Relative importance of ENSO and IOD on interannual variability of Indonesian Throughflow transport

Abstract

Introduction
The Indonesian Throughflow (ITF) connects the Pacific Ocean and the Indian Ocean. It plays an important role in the global ocean circulation system. The interannual variability of ITF transport is largely modulated by climate modes, such as Central-Pacific (CP) and Eastern-Pacific (EP) El Niño and Indian Ocean Dipole (IOD). However, the relative importance of these climate modes importing on the ITF is not well clarified.

Methods
Dominant roles of the climate modes on ITF in specific periods are quantified by combining a machine learning algorithm of the random forest (RF) model with a variety of reanalysis datasets.

Results
The results reveal that during the period from 1993 to 2019, the average ITF transport derived from high-resolution reanalysis datasets is -14.97 Sv with an intensification trend of -0.06 Sv year⁻¹, which mainly occurred in the upper layer. Four periods, which are 1993–2000, 2002–2008, 2009–2012 and 2013–2019, are identified as Niño 3.4, Dipole Mode Index (DMI), no significant dominant index, and DMI dominated, respectively.

Discussion
The corresponding sea surface height differences between the Northwest Tropical Pacific Ocean (NWP) and Southeast Indian Ocean (SEI) in these three periods when exist dominant index are -0.50 cm, 0.99 cm and -3.22 cm, respectively, which are responsible for the dominance of the climate modes. The study provides a new insight to quantify the response of ITF transport to climate drivers.

Full text

Relative importance of ENSO
and IOD on interannual
variability of Indonesian
Throughflow transport
Aojie Li
1
, Yongchui Zhang
1
*, Mei Hong
1
, Jian Shi
1
and Jing Wang
2
1
College of Meteorology and Oceanography, National University of Defense Technology,
Changsha, China,
2
CAS Key Laboratory of Ocean Circulation and Waves, Center for Ocean Mega-
Science, Institute of Oceanology, Chinese Academy of Science, Qingdao, China
Introduction: The Indonesian Throughflow (ITF) connects the Pacific Ocean and
the Indian Ocean. It plays an important role in the global ocean circulation
system. The interannual variability of ITF transport is largely modulated by climate
modes, such as Central-Pacific (CP) and Eastern-Pacific (EP) El Niño and Indian
Ocean Dipole (IOD). However, the relative importance of these climate modes
importing on the ITF is not well clarified.
Methods: Dominant roles of the climate modes on ITF in specific periods are
quantified by combining a machine learning algorithm of the random forest (RF)
model with a variety of reanalysis datasets.
Results: The results reveal that during the period from 1993 to 2019, the average
ITF transport derived from high-resolution reanalysis datasets is -14.97 Sv with an
intensification trend of -0.06 Sv year
-1
, which mainly occurred in the upper layer.
Four periods, which are 1993–2000, 2002–2008, 2009–2012 and 2013–2019,
are identified as Niño 3.4, Dipole Mode Index (DMI), no significant dominant
index, and DMI dominated, respectively.
Discussion: The corresponding sea surface height differences between the
Northwest Tropical Pacific Ocean (NWP) and Southeast Indian Ocean (SEI) in
these three periods when exist dominant index are -0.50 cm, 0.99 cm and
-3.22 cm, respectively, which are responsible for the dominance of the climate
modes. The study provides a new insight to quantify the response of ITF transport
to climate drivers.
KEYWORDS
Indonesian throughflow (ITF), upper layer, lower layer, El Niño-Southern Oscillation
(ENSO), Indian Ocean Dipole (IOD), random forest (RF) model
Frontiers in Marine Science frontiersin.org01
OPEN ACCESS
EDITED BY
Zexun Wei,
Ministry of Natural Resources, China
REVIEWED BY
Qin-Yan Liu,
Chinese Academy of Sciences (CAS), China
Xiaohui Liu,
Ministry of Natural Resources, China
*CORRESPONDENCE
Yongchui Zhang
zyc@nudt.edu.cn
RECEIVED 08 March 2023
ACCEPTED 28 April 2023
PUBLISHED 12 May 2023
CITATION
Li A, Zhang Y, Hong M, Shi J and Wang J
(2023) Relative importance of ENSO and
IOD on interannual variability of Indonesian
Throughflow transport.
Front. Mar. Sci. 10:1182255.
doi: 10.3389/fmars.2023.1182255
COPYRIGHT
© 2023 Li, Zhang, Hong, Shi and Wang. This
is an open-access article distributed under
the terms of the Creative Commons
Attribution License (CC BY). The use,
distribution or reproduction in other
forums is permitted, provided the original
author(s) and the copyright owner(s) are
credited and that the original publication in
this journal is cited, in accordance with
accepted academic practice. No use,
distribution or reproduction is permitted
which does not comply with these terms.
TYPE Original Research
PUBLISHED 12 May 2023
DOI 10.3389/fmars.2023.1182255

1 Introduction
The Indonesian Throughflow (ITF) originates from the western
Pacific Ocean, and then passes through the Indonesian Sea to enter
the Indian Ocean. The ITF carries a large amount of warmer and
fresher waters with an annual average volume transport of
approximately 15 Sv (1 Sv=10
6
m
3
s
-1
)(Wijffels et al., 2008;
Sprintall et al., 2009;Gordon et al., 2010;Susanto et al., 2012;Liu
et al., 2015;Sprintall et al., 2019), and heat transport of 0.24–1.15
PW (1 PW = 10
15
W) from the Pacific to the Indian Ocean (Hirst
and Godfrey, 1993;Vranes et al., 2002;Tillinger and Gordon, 2009;
Xie et al., 2019;Zhang et al., 2019). It has a significant impact on the
thermohaline structure and velocity profiles flowing through the
oceans (Lee et al., 2019;Pang et al., 2022). This provides an
important low-latitude ocean channel for the transmission of
climate signals and anomalies in the global thermohaline
circulation. The ITF regulates the local atmosphere system by
influencing air-sea exchange and precipitation at different time
scales, which in turn has a far-reaching effect on the global climate
(Gordon, 1986;Godfrey, 1996;Gordon, 2005;Sprintall et al., 2014;
Hu et al., 2019;Yuan et al., 2022). Thus, the variability of ITF
transport has to be understood to interpret climate change.
The ITF transport is not well determined observationally due to
several inflow and outflow channels, such as the Makassar Strait,
Maluku Strait, Halmahera Strait, Lombok Strait, Ombai Strait, and
Timor Strait. Among them, the Makassar Strait accounts for
approximately 77% of the ITF transport and is thus considered
the main inflow channel of the ITF (Du and Qu, 2010;Gordon et al.,
2010;Gordon et al., 2019). Therefore, several international
observation programs have been conducted there to detect the
changes in ITF transport, such as the Arlindo Mixing program, the
International Nusantara Stratification and Transport program
(INSTANT), and the Monitoring the ITF program (MITF). The
long-term mooring data of the Makassar Strait reveals that the
average thermocline (0–300 m) southward transport (9.1 Sv)
contributed about 73% of the total transport (12.5 Sv) (Gordon
et al., 2019). In addition to the mooring observation data,
temperature data measured by repeated expendable
bathythermograph (XBT) and Argo buoy data of IX1 section
were also used to deduce the geostrophic current transport of the
ITF (Wijffels et al., 2008;Liu et al., 2015). Based on 30 years of XBT
data –from 1984 to 2013 ITF geostrophic transport experienced a
strengthening trend of ~0.1 Sv year
-1
(Liu et al., 2015). In addition,
numerous numerical models and reanalysis data demonstrate a
relatively consistent interannual variability with the observed data
(Masumoto et al., 2004;Feng et al., 2013;Yuan et al., 2013).
The ITF variability is mainly driven by large scale sea level
gradients between the Pacific and Indian Ocean basins (Wyrtki,
1987). Specifically, the sea surface height (SSH) difference between
the northwest tropical Pacific (NWP) and the southeast Indian
Ocean (SEI) is a favorable indicator of ITF transport, which is
largely dominated by the El Niño-Southern Oscillation (ENSO) and
the Indian Ocean Dipole (IOD), respectively. The ITF is generally
strong (weak) during La Niña (El Niño) events (Meyers, 1996). This
is because the Pacific trade winds and the Walker circulation
strengthening (weakening), which leads to an increase (decrease)
of sea level in the western Pacific(Meyers, 1996;Gordon et al., 1999;
Sprintall and Revelard, 2014;Hu and Sprintall, 2016). During
negative (positive) IOD events, when the eastern and western
surface water of the tropical Indian Ocean appear abnormally
warm (cold) and cold (warm), downwelling (upwelling) occurs in
the eastern sea surface of the tropical eastern Indian Ocean. This
favors a positive (negative) sea level anomaly in that region and thus
suppresses (strengths) the ITF (Cai et al., 2011;Yuan et al., 2011).
Increasing number of studies show that IOD events have a more
significant impact on the ITF (Sprintall et al., 2009;Sprintall and
Revelard, 2014;Liu et al., 2015;Pujiana et al., 2019). However,
ENSO and IOD events often occur concurrently (Murtugudde et al.,
1998;Saji et al., 1999;Feng et al., 2001), and hence, it is hard to tease
out the individual effects of each climate mode on ITF variability.
Quantitative analysis of the influences of ENSO and IOD on ITF
changes are not yet well clarified.
In this study, four high-resolution reanalysis datasets are used
to detect the spatial-temporal variability of the ITF inflow and
outflow. A machine learning method is adopted to express whether
the ENSO or the IOD is the dominant climate driver for the ITF
variability during different periods. The study is explained as
follows. The four reanalysis datasets and the methods are
depicted in section 2. The temporal and spatial changes of ITF
transport are explained in section 3. The dominant climate indices
affecting the ITF in different periods are studied in section 4. The
possible mechanisms are discussed in section 5, and the conclusions
are given in section 6.
2 Data and methods
2.1 Mooring data
Mooring data are obtained from three straits: the Makassar
Strait, Ombai Strait, and Timor Strait. Among them, the Makassar
mooring data is procured from the INSTANT and MITF
observation projects, whereas the Ombai and Timor are from the
Integrated Marine Observing System (IMOS, 2022).
2.1.1 INSTANT
In January 2004, two moorings were deployed in the Labani
channel in the Makassar Strait as part of an international program
to monitor the major ITF inflow routes: 2°51.9′S, 118°27.3′E, and 2°
51.5′S, 118°37.7′E(Gordon et al., 2008;Gordon et al., 2010). In July
2005 and November 27, 2006, the moorings were repeatedly
recovered and redeployed. The moorings measured the three-
dimensional velocity components at 30 min intervals.
2.1.2 MITF
Only one mooring was deployed at 2°51.9′S, 118°27.3′E as part
of the MITF program on 22 November 2006, after the INSTANT
project. MITF was designed to receive data and was redeployed
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org02

every two years. Due to equipment transportation problems, there
are no data from August 2011 to August 2013. At the mooring
location, upward-looking and downward-looking acoustic Doppler
current profilers (ADCPs) were placed at 463 m and 487 m to
record flow data for the entire channel at a depth of 680 m. The
positions varied slightly throughout the observation period but
remained roughly the same. In August 2015, two ADCPs were
placed at the same buoy station at 498 m, and two c‐pods (small
autonomous instruments), which can measure the temperature
gradient spectrum using a fast thermistor, were placed at the
same position to measure the temperature microstructure
(Gordon et al., 2019). The MITF data up to August 2017 is used
in this study, though the project is still ongoing.
2.1.3 IMOS
The IMOS has moorings across both its National Mooring
Network and Deep Water Moorings facilities. This system provides
parameters such as temperature, salinity, dissolved oxygen,
chlorophyll estimates, turbidity, down-welling photosynthetic
photon flux (PAR), and current velocity, accompanied by depth
and pressure when available. The observations were made using a
range of temperature loggers, conductivity-temperature-depth
(CTD) instruments, water-quality monitors (WQM), ADCPs, and
single-point current meters. In this study, we use the single-point
mooring data (OMB) for the Ombai Strait at 125.08°E, 8.52°S,
which ranges from June 19, 2011, to October 21, 2015. For the
Timor Strait, we mainly use three mooring data: Timor North
(TNorth), Timor North Slope (TNSlope), and Timor South
(TSouth), which cover from June 14, 2011 to April 15, 2014.
These are the hourly mooring data, and the vertical depth
reaches 520.95m.
2.2 Reanalysis data
Four reanalysis datasets are used: Copernicus Marine
Environment Monitoring Service (CMEMS), OGCM For Earth
Simulator (OFES), Hybrid Coordinate Ocean Model (HYCOM),
and simple Ocean Data Assimilation Ocean/Sea Ice Reanalysis
(SODA), respectively.
2.2.1 CMEMS
For the CMEMS, the version GLOBAL_MULTIYEAR_
PHY_001_030 global reanalysis data with a monthly average from
1993 to 2019 is used. The data have a spatial resolution of 1/12° × 1/
12° (about 8 km × 8 km at the equator) and a total of 50 standard
layers in the vertical direction. The reanalysis data assimilate many
available observations. The time range is from 1993 to 2020, which
covers the most recent period of altimeter data (beginning with the
launch of TOPEX/Poseidon and ERS-1 satellites in the early 1990s).
2.2.2 OFES
OFES data is a global 0.1° × 0.1° × 54 layers model data forced
by NCEP winds. The output is an integration of more than 50 years.
The OFES data used is the monthly average from 1993 to 2017.
2.2.3 HYCOM
The HYCOM versions of GLBu0.08/expt_19.0, GLBu0.08/
expt_90.9, and GLBv0.08/expt_93.0 cover the time range 1993–
2012, 2013–2017 and 2018–2019, respectively. The global 0.08° ×
0.08° horizontal resolution and 40 vertical resolution of depth levels
are daily reanalysis data from 1993 to 2019.
2.2.4 SODA
SODA is a reanalysis data set that covers the global ocean
(except some polar sea areas) jointly developed by the University of
Maryland and Texas A&M University. The latest version, SODA
3.4.2, which adopts the Modular Ocean Model (MOM5) of 0.5° ×
0.5° × 50 layers (horizontal spacing at the equator 28 km, polar
location less than 10 km) is used in this study. The time range of the
monthly reanalysis data is from 1993 to 2019.
2.3 Methods
2.3.1 Transport calculation
Many methods are used to calculate the flow in channels,
including the volume flux (Anderson et al., 1986)andtheP-
vector methods (Chu, 1995). The volume flux method is used to
take full advantage of the high-resolution datasets:
Fv=o
nz
k=1
o
ns
i=1
vik
!·dx
i·dz
k:(1)
Among them, vi,k
!is perpendicular to the ith horizontal grid
and the kth vertical grid on the cross-section flow velocity, iis the
location of the cross-section grid point number, ns is the number of
grid points, xi is the distance between two adjacent grid points, kis
the number of vertical layers, nz is the number of vertical layers, and
dzk is the distance between two adjacent vertical layers.
2.3.2 Removal of potential dependency between
climate drivers
ENSO and the IOD events often occur synchronously; hence,
they can interact with each other. Therefore, linear regression is
used to eliminate the possible influence of Niño 3.4 on IOD (Saji
and Yamagata, 2003), as follows:
d
DMI =aNino 3:4+b, (2)
DMInew =DMI −
d
DMI :(3)
where d
DMI represents the linear fitting term of Nino3:4, aand
brepresent the trend and offset, respectively, and DMInew indicates
the new DMI, excluding the Nino3:4 trend item.
2.3.3 Random Forest model
A machine learning decision method, named Random Forest
(RF) is employed to determine the contribution of climate drivers to
the interannual variation of the ITF. RF is a kind of ensemble
learning algorithm. The general idea of RF is to train multiple weak
models to pack together to form a strong model. The performance
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org03

of the strong model is much better than that of a single weak model.
Hence, the results of the multi-models have higher accuracy and
generalization performance.
Unlike the simple linear correlation and regression methods, RF
methods can be used to study complex relationships between
variables and can reveal nonlinear and hierarchical relationships
between responses and predictors (Feng et al., 2022). RF builds the
model by combining predictors and evaluates the relative
importance of each predictor. In this study, an out-of-bag (OOB)
generated accuracy-based materiality measure is used. When
building the model, approximately one-third of the relevant data
was randomly selected for model verification. When variables in the
OOB samples are randomly disturbed, the average prediction
accuracy is defined as the important value of the corresponding
variable (Heung et al., 2014), which is expressed as the mean square
error:
MSEOOB =1
No
N
K=1
(Oi−PKOOB)2
:(4)
where Nis the number of observations, Oirepresents actual
data and PKOOB indicates the average of all OOB predictions across
all trees.
3 Variability of ITF
Mooring observation in the Makassar Strait reveals different
water masses in the upper (0–300 m) and lower (300–760 m) layers,
respectively. The ITF is separated into the upper and lower parts,
with a depth boundary of 300 m, to reveal the impacts of climate
modes on the vertical structures. To facilitate later research, the 0–
300 m layer is defined as the upper layer and the 300–760 m layer as
the lower layer. Due to the depth limitation of the Ombai and Timor
Straits, the lower layer is 300–520 m. In accordance with Li et al.
(2020), the inflow is defined as that the sections of Sulawesi Sea
(125°E, 1°N–6°N), Maluku Sea (125°E–127.5°E, 0.5°N), and
Halmahera Sea (128°E–131°E, 0.5°S), respectively, and the
outflow is defined as that across the eastern tropical Indian Ocean
section (114°E, 8°S–22°S) (Figure 1).
3.1 Validation of model data
Performances of the four reanalysis datasets are validated by
using the mooring observation located at the Makassar Strait,
namely the INSTANT and MITF programs. In accordance with
the two periods of INSTANT and MITF, the comparative results are
divided into two stages, which are January 2004 to late November
2006 and November 2006 to August 2017, respectively. An
approximate position at the same latitude as the mooring position
is selected to conducted verification when calculating the Makassar
Strait flow, which is 117°E–119°E, 2.5°S (Figure 1).
The comparative results are illustrated in Figures 2A,B. In the
upper layer, the mean mooring transport is -7.94 Sv (negative
means southward transport). It shows a prominent interannual
variability. The transport intensified from 2004 to late 2008, with an
intensification of -1.53 Sv, whereas from 2009 to mid–2011, it
slightly weakened by 0.32 Sv. The transport quickly weakened from
mid–2013 to mid–2016, whereas the ITF quickly intensified. The
relativity between the upper and lower layers was negative and
insignificant (correlation coefficient of -0.30). In the lower layer, the
mean transport is -3.10 Sv. The trends of mean transport during
2004–2011 and 2013–2017 were the opposite: there is an intensified
(weakened) trend in the former period and a weakened (intensified)
trend in the latter period in the upper (lower) layer. In addition, in
terms of interdecadal variability, the correlation between the upper
layer Makassar flow and PDO and NPO were significantly opposite
(Figures not shown). During 2004–2011 and 2013–2017, the
correlation coefficients between PDO and flow in the upper layer
were 0.78 and 0.75, respectively; whereas the correlation coefficients
between NPO and flow changes in the upper layer were -0.44 and
-0.35, respectively. All the correlation coefficients pass the 95%
significance test. However, the contributions of PDO in the ITF
transport are limited in the upper layer inflow and in the decadal
timescales, which will not be further considered in this study.
In the upper layer, the correlation coefficients of SODA and
CMEMS in the first and second periods are R1 = 0.8 and R2 = 0.96,
respectively, and the RMSE is 1.12 Sv, in comparison with the
observed data. Similarly, the OFES data in this layer exhibit low
correlation coefficients with the observed data in the first half and
the second half, where R1 = 0.25 and R2 = 0.7, respectively, and the
FIGURE 1
The topographic map of the Indonesian sea and the main flow system
of the ITF. The pathways of ITF are shown by orange lines with
arrows. The red  indicates the mooring station in the Makassar
Strait. The four red dots that appear in the enlarged view in the small
window of the lower right corner denote the mooring positions of the
Ombai and Timor Straits, and from north to south are Ombai Strait
(OMB), Timor North (TNorth), Timor North Slope (TNSlope), and Timor
South (TSouth), the first of which is the mooring position of the
Ombai Strait and the remaining three are the mooring positions of the
Timor Strait. The purple lines are the interception position of the
corresponding strait data validation. The dotted arrows in orange
represent the PacificOceanflow. The red and green solid lines
indicate the inflow and outflow cross-sections, respectively.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org04

RMSE is 0.82 Sv. It can be found that between July 2005 and
November 2006, the reanalysis data declined more than the
mooring observation data, which may be explained by that single
point data of Makassar mooring could not accurately reflect the
wholechangeintheentirechannel.Inthelowerlayer,the
correlation coefficients of OFES and CMEMS with the observed
data in the first and second half are R1 = 0.73 and R2 = 0.98,
respectively, and the RMSE is 0.89 Sv. The HYCOM data in this
layer has relatively low correlation coefficients of R1 = 0.49 and
R2 = 0.77, respectively, and the RMSE is 1.35 Sv, in comparison
with the observed data. All the correlation coefficients pass the 95%
significance test. It can be found that in the upper and lower layers,
the correlation between the reanalysis and mooring observation
data in the second period is better than in the first period, which
may be attributed to the reanalysis datasets dependency and the
limitation of mooring observations.
The mooring data of the Ombai and Timor Straits are
comprehensively collected to verify the applicability of reanalysis data
in the outflow area. Two sections along 125.08°E, 8.33°S–8.83°S and
127.35°E, 8.71°S–10.02°S in the Ombai and Timor Straits, respectively,
are selected to conduct the calculation. The low resolution of SODA
(only 0.1°) leads to a lack of lower layer data in the Ombai Strait, and
hence, only the upper layer of SODA is given below.
Figures 3A,Billustrate the comparison results in the Ombai
Strait. For SODA, the correlation with the observed data is low due
to the low resolution in the upper layer (R = 0.31, which passes the
95% significance test; RMSE = 0.61 Sv). Apart that, the correlations
of the other three reanalysis data are all high. Among them,
correlation coefficient of OFES is maximum which reaches 0.90
and RMSE is 0.45 Sv. In the lower layer, HYCOM has the maximum
correlation coefficient of 0.80 and lowest RMSE of 0.28 Sv, whereas
OFES has a relatively low correlation coefficient of 0.69 and RMSE
of 0.43 Sv. The validation results of the Timor Strait data are shown
in Figures 3C,D. It is found that CMEMS has the maximum
correlation coefficient of 0.71 and lowest RMSE of 0.14 Sv in the
upper layer, whereas HYCOM has a relatively low correlation
coefficient of 0.38 (passing the 95% significance test) and RMSE of
0.26 Sv. In the lower layer, CMEMS also exhibits a good correlation,
with a correlation coefficient of 0.63 and RMSEof 0.11 Sv. However, the
correlation coefficientsof the other three data arelow. Note that, all the
correlation coefficients pass the 95% significance test, which
demonstrate that all the four reanalysis datasets show consistency
with the observations,which gives the confidences to use the reanalysis
datasets to show the variability and mechanisms.
3.2 Spatial and temporal variability
of the ITF
Good performance of the eddy-resolving reanalysis datasets
reveals, the detailed spatial structures of the ITF in the inflow
and outflow.
Among the three inflow cross-sections (Figures 4A–C), the
Sulawesi Sea has the widest entrance, which spans 3.58°N–5.5°N,
A
B
FIGURE 2
Transport anomaly of the four reanalysis datasets and the mooring observation at Makassar. (A) Transport in the upper layer. The red line is the
mooring observation of INSTANT and MITF after a 13-month moving average. The black, blue, green, and pink lines are CMEMS, OFES, SODA, and
HYCOM, respectively. Due to the discontinuity of mooring data between 2004 and 2017, the correlation coefficients of the first and second halves
of the mooring after the 13-month moving average are calculated, respectively. (B) Same as (A) but for the lower layer. Negative values mean
southward transport anomaly.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org05

125°E. There are two opposite flows, which are westward in the
northern channel and eastward in the southern channel,
respectively. As the depth deepens, the eastward flow gradually
expands to the north, while the westward flow decreases in scope.
The eastward flow mainly exists in the range of 125°E, 1.83°N–3.5°
N (the maximum depth is 350 m south of 3.25°N, while it can be
extended to 760 m north of 3.25°N), with an average speed of
0.07 m s
-1
and a maximum speed of 0.23 m s
-1
at a depth of 20m
near 2.25°N. The westward flow in the 3.58°N–4.66°N region
mainly exists above the 115 m layer with an average speed of
-0.14 m s
-1
. The large value zone of westward flow mainly exists in
4.5°N–5.33°N (the depth can be extended to 760 m) and a
maximum speed is -0.59 m s
-1
at a depth of 85 m near 5°N.
There is a weak eastward flow near Mindanao Island. The standard
deviation of the flow gradually decreases with the increase of depth
overall. There has a larger standard deviation in the westward flow,
which can reach 0.13 m s
-1
, mainly concentrated in the upper 90 m
layer of 4.58°N–5.08°N. In the eastward flow range, the standard
deviation of the flow can reach a maximum of 0.09 m s
-1
, mainly
concentrated in the surface layer of 1.91°N–2.83°N.
The opposite flow structure is the same as the Maluku Sea
meridional sections (Figure 4B). In the western Maluku Sea (125°
E–126°E, 0.5°N): above 60 m layer, it flows northward with an
average speed of 0.05 m s
-1
, and the maximum is 0.16 m s
-1
at a
depth of 10 m near 125°E; within a depth of 60 m–300 m, it flows
southward with an average speed of -0.02 m s
-1
, and the flow
gradually decreases as the depth increases; within a depth of 300
m–450 m, there is a weak northerly flow; under 450 m, there is a
A
B
D
C
FIGURE 3
Transport anomaly of the four reanalysis datasets and the mooring observation in the Ombai and Timor Straits. (A) Transport in the upper layer of
Ombai Strait. The red line is the mooring observation of Ombai Strait after a 13-month moving average. The black, blue, green and pink lines are
CMEMS, OFES, SODA and HYCOM, respectively. (B) Same as (A) but for the lower layer of Ombai Strait. (C, D) Same as (A, B) but for Timor Strait.
Negative values mean southward transport anomaly.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org06

southward flow, which corresponds well to the observational finding
of the Maluku Sea intermediate western boundary current (Yuan
et al., 2022). In the eastern Maluku Sea (126°E–127.5°E, 0.5°N): above
60 m layer, there is a weak southward flow with an average speed of
-0.03 m s
-1
and the maximum southward flow velocity is -0.07 m s
-1
;
under 60 m, except for the presence of weak southward flow in depth
of 350 m–720 m and range of 126°E–126.41°E, it flows northward in
the western side (126°E–126.83°E, 0.5°N) and southward on the
eastern side (126.83°E–127.5°E, 0.5°N). The large standard deviation
of the flow velocity is mainly concentrated in the western upper layer
of Maluku Sea (125°E–126°E, 0.5°N), which can reach a maximum of
0.07 m s
-1
. With the increase of depth, the standard deviation
gradually decreases. However, below 450 m in the western of
Maluku Sea, there is still a relatively large variation zone with a
standard deviation of 0.02 m s
-1
.
In the Halmahera Sea section (0.5°S) (Figure 4C), the flow
structures are complicated. In the range of 128.5°E–129.75°E and
129.91°E–130.08°E, there are southward flows, which can reach a
depth of 200 m. The maximum velocity (-0.36 m s
-1
) appears at
128.83°E and 50 m depth. The flow is northward at both sides of the
Halmahera Sea section and in the lower layer. However, the
northward flow is relatively weak, with a maximum of 0.08 m s
-1
.
The large standard deviation of flow velocity is concentrated in the
large value region of southward flow, which can reach a maximum
of 0.14 m s
-1
in the upper layer of 128.83°E. With the increase of
depth, the standard deviation gradually decreases. At the maximum
southward flow (at 128.83°E and 50 m depth), the standard
deviation reaches 0.12 m s
-1
.
In the exit of the eastern equatorial Indian Ocean (Figure 4D),
the outflow is a significant eastward flow in the 0–110 m range of
114°E, 8.75°S–9.33°S, and the maximum velocity is 0.22 m s
-1
at a
depth of 60 m at 8.83°S. The eastward flow corresponds to the upper
ocean South Java Coastal Current (SJCC) (Atmadipoera et al., 2009;
Sprintall et al., 2009;Liang and Xue, 2020). In the middle section of
114°E, 9.5°S–13.41°S, most of the flow is westward with a maximum
of -0.28 m s
-1
, which exists at a depth of 10 m at 10.5°S. The
eastward flow south of 14°S is closely related with the Eastern Gyral
Current (EGC) and the Northwest Shelf Inflow (NWS-inflow)
(Liang and Xue, 2020). The large standard deviation values of
flow velocity are concentrated in the range of westward flow,
which can reach 0.06 m s
-1
in the upper layer.
The comparison results reveal a narrower and stronger inflow
and a wider and weaker outflow. This opposite flow among the
cross-sections implies complicated flow structures in the high-
resolution reanalysis datasets. Detailed flow structures along with
additional field experiments should be verified.
AB
D
C
FIGURE 4
The four reanalysis datasets average velocity in the inflow (upper panel) and outflow (lower panel) cross-sections from 1993 to 2019. The profile
color fill is the East-West (North-South) velocity of the flow perpendicular to the longitude (latitude). (A, D) are the zonal flow in the Sulawesi Sea
and the eastern equatorial Indian Ocean cross-sections, respectively. (B, C) are zonal flows in the Maluku Sea and the Halmahera Sea cross-sections,
respectively. The magenta contour is represented as the standard deviation of the flow velocity of each section on the interannual scale. The black
contour line in the figures indicate that the velocity is 0. A negative value represents southward or westward flow.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org07

To understand the interannual variations of the ITF transport,
the upper, lower and full depth layers of the inflow and outflow
transport anomaly are illustrated in Figure 5. The results are
averaged from the four reanalysis datasets. In the upper layer
(Figure 5A),thelineartrendsofinflow and outflow are -0.08 Sv
year
-1
and -0.04 Sv year
-1
, respectively (negative linear trend
indicatesanincreaseinsouthwesterlyflow). During the strong
El Niño event (1997/1998), there was a significant decrease in the
upper layer inflow, with a value of -3.38 ± 4.15 Sv. It applies to
2015/2016 as well, when the upper layer inflow is -4.90 ± 1.98 Sv.
During positive IOD events (2004–2008), the upper layer inflow
increased significantly, from -5.31 ± 2.76 Sv in 2004 to -14.31 ±
4.14 Sv in 2008. However, the upper layer outflow intensified only
during 2005–2006, which was significantly weaker than the inflow.
In the upper layer outflow, the interannual variation is similar to
those of the inflow, the average correlation coefficient between the
upper layer inflow and outflowofthefourmodelsis0.63,which
passes the 99% significancetest.Theaverageflow is more
intensified than the upper layer inflow, and the difference is 3.23
Sv. In the upper layer, it can be found that the inflow is smaller
than the outflow, which is mainly due to the reason that the South
China Sea (SCS) branch of the ITF is not taken into account. Note
that, the depth of the Karimata Strait that connects the SCS and
Indonesian Sea is less than 50 m, so its inflow mainly contributes
to the upper layer.
Unlike the upper layer, the lower layer flow shows no significant
trend in the interannual scale (Figure 5B). For the lower layer
inflow, the flow decreased in 2015 and increased in 2016, but with a
A
B
C
FIGURE 5
Interannual variability of the ITF transport in different layers. (A) The four reanalysis datasets average the upper layer inflow and outflow from 1993 to
2019. The red and blue lines represent inflow and outflow, respectively. The shaded areas indicate the standard deviation. (B, C) Same as (A) but for
the lower layer and full depth (0 to the maximum modeling depth), respectively. Negative values mean southward or westward transport anomalies
(enhanced ITF).
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org08

much smaller amplitude than that of the upper layer. The lower
layer outflow in the same period was also similar. In 1997, the lower
layer outflow showed a similar change as the upper layer. In the
lower layer, it can be found that the outflow is greater than the
inflow, which is closely related to the large inflow of HYCOM.
The total ITF transport is obtained in Figure 5C. The results
show that the average ITF transport is -13.33 Sv and -16.62 Sv of the
inflow and outflow, with linear strengthening trends of -0.06 Sv
year
-1
and -0.05 Sv year
-1
, respectively. On average, the outflow
value is larger than the inflow. The outflow volume transport
estimated in this study is -16.62 Sv, which is close to that by
SODA (-16.9 Sv) and the multi-model ensemble means of phase 5
of the Coupled Model Intercomparison Project (CMIP5) (-15.3 Sv)
(Santoso et al., 2022). In the Makassar Strait, the mooring average
flow of the upper (-9.10 Sv) and lower layers (-3.40 Sv) from 2004 to
2017 (Gordon et al., 2019) are consistent with the reanalysis
datasets, which is -9.36 Sv in the upper layer and -1.88 Sv in the
lower layer. The difference in the outflow and inflow (3.29 Sv) is
larger than in the Makassar Strait (1.26 Sv). Two possible reasons
contribute to the mismatch between the inflow and the outflow.
First, the SCS branch of the ITF and the inflow along the northern
Australia coast are not taken into account in the inflow and outflow
calculation, which results in small and large estimates of the inflow
and outflow, respectively. Results of the previous particle tracking
experiments present the interannual average of the ITF branch in
the South China Sea to be approximately 1.6–1.98 Sv (He et al.,
2015;Xu et al., 2021). Most of the reanalysis data show that the
inflow after adding Karimata Strait flow is comparable to the
outflow. And for the high spatial resolution data (CMEMS,
HYCOM), the inflow adding Karimata Strait flow is comparable
to the outflow after subtracting the inflow along the northern
Australia coast. Second, different reanalysis data show large
dispersions in the ITF estimation. The full-depth ITF obtained by
four kinds of reanalysis data (CMEMS, SODA, OFES, HYCOM) at
the inflow position is -14.22 Sv, -15.01 Sv, -6.44 Sv, and -16.94 Sv,
respectively. At the outflow position, they are -17.28 Sv, -15.20 Sv,
-8.95 Sv, and -24.29 Sv, respectively. The differences are 3.06 Sv,
0.19 Sv, 2.51 Sv, and 7.35 Sv, respectively. The largest deviation
between the outflow and the inflow is derived from the HYCOM
dataset. In addition, the selection of the outflow section also affects
the amount of transport. Overall, it is reasonable to use reanalysis
data in this study to reflect the interannual changes of ITF.
4 Contribution of ENSO and IOD to
the ITF transport
4.1 Relationships between climate modes
and ITF transport
Figure 6 illustrates the variabilities of ITF transport in different
layers that further explain the correlations between the climate
modes and the detailed flow structure. For the upper layer inflow,
the correlation coefficients between flow data and Niño 3.4, DMI,
and CP indices are 0.73, -0.28, and 0.56, respectively. Among the
above results, Niño 3.4 and CP pass the 99% significance test,
whereas DMI does not pass the 95% significance test. In the lower
layer inflow, the correlation coefficient between flow anomaly and
Niño 3.4, DMI and CP indices are 0.23, 0.01 and 0.12, respectively,
none of them pass the 95% significance test. This concludes that the
influence of the climate index in the upper layer is greater than that
in the lower layer.
In the upper layer outflow, the correlation coefficients between
flow data and Niño 3.4, DMI, and CP indices are 0.52, -0.55, and
0.32, respectively. Among the above results, Niño 3.4 and DMI pass
the 99% significance test, whereas CP does not pass the 90%
significance test. In the lower layer outflow, the correlation
coefficients of flow anomaly with Niño 3.4 and CP indices are
0.65 and 0.41, respectively, and they pass the 99% significance test,
whereas the correlation coefficient with the DMI shows a low
correlation (0.12). The correlation between Niño 3.4 and CP
indices, and the lower layer is greater than the correlation of the
upper layer inflow. Whereas, the correlation between the DMI
index and the lower layer with that of the upper layer inflow is
the opposite.
Niño 3.4 and DMI have opposite correlations for ITF in
comparison with the upper layer inflow and outflow. The linear
correlation coefficients between Niño 3.4 and ITF anomalies
remained high at 0.73 and 0.52 for the inflow and outflow,
respectively. However, the DMI is -0.28 and -0.55 in the inflow
and outflow, respectively, and the correlation changes are relatively
large. The conclusion is consistent with Li et al. (2020). In the upper
layer, the CP index maintains a positive correlation with both inflow
and outflow, which is similar to the impact of the Niño 3.4 index.
The difference with Li et al. (2020) is that the Niño 3.4 and CP
indices have a strong positive correlation with the lower layer
outflow, which may be related to the period selected.
4.2 Relative importance of climate modes
in the ITF variability
To quantitively expose the climate drivers of ITF transport, the
RF method is adopted in this subsection. Before that, the dominant
periods of the ITF transport should be clarified. The average multi-
models flow data in different levels (Figure 7) help to obtain the
power spectra of the upper layer and the lower layer inflow and
outflow during 1993–2019. Power spectral analysis reveals that the
upper layer and the lower layer inflow exhibit peak periods of 4–9
and 5–7 years, respectively. The upper layer and the lower layer
outflow exhibit peak periods of 5–7 and 2–6 years, respectively.
Overall, the results characterize a common peak period flow of 5–7
years at four different locations. Comparatively, Niño 3.4, as well as
DMI and CP indices have peak periods of 2–7, 2–4, and within 10
years (Sullivan et al., 2016;Santoso et al., 2022), respectively.
The calculation of the power spectrum of the four levels of flow
data and climate indices concludes that the main change cycle is
concentrated in 5–7 years; hence, it is taken as the cycle for RF
model training. In addition, to make full use of the data and the
experimental results more reliable, the data is recycled (consider the
6-year cycle as an example: 1993–1998, 1994–1999…). Niño and
IOD events often occur simultaneously during RF model training.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org09

AB
FIGURE 7
The power spectrum of inflow and outflow at different levels. The flow data is the multi-models average monthly data after a 13-month sliding
average from 1993 to 2019. (A) The black and blue solid lines represent the result of the upper layer (0–300 m) and the lower layer (300–760m)
inflow, respectively. The black and blue dotted lines represent their 95% confidence level. (B) same as (A) but for outflow.
A
B
D
C
FIGURE 6
Interannual variations of the ITF transport anomaly and the climate indices. (A) The four datasets average the upper layer inflow transport anomaly (purple
shadow), superimposed with the three indices (Niño 3.4-blue, DMI-orange, CP-green) after a 13‐month running mean. The DMI is the DMI new after
removing the linear influence of Niño 3.4. (B) Same as (A) but for the lower layer inflow. (C, D) The upper layer and the lower layer outflow transport.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org10

Therefore, the DMInew, which removes the influence of the linear
trend of Niño 3.4, is adopted and the corresponding RF models by
using four model data training are obtained. The relative
importance of the RF training results in the period of 5, 6, and 7
years, respectively, are considered. Significant differences are not
revealed in the results (figures not given). Therefore, the 6–year
period is considered to reveal the relative importance of different
climate indices at different levels, beginning with different starting
years (Figure 8). Due to the OFES data is up to 2017, the results for
the starting year 2013–2014 are obtained from CMEMS, HYCOM,
and SODA data. In this study, the dominant index is defined as the
importance higher than 33% and which exceeds the other two
indices without overlapping.
The RF results show that in the upper layer inflow, starting from
1993–1995 (end of 2008 and 2000), the importance of the Niño 3.4
was significantly higher than that of the other two indices which was
greater than 40%. In the 1996–2001 period (end of 2001 and 2006),
Niño 3.4 showed more important but insignificant results (shadow
regions cover each other) under the average training results of
multi-model data. In the starting year of 2002–2003 (end of 2007
and 2008), the DMI became the important index (without overlap
of the other two indices). In the periods of 2004–2012 (end of 2009
and 2017), the relative importance of the three indices were
comparative and the overlapping shadow covered each other. In
the periods of 2013–2014 (end of 2018 and 2019), the importance of
the DMI increased, whereas the importance of the Niño
3.4 decreased.
Different from the upper layer, in the lower layer inflow, in the
starting year of 1993 (end of 1998), the DMI was relatively
important. However, in the starting year of 1994 (end of 1999),
the importance of DMI decreased and CP was relatively important.
In the starting years of 1995–1999 (end of 2000 and 2004), the
shadows of the three indices overlapped each other. On average, the
CP was relatively important but insignificant. In the initial years of
2000–2007 (end of 2005 and 2012), the importance of DMI
gradually increased, and during 2002–2003 (end of 2007 and
2008), DMI had significant importance. In the later period,
starting from 2008–2014 (end of 2013 and 2019), the error bars
of the three indices overlapped, and again, there was no
dominant index.
In the upper layer outflow, the Niño 3.4 was always at a high
level during the 1993–1999 period (end of 1998 and 2004), which
implies the dominance of Niño 3.4. During 2000–2001 (end of 2005
and 2006), the influence of the DMI on the upper layer outflow
increased, but the importance of the Niño 3.4 decreased. In the
starting years of 2002–2004 (end of 2007 and 2009), the DMI
became the dominant factor, with an average relative importance of
more than 50% between 2002 and 2007. In the initial years of 2005–
2007 (end of 2010 and 2012), the average relative importance of the
CP increased, but the shadow region coincided with the shadow
region of the DMI. In the starting years 2008–2012 (end of 2013 and
2017), the importance of all three indices was relatively close. This
indicates the complexity of the influencing factors of ITF during this
period, and a significant index was not identified. In the following
starting years 2013–2014 (end of 2018 and 2019), the DMI became
the dominant factor again, with an average relative importance of
more than 40%.
In the lower layer outflow, starting from 1993–1995 (end of
1998 and 2000), the error bars of the three indices coincided, and
the significant index was not dominated. In the starting years of
1996–1998 (end of 2001 and 2003) and 2013–2014 (end of 2018 and
2019), the CP was the important index, with an average proportion
of 40%. In the starting years of 1999–2000 (end of 2004 and 2005)
and 2009–2012 (end of 2014 and 2017), there was no dominating
index. In the initial years of 2001–2005 (end of 2006 and 2010), the
DMI was the important influence index, with an average proportion
of 50%. In the starting years of 2006–2008 (end of 2011 and 2013),
the Niño 3.4 was the dominant influence index, with an average
proportion of 50% in 2007.
Figure 9 illustrates the relative importance of the different
climatic factor indices and their partial dependence in different
periods. In the partial dependence plots, the steeper the curve
change, the greater the influence of the corresponding index. The
results reveal that the curve of Niño 3.4 to be the steepest and the
influence on the upper layer inflow to be the greatest during 1993–
1998. In 2002–2007, the DMI curve was the steepest and became the
dominant index. Similarly, in 2008–2013, the CP and Niño 3.4 were
relatively important. During 2014–2019, the importance of DMI
increased. The results are consistent with those in Figure 8. The
partial dependences of the climate indices on the lower layer inflow,
as well as the upper layer and lower layer outflows are similar to the
upper layer inflow (figures not given).
5 Discussion
The RF model shows the domination of different climate modes
in different periods. To explore the underlying mechanisms, four
periods of 1993–2000, 2002–2008, 2009–2012, and 2013–2019 are
composited, which corresponds to the dominant factors of Niño
3.4, DMI, no significant dominant index, and DMI, respectively.
Among them, the dominant DMI from 2013 to 2019 mainly exists
in the upper layer outflow. Previous studies reveal that the sea
surface height anomaly (SSHA) between two oceans (NWP and
SEI) is regarded as a good indicator of ITF transport (Wyrtki, 1987;
Shilimkar et al., 2022). This viewpoint is adopted to demonstrate the
underlying mechanism.
Figure 10 presents the average of SSHA for different periods,
with the spatial location of NWP and SEI marked. During the
ENSO domination period (1993–2000), the mean SSH in the NWP
and SEI are -0.68 cm and -0.18 cm, respectively, and the SSH
difference is -0.50 cm. During the first IOD domination period
(2002–2008), the NWP and SEI are 0.49 cm and -0.50 cm,
respectively, and the SSH difference is 0.99 cm. During the period
with no dominant climate indices (2009–2012), the NWP and SEI
are 5.58 cm and 1.24 cm, respectively, and the SSH difference is
4.34 cm. During the second IOD domination period (2013–2019),
the NWP and SEI are -3.59 cm and -0.37 cm, respectively, and the
SSH difference is -3.22 cm. The SSH field variation in the two
regions is related to the ENSO and IOD events, especially the El
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org11

Niño and negative IOD events, respectively. The SSH differences
between the different periods are consistent with the mechanism,
that during El Niño (negative IOD) events, it reduces the Pac-
Indian Ocean pressure gradient and makes a weak ITF. On the
contrary, La Niña and positive IOD events help to enhance the ITF.
To further explore the mechanisms, the ENSO and IOD events
in the years 1993–2019 are summarized in Table 1. During 1993–
2000 when ENSO dominates, a strong El Niño event occurred in
1997/1998 accompanied by a positive IOD event (Figure 8A). This
strong El Niño event led to reduced precipitation in the western
Pacific and Indonesia, which directly led to a lower SSH at the ITF
inflow position (Chandra et al., 1998;Gordon et al., 1999).
Moreover, the occurrence of the positive IOD event reduced the
SSH at the ITF outflow location (Cai et al., 2011). Thus, both the
A
B
D
C
FIGURE 8
The importance of climate indices on the ITF transport. (A) The importance of climate indices (Niño3.4 blue, DMI red, CP green) in the upper layer
inflow with a 6–year cycle. The triangles represent the major ENSO and IOD events that occur in the corresponding starting year, where the upper
and lower edges represent ENSO and IOD events, respectively, and the solid and hollow indicate positive and negative anomalies, respectively. (B–
D) same as (A) but for the lower layer inflow, the upper layer outflow, and the lower layer outflow, respectively. The shaded areas indicate the
standard deviation. The dominant driver defined in this study is the one with the largest proportion and does not overlap with the other two indices.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org12

upper layer inflow and outflow of the ITF showed a strong
reduction of 4.21 Sv and 1.21 Sv, respectively (Figure 5A). After
experiencing a significantly weakened ITF in 1997/1998, the ITF
returned to its previous level in late 1998 under the combined effect
of La Niña and negative IOD events. Overall, it is inferred that the
occurrence of the El Niño event in 1997/1998 played a dominant
role in the 1993–2000 period.
During 2002–2008 when IOD dominates, two successive
positive IOD events occurred in 2006 and 2007, corresponding to
the El Niño and La Niña events, respectively (Figure 8A). The
positive IOD events are strongly associated with the Indian Ocean
surface temperature anomalies (unusually cold in the east and warm
in the west), further leading to a low SSH in the eastern Indian
Ocean (Behera et al., 2008). During 2004–2008, both the upper layer
inflow and outflow were enhanced 2.08–2.79 Sv (Figure 5A). During
2002–2008, the SSH differences between NWP and SEI was 0.99 cm,
which partly explains the dominant effect of the IOD events. This
concluded that the IOD events dominated the variability of the ITF
transport during 2002–2008.
During 2009–2012 when no climate indices dominate, the
successive La Niña events in 2010 and 2011 and the successive
positive IOD events in 2011 and 2012 (Figure 8A), which partly
explained that there were no single climate mode playing the
dominant role during this period.
During 2013–2019 when IOD dominates in the upper layer
outflow, it also corresponded to the sharp decrease in the upper
layer outflow during 2015–2016 (Figure 5A) and the SSH difference
between NWP and SEI was -3.22cm. During this period, relatively
strong negative IOD event occurred in 2015/2016 (Figure 8A).
Pujiana et al. (2019) also revealed that the sharp decrease in the
upper layer inflow and outflow in 2015/2016 was attributed to the
strong negative IOD events. In addition, successive positive IOD
events occurred from 2017 to 2019, which further illustrated the
dominant effect of IOD events.
6 Conclusions
In this study, the spatial and temporal variabilities of ITF are
obtained in the four high-resolution reanalysis datasets (three are
eddy-resolving and one is eddy-permitting models). The ITF is
further divided into the upper layer and the lower layer flow,
respectively. The results of spatial structure analysis reveal that
among the three inflow passages, the Sulawesi Sea is the main inflow
area in terms of inflow and the maximum westward flow velocity
reaches -0.59 m s
-1
. In terms of outflow, the range of 9.5°S–13.41°S
and 114°E in the eastern equatorial Indian Ocean cross-section is
the main outflow area, and the maximum westward flow velocity
AB
DC
FIGURE 9
The partial dependence of the upper layer inflow in different sub-period from the RF model on each climatic factor index. (A) 1993–1998, (B) 2002–
2007, (C) 2008–2013 and (D) 2014–2019. The blue, orange, and green lines are Niño 3.4, DMI and CP indices, respectively. The trend of the curve
describes the correlation between the corresponding factor and the predictor. The shaded part represents a 95% confidence interval.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org13

reaches -0.28 m s
-1
. The temporal analysis results reveal that the
southwest flow of the upper layer inflow and outflow have
interannual enhanced trends, which are -0.08 Sv year
-1
and -0.04
Sv year
-1
, respectively, whereas the lower layer inflow and outflow
do not have obvious linear trends on the interannual scale. The
interannual average of ITF full-depth flow is -14.97 Sv, whereas the
linear enhanced trend is -0.06 Sv year
-1
.
During 1993–2019, the linear correlation coefficients of Niño
3.4 index and flow change are greater than those of DMI index and
CP index at the upper layer and the lower layer inflow and the lower
layer outflow, while the upper layer outflow shows a strong
correlation with DMI index. To quantify the relative importance
of the three climate factors to ITF changes over time, RF models are
conducted at different levels. Through the analysis of the model
results, the dominant climate factors of the upper layer inflow and
outflow are clear in the three periods of 1993–2000, 2002–2008, and
2009–2012, which are dominated by Niño 3.4 (relative importance
reaching 40%), DMI (relative importance exceeding 50%), and no
significant dominant index. During 1993–2000, the El Niño event in
1997/1998 and the -0.5 cm SSH difference between NWP and SEI
dominated the Niño 3.4 index. During 2002–2008, the dominant
DMI was reinforced by successive positive IOD events in 2006–
2007 and the 0.99 cm SSH difference between NWP and SEI. In the
upper layer outflow, the dominant climate factor is clear in 2013-
2019 period, which are dominated by DMI (relative importance
reaching 40%). In the upper layer outflow, during 2013–2019,
TABLE 1 Classification of years when El Niño or La Niña is contacted with positive IOD or negative IOD events from 1993 to 2019.
Negative IOD No IOD event Positive IOD
El Niño 2002,2004,2009 1994,1997,2006,2015,
2019
No ENSO event 1996 1993,2001,2003,2008,
2013,2014 2012,2017
La Niña 1998,2010,2016 1995,1999,2000,2005 2007,2011,2018
A
B
D
C
FIGURE 10
SSHA in different periods. (A) 1993–2000, (B) 2002–2008, (C) 2009–2012, (D) 2013–2019. The ranges in the red and green boxes in (A) are NWP (6°
N–16°N, 125°E–155°E) and SEI (6°S–16°S, 85°E–115°E), respectively. The solid gray line is the contour line with SSHA of zero.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org14

relatively strong negative IOD event in 2015/2016, the -3.22cm SSH
difference between NWP and SEI, and successive positive IOD
events occurred from 2017 to 2019 dominated the DMI index.
The climate drivers can markedly regulate the ITF transport.
However, due to the complexity of the climate modes and the
interaction between them, it is difficult to clarify the relative
importance of each period (Wang et al., 2022). This study provides a
new insight to quantify the response of ITF transport to climate drivers.
However, the effect of climate factors on ITF in specificyearscouldnot
be analyzed in detail. The influence of these factors on ITF changes in a
specific climatic event can be studied with this method using identified
significantly important climate factors.
Data availability statement
The original contributions presented in the study are included
in the article/supplementary material. Further inquiries can be
directed to the corresponding author.
Author contributions
YZ convinced the work, and AL performed the data analysis
and wrote the original manuscript. YZ and AL improved the
manuscript. All authors discussed and contributed to the writing.
Funding
This work was financially supported by Laoshan Laboratory
(No. LSKJ202202704).
Acknowledgments
We sincerely thank Integrated Marine Observing System (IMOS)
for providing the mooring data in the Ombai and Timor Straits.
Conflict of interest
The authors declare that the research was conducted in the
absence of any commercial or financial relationships that could be
construed as a potential conflict of interest.
Publisher’s note
All claims expressed in this article are solely those of the authors
and do not necessarily represent those of their affiliated organizations,
or those of the publisher, the editors and the reviewers. Any product
that may be evaluated in this article, or claim that may be made by its
manufacturer, is not guaranteed or endorsed by the publisher.
References
Anderson, W. K., Thomas, J. L., and Van Leer, B. (1986). Comparison of finite
volume flux vector splittings for the Euler equations. AIAA J. 24 (9), 1453–1460.
doi: 10.2514/3.9465
Atmadipoera, A., Molcard, R., Madec, G., Wijffels, S., Sprintall, J., Koch-Larrouy, A.,
et al. (2009). Characteristics and variability of the Indonesian throughflow water at the
outflow straits. Deep Sea Res. Part I: Oceanographic Res. Papers 56 (11), 1942–1954.
doi: 10.1016/j.dsr.2009.06.004
Behera, S. K., Luo, J. J., and Yamagata, T. (2008). Unusual IOD event of 2007.
Geophysical Res. Lett. 35 (14). doi: 10.1029/2008GL034122
Cai, W., Sullivan, A., and Cowan, T. (2011). Interactions of ENSO, the IOD, and the
SAM in CMIP3 models. J. Clim. 24 (6), 1688–1704. doi: 10.1175/2010JCLI3744.1
Chandra, S., Ziemke, J., Min, W., and Read, W. (1998). Effects of 1997–1998 El nino
on tropospheric ozone and water vapor. Geophys. Res. Lett. 25 (20), 3867–3870.
doi: 10.1029/98GL02695
Chu, P. C. (1995). P-vector method for determining absolute velocity from
hydrographic data. Marine Tech. Soc. J. 29 (2), 3–14.
Du, Y., and Qu, T. (2010). Three inflow pathways of the Indonesian throughflow as
seen from the simple ocean data assimilation. Dyn. Atmos. Oceans 50 (2), 233–256.
doi: 10.1016/j.dynatmoce.2010.04.001
Feng, X., Liu, H., Wang, F., Yu, Y., and Yuan, D. (2013). Indonesian Throughflow in
an eddy-resolving ocean model. Chin. Sci. Bull. 58, 4504–4514. doi: 10.1007/s11434-
013-5988-7
Feng, M., Meyers, G., and Wijffels, S. (2001). Interannual upper ocean variability in
the tropical Indian ocean. Geophys. Res. Lett. 28 (21), 4151–4154. doi: 10.1029/
2001GL013475
Feng, P., Wang, B., Macadam, I., Taschetto, A. S., Abram, N. J., Luo, J.-J., et al.
(2022). Increasing dominance of Indian ocean variability impacts Australian wheat
yields. Nat. Food 3 (10), 862–870. doi: 10.1038/s43016-022-00613-9
Godfrey, J. (1996). The effect of the Indonesian throughflow on ocean circulation
and heat exchange with the atmosphere: a review. J. Geophys. Res. Oceans. 101 (C5),
12217–12237. doi: 10.1029/95JC03860
Gordon, A. L. (1986). Interocean exchange of thermocline water. J. Geophys. Res.
Oceans. 91 (C4), 5037–5046. doi: 10.1029/JC091iC04p05037
Gordon, A. L. (2005). Oceanography of the Indonesian seas and their throughflow.
Oceanography 18 (4), 14–27. doi: 10.5670/oceanog.2005.01
Gordon, A. L., Napitu, A., Huber, B. A., Gruenburg, L. K., Pujiana, K.,
Agustiadi, T., et al. (2019). Makassar strait throughflow seasonal and interannual
variability: an overview. J. Geophys. Res. Oceans 124 (6), 3724–3736. doi: 10.1029/
2018JC014502
Gordon, A. L., Sprintall, J., Van Aken, H. M., Susanto, D., Wijffels, S., Molcard, R.,
et al. (2010). The Indonesian throughflow during 2004–2006 as observed by the
INSTANT program. Dyn. Atmos. Oceans 50 (2), 115–128. doi: 10.1016/
j.dynatmoce.2009.12.002
Gordon, A. L., Susanto, R. D., and Ffield, A. (1999). Throughflow within makassar
strait. Geophys. Res. Lett. 26 (21), 3325–3328. doi: 10.1029/1999GL002340
Gordon, A., Susanto, R., Ffield, A., Huber, B., Pranowo, W., and Wirasantosa, S.
(2008). Maka ssar strait throughflow 2004 to 2006. Geophys. Res. Lett. 35 (24).
doi: 10.1029/2008GL036372
He, Z., Feng, M., Wang, D., and Slawinski, D. (2015). Contribution of the karimata
strait transport to the Indonesian throughflow as seen from a data assimilation model.
Cont. Shelf Res. 92, 16–22. doi: 10.1016/j.csr.2014.10.007
Heung, B., Bulmer, C. E., and Schmidt, M. G. (2014). Predictive soil parent material
mapping at a regional-scale: a random forest approach. Geoderma 214, 141–154.
doi: 10.1016/j.geoderma.2013.09.016
Hirst, A. C., and Godfrey, J. (1993). The role of Indonesian throughflow in a global
ocean GCM. J. Phys. Oceanogr. 23 (6), 1057–1086. doi: 10.1175/15200485(1993)
023<1057:TROITI>2.0.CO;2
Hu, S., and Sprintall, J. (2016). Interannual variability of the Indonesian
throughflow: the salinity effect. J. Geophys. Res. Oceans 121 (4), 2596–2615.
doi: 10.1002/2015JC011495
Hu, S., Zhang, Y., Feng, M., Du, Y., Sprintall, J., Wang, F., et al. (2019). Interannual to
decadal variability of upper-ocean salinity in the southern Indian ocean and the role of
the Indonesian throughflow. J. Clim. 32 (19), 6403–6421. doi: 10.1175/JCLI-D-19-
0056.1
IMOS. (2022). IMOS - moorings - hourly time-series product (ITFOMB, ITFTIN,
ITFTSL, and ITFTIS). Available at: https://imos.org.au/data/.
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org15

Lee, T., Fournier, S., Gordon, A. L., and Sprintall, J. (2019). Maritime continent water
cycle regulates low-latitude chokepoint of global ocean circulation. Nat. Commun. 10
(1), 2103. doi: 10.1038/s41467-019-10109-z
Li, M., Gordon, A. L., Gruenburg, L. K., Wei, J., and Yang, S. (2020). Interannual to
decadal response of the Indonesian throughflow vertical profile to indo-pacific forcing.
geophysical res. Letters 47 (11), e2020GL087679. doi: 10.1029/2020GL087679
Liang, L., and Xue, H. (2020). The reversal indian ocean waters. Geophysical Res.
Letters. 47 (14), e2020GL088269. doi: 10.1029/2020GL088269
Liu, Q. Y., Feng, M., Wang, D., and Wijffels, S. (2015). Interannual variability of the I
ndonesian T hroughflow transport: a revisit based on 30 year expendable
bathythermograph data. J. Geophys. Res. Oceans 120 (12), 8270–8282. doi: 10.1002/
2015JC011351
Masumoto, Y., Sasaki, H., Kagimoto, T., Komori, N., Ishida, A., Sasai, Y., et al.
(2004). A fifty-year eddy-resolving simulation of the world ocean: preliminary
outcomes of OFES (OGCM for the earth simulator). J. Earth Simu. 1, 35–56.
Meyers, G. (1996). Variation of Indonesian throughflow and the El niño-southern
oscillation. J. Geophys. Res. Oceans. 101 (C5), 12255–12263. doi: 10.1029/95JC03729
Murtugudde, R., Busalacchi, A. J., and Beauchamp, J. (1998). Seasonal-to-
interannual effects of the Indonesian throughflow on the tropical indo-pacific basin.
J. Geophys. Res. Oceans. 103 (C10), 21425–21441. doi: 10.1029/98JC02063
Pang, C., Nikurashin, M., Pena-Molino, B., and Sloyan, B. M. (2022). Remote energy
sources for mixing in the Indonesian seas. Nat. Commun. 13 (1), 6535. doi: 10.1038/
s41467-022-34046-6
Pujiana, K., McPhaden, M. J., Gordon, A. L., and Napitu, A. M. (2019).
Unprecedented response of Indonesian throughflow to anomalous indo-pacific
climatic forcing in 2016. J. Geophys. Res. Oceans 124 (6), 3737–3754. doi: 10.1029/
2018JC014574
Saji, N., Goswami, B. N., Vinayachandran, P., and Yamagata, T. (1999). A dipole
mode in the tropical Indian ocean. Nature 401 (6751), 360–363. doi: 10.1038/43854
Saji, N., and Yamagata, T. (2003). Possible impacts of Indian ocean dipole mode
events on global climate. Clim. Res. 25 (2), 151–169. doi: 10.3354/cr025151
Santoso,A.,England,M.H.,Kajtar,J.B.,andCai,W.(2022).Indonesian
Throughflow variability and linkage to ENSO and IOD in an ensemble of CMIP5
models. J. Clim. 35 (10), 3161–3178. doi: 10.1175/jcli-d-21-0485.1
Shilimkar, V., Abe, H., Roxy, M. K., and Tanimoto, Y. (2022). Projected future changes in
the contribution of indo-pacific sea surface height variability to the Indonesian throughflow.
J. Oceanogr. 78 (5), 337–352. doi: 10.1007/s10872-022-00641-w
Sprintall, J., Gordon, A. L., Koch-Larrouy, A., Lee, T., Potemra, J. T., Pujiana, K.,
et al. (2014). The Indonesian seas and their role in the coupled ocean–climate system.
Nat. Geosci. 7 (7), 487–492. doi: 10.1038/ngeo2188
Sprintall, J., Gordon, A. L., Wijffels, S. E., Feng, M., Hu, S., Koch-Larrouy, A., et al.
(2019). Detecting change in the Indonesian seas. Front. Mar. Sci. 6 257. doi: 10.3389/
fmars.2019.00257
Sprintall, J., and Revelard, A. (2014). The Indonesian throughflow response to indo-
pacific climate variability. J. Geophys. Res. Oceans 119 (2), 1161–1175. doi: 10.1002/
2013JC009533
Sprintall, J., Wijffels, S. E., Molcard, R., and Jaya, I. (2009). Direct estimates of the
Indonesian throughflow entering the Indian ocean: 2004–2006. J. Geophys. Res. Oceans.
114 (C7). doi: 10.1029/2008JC005257
Sullivan, A., Luo, J.-J., Hirst, A. C., Bi, D., Cai, W., and He, J. (2016). Robust
contribution of decadal anomalies to the frequency of central-pacific El niño. Sci. Rep. 6
(1), 38540. doi: 10.1038/srep38540
Susanto, R. D., Ffield, A., Gordon, A. L., and Adi, T. R. (2012). Variability of
Indonesian throughflow within makassar strait 2004–2009. J. Geophys. Res. Oceans. 117
(C9). doi: 10.1029/2012JC008096
Tillinger, D., and Gordon, A. L. (2009). Fifty years of the Indonesian throughflow. J.
Clim. 22 (23), 6342–6355. doi: 10.1175/2009JCLI2981.1
Vranes, K., Gordon, A. L., and Ffield, A. (2002). The heat transport of the Indonesian
throughflow and implications for the Indian ocean heat budget. Deep Sea Res. Part II:
Topical Stud. Oceanography 49 (7-8), 1391–1410. doi: 10.1016/S0967-0645(01)00150-3
Wang, J., Zhang, Z., Li, X., Wang, Z., Li, Y., Hao, J., et al. (2022). Moored
observations of the timor passage currents in the Indonesian seas. J. Geophys. Res.
Oceans. 127 (11), e2022JC018694. doi: 10.1029/2022JC018694
Wijffels, S. E., Meyers, G., and Godfrey, J. S. (2008). A 20-yr average of the
Indonesian throughflow: regional currents and the interbasin exchange. J. Phys.
Oceanogr. 38 (9), 1965–1978. doi: 10.1175/2008JPO3987.1
Wyrtki, K. (1987). Indonesian Through flow and the associated pressure gradient. J.
Geophys. Res. Oceans. 92 (C12), 12941–12946. doi: 10.1029/JC092iC12p12941
Xie, T., Newton, R., Schlosser, P., Du, C., and Dai, M. (2019). Long-term mean mass, heat
and nutrient flux through the Indonesian seas, based on the tritium inventory in the pacific
and Indian oceans. J. Geophys. Res. Oceans 124 (6), 3859–3875. doi: 10.1029/2018JC014863
Xu, T., Wei, Z., Susanto, R., Li, S., Wang, Y., Wang, Y., et al. (2021). Observed water
exchange between the south China Sea and Java Sea through karimata strait. J. Geophys.
Res. Oceans. 126 (2). doi: 10.1029/2020JC016608
Yuan, D., Wang, J., Xu, T., Xu, P., Hui, Z., Zhao, X., et al. (2011). Forcing of the
Indian ocean dipole on the interannual variations of the tropical pacific ocean: roles of
the Indonesian throughflow. J. Clim. 24 (14), 3593–3608. doi: 10.1175/2011JCLI3649.1
Yuan, D., Yin, X., Li, X., Corvianawatie, C., Wang, Z., Li, Y., et al. (2022). A maluku
Sea intermediate western boundary current connecting pacific ocean circulation to the
Indonesian throughflow. Nat. Commun. 13 (1), 2093. doi: 10.1038/s41467-022-29617-6
Yuan, D., Zhou, H., and Zhao, X. (2013). Interannual climate variability over the
tropical pacific ocean induced by the Indian ocean dipole through the Indonesian
throughflow. J. Clim. 26 (9), 2845–2861. doi: 10.1175/JCLI-D-12-00117.1
Zhang, T., Wang, W., Xie, Q., and Chen, L. (2019). Heat contribution of the
Indonesian throughflow to the Indian ocean. Acta Oceanol. Sin. 38 (4), 72–79.
doi: 10.1007/s13131-019-1414-6
Li et al. 10.3389/fmars.2023.1182255
Frontiers in Marine Science frontiersin.org16