accuracy. Several studies have compared the
HF Radar data with other instruments for at least three months Figure 3 Comparison of
meridional current between HF Radar and ADCP: a. Meridional current, b. Zonal
current.

Although the verification results on 24 April to 1 May 2019 HF Radar against ADCP
showed less accuracy in the zonal component, we still used it to describe the overall
monthly current characteristics. Because the data studied in this study is for July 2020,
we assumed that it is more accurate than the previous one in 2019 because the HF
Radar has been maintained every year, such as the antenna pattern calibration. It is
necessary to verify again to see how accurate the HF Radar after maintenance. The
mean and standard deviation of surface currents in the Bali Strait in July 2020 showed a
high velocity in the middle of the strait with a southward direction. The mean and
standard deviation of 1.09 m/s and 0.82 m/s, respectively. The strong current is in the
middle because it is the main channel of the Bali Strait waterway. The Indonesian
throughflow (ITF) also reaches its peak in July [14] caused the southward current in the
Bali Strait. The high standard deviation value indicates that the velocity variation in the
middle of the strait is highly varied. The meridional current on the mid-eastern side is
southward with a maximum speed of 1.08 m/s, while on the western side is northward
with a relatively low speed (max ~0.5 m/s). For the zonal current, the current is more
dominant towards the eastern with a maximum speed of 0.66 m/s. The monthly mean of
meridional current (figure 4d) shown that the dominant current in the mid-eastern side of
the strait is the tidal current, while on the western side influenced by residual current.
The distribution of sea surface currents in the Bali Strait (current rose) shows that in
July, the current is more dominant to south and southeast for speed of more than 1.00
m/s, while for speed less than 1.00 m/s is distributed in any direction Figure 4
Characteristics of current in Bali Strait in July 2020. The colorbar is not the same for all
maps for better visualization.

HF Radar data verification against ADCP data shows a good agreement for
the meridional current but not suitable for the zonal current. Although the accuracy of
HF radar needs to be improved, this research shows that the HF Radar data can
comprehensively describe the Bali Strait's current characteristics. The dominant current
is the tidal currents with different effects from various locations. On the mid-eastern
side, it is strongly influenced by harmonic currents in a more dominant southward
direction. The velocity pattern of meridional harmonic currents during high tide (flood)
and low tide (ebb) is different. The northward maximum current velocity occurs between
the two highest tides for high conditions, while at ebb condition, the southward
maximum current velocity occurs at the lowest tide. On the western side, the meridional
current's velocity is not as large as that of the mid-eastern side, and its direction is more
dominant to the north. The spectral density analysis result showed that the 0.5 and 1-
hour oscillation was found in both wind and current data suggested that the wind causes
residual current.