Role of the Indian Ocean dynamics in the Indonesian Throughflow variability and extremes

Role of the Indian Ocean dynamics in the Indonesian Throughflow variability and extremes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Role of the Indian Ocean dynamics in the Indonesian Throughflow variability and extremes Yuanlong Li, Rui Li, Yilong Lyu, Janet Sprintall, Fan Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4745867/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The Indonesian throughflow (ITF) regulates heat and freshwater distributions of the Indo-Pacific Oceans and fundamentally affects the climate. The past decade has witnessed acute interannual variations in the Makassar Strait – the main ITF inflow passage, reaching monthly extremes of 1.9 Sv (1 Sv ≡ 10 6 m 3 s -1 ) in 2015 and 16.6 Sv in 2017, compared with a mean transport of ~12 Sv. The Pacific Ocean dynamics dictated by El Niño/Southern Oscillation (ENSO) cannot fully explain these variations and the role of the Indian Ocean (IO) dynamics remains uncertain. Here, we use a 0.1°, quasi-global ocean model to cleanly isolate the impact of the IO dynamics on the ITF. The wind-driven IO dynamics are found to play a significant role in either buffering or driving ITF variability. The buffering effect is commonly seen during strong ENSO events, while the driving effect arises from Indian Ocean dipole (IOD) events independent of ENSO. Notably, the IO dynamics buffered the weak ITF extreme of 2015 by ~35% and contributed to the strong ITF extreme of 2017 by ~23%. Our study aids in the prediction of regional climate extremes under the intensifying ENSO and IOD scenarios expected in the future. Earth and environmental sciences/Ocean sciences/Physical oceanography Earth and environmental sciences/Climate sciences/Ocean sciences/Physical oceanography Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction The Indonesian Throughflow (ITF) refers to a complex ocean current system flowing from the Pacific (PO) to the Indian Ocean (IO) through seas and channels within the Maritime Continent 1 - 3 . As a vital component of the global conveyor belt 1 , 4 , the ITF serves as the primary conduit for the inter-basin exchange between the PO and IO. In situ measurements in the Makassar Strait 5 – the primary inflow passage of the ITF with a mean volume transport of ~12 Sv 6 – have documented pronounced variability in the ITF strength during the past decade, such as an abrupt drop during 2015 and 2016 7 , 8 and a surge to nearly 20 Sv in 2017 5 . These changes, probably accompanied by property changes of the ITF water 9 , have led to dramatic heat and freshwater redistributions over the Indo-Pacific Oceans 8 , 10 - 14 and exert regional impacts on climate extremes 15 , 16 , sea level 13 , 17 , 18 , and biogeochemical cycles 19 . Yet, mechanisms governing the ITF variability and extremes have not been fully established, inhibiting accurate prediction of the ITF variability and its impacts. Year-to-year variability of the ITF has been linked to El Niño/Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD), with both modes causing oceanic wave adjustment in the Indo-Pacific Ocean circulations and thereby affecting the ITF 7 , 20 - 23 . Changes in Pacific trade winds associated with ENSO modulate the Indo-Pacific sea-level gradient and the ITF by evoking oceanic Rossby waves, rendering a weaker ITF in El Niño condition and a stronger ITF in La Niña conditions 24 - 26 . IOD events can generate equatorial Kelvin waves that propagate to the Sumatra‐Java coasts and then into the Indonesian Seas 24 , 27 , 28 , which also perturb the ITF strength 7 , 29 - 34 . Notably, a developing El Niño can frequently, though not always, trigger a positive IOD through atmospheric teleconnection 7 , 34 - 38 , which then drives upwelling Kelvin waves that act to alleviate the ITF weakening induced by the El Niño 7 , 25 , 34 , 39 . As such, the effect of the IO greatly complicates the relationship between ENSO and the ITF and brings difficulties to the prediction of the ITF variability. Clarifying the role of the IO in the ITF variability is challenging. Climates of the tropical Indo-Pacific Oceans are intrinsically coupled 38 , particularly in the vital interaction between ENSO and IOD 37 , 40 - 42 . The observed correlation of ~0.3-0.5 between the ITF and IOD stems largely from the impacts of ENSO on both 7 , 25 , 43 . As such, it is difficult to cleanly isolate the IO effect by analyzing short observational records. Moreover, resolutions of the Coupled Model Intercomparison Project Phase 6 (CMIP6) models are typically ~100 km and so do not resolve the narrow passages of the ITF. With unrealistic topography in the Maritime Continent, these models cannot faithfully simulate the ITF structure in climatology 22 , 26 . In comparison, high-resolution (e.g., ~10 km or finer) ocean model simulations achieve a more favorable representation of the ITF and wave propagation into the Indonesian Seas 28 , 44 - 48 . These issues highlight the need to reproduce the ITF transport and its changes consistent with observations through high-resolution model simulations and explore the underlying mechanisms. Encouraged by the success in reproducing observed ITF variations since the 1990s, here we use a quasi-global, eddy-resolving ocean model constrained by realistic topography in the Maritime Continent ( Methods ) to better understand the competing PO and IO mechanisms governing the ITF variability. Specifically, we carried out a series of regional forcing experiments to isolate the effect of wind-driven dynamical processes in the IO on the ITF. Our results reveal a diverse role of the IO in the ITF variability, emerging as a buffering effect for ~56% and a driving effect for ~44% of the time. Both effects are considerable in modulating the ITF transport and its extremes. The buffering effect tends to emerge during strong ENSO events, while the driving effect is usually exerted by IOD events that occur independent of ENSO. These findings will improve the predictability of the ITF variability and its broad impacts, such as regional climate extremes surrounding the Maritime Continent. Our work also highlights the inter-basin climate interaction between the Indo-Pacific Oceans through the ITF, which has potential implications for climate model simulations. Results Interannual variability and extremes of the ITF Here, we utilize a quasi-global (75°S-75°N; Fig. 1a ) hindcast of the Hybrid Coordinate Ocean Model (HYCOM) to simulate the ITF variability and understand the underlying processes ( Methods ). By adopting eddy-resolving horizontal resolutions of 0.1°×0.1° and applying topography correction for the Maritime Continent, the control run (CTRL) of HYCOM has produced the realistic flow pathways and structure of the ITF ( Fig. 1b, c ), as referenced to existing observation-based knowledge 5 , 6 , 26 , 27 . The simulated mean transport and standard deviation in CTRL agree with available mooring observations in key channels of the ITF (Makassar, Lombok, Ombai, and Timor Straits; Extended data table S1 ), particularly in the Makassar Strait (11.3 ± 3.71 Sv in HYCOM versus 11.1 ± 3.29 Sv in mooring observation). The simulated Makassar Strait transport achieves a correlation of r = 0.70 with the 13.3 years of mooring observations since 1997, which is significant at a 99% confidence level, and is also consistent with that produced by the Global Ocean Reanalysis and Simulations product (GLORYS12V1) reanalysis data ( r = 0.73) since 1993 ( Fig. 2a, b ). Mooring observations in the Makassar Strait have experienced pronounced ITF variability during the past decade. The volume transport abruptly dropped to ~2.0 Sv in the boreal winters (“boreal” omitted hereafter) of 2014-2015 and 2015-2016 and consequently rebounded to >16.0 Sv in the summers of 2016 and 2017 ( Fig. 2a ). These extremes were formed by the superimposition of interannual variations upon the seasonal cycle that features a stronger Makassar Strait transport in summer than in winter 48 - 50 . The notable impacts of these extremes on inter-basin heat and freshwater redistributions and the regional ecosystems have been documented previously 8 , 10 , 18 . The CTRL run of HYCOM has agreeably reproduced these variabilities and extremes, although the simulated extremes are slightly stronger in amplitude than in observations or GLORYS12V1. We define extremes when the CTRL transport exceeds the one standard deviation range (8.3-15.3 Sv; dashed lines in Fig. 2a ) for at least three months. During the 31-year period of 1992-2022, the ITF experienced 17 extremes (5 strong and 12 weak extremes), with 7 of them (1 strong and 6 weak extremes) occurring in the past decade (2013-2022). There were 29 months of extreme ITF transports in all during 2013-2022, whereas during the preceding decade of 2001-2010, there were only 13 months of extreme events. In December 2015, the simulated ITF transport was close to zero, indicating the Makassar Strait throughflow had nearly disappeared, and then subsequently strengthened to as large as 17.3 Sv (enhanced by 53%) in September 2017. After the removal of the seasonal cycle, interannual variations stand out with a standard deviation of 2.5 Sv and anomalies up to ~6.0 Sv ( Fig. 2b ), confirming the key role of interannual variability in the occurrence of extremes. To explore the dynamical processes underlying the ITF variability and extremes, sensitivity experiments of HYCOM were carried out, with WND, PWND and IWND experiments representing the dynamical response of the ITF to global, tropical PO and tropical IO wind variabilities, respectively ( Methods ). Existing studies have suggested that the salinity and temperature components account for ~30% and ~70% of the interannual variability of the ITF geostrophic transport, respectively 43 , 51 and emphasized the impact of precipitation-induced salinity changes 52 , 53 . It is important to recognize that both the salinity and temperature components are modulated by wind stress to some extent. Wind stress, through its influence on large-scale circulation, dictates the advection and leads to salinity variations of the southeastern IO 28 , 54 , 55 . Meanwhile, the heaving process driven by Ekman pumping 56 dominates upper-ocean temperature variability in the Indian Ocean 57 . Hence, it is not surprising to see the dominance of wind forcing in the ITF variability in our HYCOM experiments ( Fig. 2c ). The ITF transport anomaly in WND ( ITF WND ) greatly resembles that of the CTRL ( ITF CTRL ), and their correlation coefficient is up to 0.95. Wijffels and Meyers (2004) found that about 60–90% of sea level variability within the Indonesian Seas and southeast IO can be explained by free Kelvin and Rossby waves generated by remote zonal winds along the equator of the Indian and Pacific Oceans. In addition, certain studies have emphasized the role of local winds in the Maritime Continent 58 , 59 , though their primary significance lies within the realm of seasonal and higher-frequency timescales 53 . Here, we use PWND and IWND to isolate the roles of PO and IO wind forcing, respectively ( Methods; Extended Data Fig. S1 ). The results confirm that the remote wind forcing from the equatorial PO and IO largely accounts for the ITF variability, as evidenced by a correlation of 0.85 between ITF PWND + ITF IWND (the sum of ITF transport anomalies in PWND and IWND) and ITF WND . The missing variance suggests the possible role of local winds, but that is beyond the scope of our study. ITF PWND surpasses ITF CTRL in magnitude in most cases ( Fig. 2c, d ), indicating an overall counteracting effect of the IO dynamics. Despite an overall stronger effect of the PO winds, the IO’s influence cannot be disregarded 25 , 31 , and it can even assume a dominant role during certain periods 7 , 9 . In addition to the overall transport, our model also characterizes the vertical structure of the ITF’s variability in the Makassar Strait ( Extended Data Fig. S2 ). Both observations 5 and numerical models 23 , 31 have suggested that the ITF may exhibit opposing transport anomalies between the upper and the deeper layers, which are linked to different dynamic processes 23 , 31 , 60 . This is also seen in our HYCOM simulations which show a clear baroclinic structure of the ITF anomalies and distinct features in PWND and IWND ( Extended Data Fig. S2 ). Role of the IO dynamics Next, we attempt to understand the role of IO dynamics in modulating the ITF variability. This is first pursued by looking at two extreme cases of the Makassar Strait transport ( Fig. 3 ). In December 2015, the ITF transport in CTRL was drastically weakened to <1.0 Sv in the Makassar Strait ( Fig. 3a ). In PWND, the weakening of the ITF is even more striking. Owing to the strong PO influence associated with the 2015-2016 super El Niño, the ITF almost disappeared in the Makassar Strait and near the outlet regions ( Fig. 3c ). By contrast, in IWND, the southward flow in the Makassar Strait was stronger than normal in December 2015 (by 16% above average in volume transport), indicating a buffering effect of the IO winds for the weakening of the ITF ( Fig. 3e ). In September 2017, the ITF was strengthened in both PWND and IWND ( Fig. 3d, f ). In this case, the IO and PO dynamics thus act mutually to drive an enhancement of the ITF, in which the contributions of the IO and PO were 52% and 48%, respectively. The analysis of extremes presented above indicates that the IO dynamics may either buffer or drive the ITF variability. The relationship between the IO and PO effects can be visualized by a scatter plot of ITF PWND versus ITF IWND ( Fig. 4a ). Most monthly data points fall in the quadrants 2 and 4, indicating the opposite effects of the PO and IO on the ITF change. When ITF IWND is weaker than ITF PWND in magnitude, the IO acts to “buffer” the ITF change dominated by the PO effect. This situation accounts for 56.3% of the time during 2014-2022, indicated by red areas in Fig. 4a . Notably, there are many weak extremes (10 months) located in the quadrant 2 with a stronger negative ITF PWND and a weaker positive ITF IWND . They mainly represent the buffering effect of the IO dynamics on the weakened ITF during strong El Niño events, as in the 2015-2016 winter ( Fig. 3a-c ). By contrast, there are only three months of strong ITF extreme in quadrant 4. In that case, the PO enhanced the ITF transport owing to the La Niña event, which was buffered by the IO. On average, the buffering effect of the IO on the ITF change is 1.4 Sv, as quantified by the standard deviation of ITF IWND in the buffering situation. This accounts for ~41% of the corresponding ITF change driven by the PO (3.4 Sv as the standard deviation of ITF PWND ). When the IO’s buffering effect operates, the total ITF anomaly ( ITF CTRL ) is 1.8 Sv in standard deviation and significantly weaker than ITF PWND . When ITF IWND is stronger in magnitude than the opposing ITF PWND (in quadrants 2 and 4) or of the same sign as ITF PWND (in quadrants 1 and 3), the IO contributes constructively to the total ITF change, which is regarded as a “driving” effect. This situation constitutes 43.6% of the time during 2014-2022, marked as blue areas in Fig. 4a . In this situation, the standard deviation of ITF IWND is 1.1 Sv, weaker than the corresponding ITF PWND (2.2 Sv in standard deviation). The standard deviation of ITF CTRL is 1.6 Sv, which is smaller than the sum of the two effects above. This is because ITF PWND does not always have the same sign as ITF IWND in driving cases. Interestingly, the joint strengthening of the ITF by the PO and IO (quadrant 1) occurs most frequently – a dominant regime for the occurrence of strong ITF extremes. By contrast, a joint weakening (quadrant 3) is seldom observed. This is linked to the complexity of the relationship between ENSO and IOD, which will be discussed in the following subsection. On interannual timescales, changes in the ITF strength are mainly controlled by the interbasin sea-level gradient between the PO and IO 2 , which is in turn perturbed by wind-forced planetary waves 24 . The diverse effects of the IO dynamics can also be understood in this framework. A composite analysis suggests that the driving effect on the ITF is linked to sea-level anomalies (SLAs) in the eastern tropical IO ( Fig. 4b ) that alter the inter-basin sea-level gradient and the ITF strength. These SLAs near the exit region of the ITF are established by equatorial winds through eastward-propagating Kelvin waves 7 , 61 . A buffering scenario is usually linked to a pattern of in-phase SLAs in the entrance and exit regions (the western PO and the eastern IO, respectively) that are driven by opposing equatorial wind anomalies in the two basins. In Fig. 4c for example, the positive SLAs on the IO side act to dampen the enhancements of the sea-level gradient and the ITF induced by the stronger positive SLAs in the PO side. The relationship between the ITF with ENSO and IOD events Next, we explore what determines the buffering or driving role of the IO dynamics. The above analysis has highlighted the key role of SLA in the ITF variability arising from IO and PO. Various SLA-based proxies for the ITF strength have been proposed by existing studies to understand how ENSO and IOD give rise to ITF changes 30 , 45 , 62 , 63 . In these studies, SLAs in key regions on the IO and PO sides, usually identified by the high correlations between the ITF transport and SLA, are used to construct the inter-basin SLA gradient to represent the ITF strength. In this study, this approach is applied separately to PWND and IWND experimental output ( Methods ). We identified the key regions with maximum correlation coefficients ( Fig. 5a, b ). In PWND, a rectangle region of 130°E-160°E, 5°N-15°N stands out with a maximum correlation of >0.7, whereas in IWND, a diamond-shape region surrounding the Sumatra-Java island chain is identified as the key region ( r > 0.5) on the IO side of the ITF ( Fig. 5b ). Note that the key region on the IO side differs from all existing studies that use the total ITF anomaly to seek key regions 30 , 63 , 64 . This can be understood by its correlations with the ITF in different experiments ( Extended Data Fig. S3 ): In IWND, the SLA in the IO key region shows a high correlation with ITF IWND , but in CTRL, mimicking the observed ocean, its correlation with ITF CTRL is minimal because of the dominance of the PO dynamics in the ITF variability. As such, it is difficult to identify the “true” key region on the IO side through correlation analysis of observational data. By contrast, the key region on the PO side shows a high correlation with the ITF in both PWND and CTRL ( Extended Data Fig. S3 ). Using the SLAs of the two key regions, we construct proxies of ITF PWND and ITF IWND through least-square fitting in PWND and IWND, respectively. Further, by applying SLAs of CTRL to the proxy algorithm, we obtain proxies of ITF PWND and ITF IWND for the entire 1979-2022 period, denoted as ITF P and ITF I , respectively ( Fig. 5c ). Note that altimetric SLAs since 1992 can also be used in this algorithm to generate observation-based proxies, which are consistent with the model-based proxies ( Extended Data Fig. S4 ). The sum of ITF P and ITF I compares favorably well with ITF CTRL during 1979-2022, showing a correlation of 0.82 ( Fig. 5d ). This indicates that the proxies have captured the primary mechanisms governing the interannual variability of the ITF. Given that the ENSO and IOD are the two primary origins of interannual wind variabilities over the equatorial Indo-Pacific Oceans, we hypothesize that the complexity of the ENSO-IOD relationship is deterministic in the diverse role of IO dynamics. Compared to ITF IWND , the lengthened ITF I allows us to explore the diverse role of IO dynamics under the impacts of ENSO and IOD events more robustly. Among the 525 months of 1979-2022, there were 211 months with significant IO-driven ITF changes (defined as exceeding the ±0.4 standard deviation: ±0.4 Sv for ITF I and ±0.7 Sv for ITF I + ITF P ) and used for our analysis. The results suggest that the IO’s effect on the ITF, represented by ITF I , is modulated by both ENSO and IOD ( Fig. 6a ). Generally, El Niño and positive IOD (pIOD) conditions render a positive ITF I (the IO enhancing of the ITF), while La Niña and negative IOD (nIOD) conditions favor a negative ITF I . It is interesting to note that in quadrant 4, even in nIOD condition, there are positive ITF I values dictated by the El Niño. This clearly points to the essence of ENSO, in addition to the IOD, in driving the IO dynamics. A buffering effect of the IO mainly operates in two situations. The first is during the co-occurrence of in-phase ENSO and IOD conditions, e.g., an El Niño accompanied by a pIOD (pIOD) (or a La Nina plus a nIOD), which accounts for 35.0% of all cases. The positive and negative scenarios are nearly symmetric in occurrence frequency, 17.0% versus 18.0% of all cases (in quadrants 1 and 3 of Fig. 6a ), respectively. The composites show that westerly winds in the equatorial PO associated with El Niño attenuate the ITF resulting in a strong sea-level decrease on the PO side, while the easterly winds of the co-occurring pIOD act to enhance the ITF by causing weaker sea-level falling on the IO side (Fig. 6c) . The stronger impact of the El Nino results in a IO buffering effect that counteracts the weakening of the ITF. This scenario was commonly observed during strong El Niño events that are more apt to trigger the pIOD, such as 2015-2016 one ( Fig. 2a-c ). The negative scenario of La Niña plus nIOD leads to a similar pattern of the opposite sign ( Fig. 6f ). The other situation is the ENSO events occurring in a neutral condition of IOD ( Fig. 6d, g ). This scenario was also commonly observed, accounting for 22.5% of all cases (8.5% for El Niño plus 14.0% for La Niña; Fig. 6a ). An ENSO event gives rise to wind anomalies in the eastern equatorial IO through atmospheric teleconnections 35 , 40 , causing SLAs on the IO side ( Fig. 6d, g ) and buffers the ENSO-induced ITF change. A driving effect of the IO arises primarily from IOD events that occur “independent” of ENSO, which makes up 24.4% of all cases. Westerly equatorial winds in the IO associated with an independent nIOD event attenuate the ITF by evoking downwelling Kelvin waves ( Fig. 6e ), whereas a pIOD involves easterly winds and drives a strengthening of the ITF ( Fig. 6b ). Note that the independent nIOD is more frequently observed than the independent pIOD (18.0% versus 6.4% in all cases, respectively). Another situation for the driving effect is for out-of-phase ENSO and IOD events, representing 7.6% of all cases, including La Niña plus pIOD (4.2%) and El Niño plus nIOD (3.4%). In this situation, ENSO and IOD operate mutually to drive the ITF change, with the strong ITF extreme in 2017 summer as an example. This situation has much fewer samples than the in-phase scenarios (35.0%), reflecting the overall positive ENSO-IOD correlation, and therefore their composites are weak and insignificant (not shown). This also determines that the buffering effect of the Indian Ocean prevails over its driving effect. We should state that the four situations described above are the main rather than the whole story. They add up to 89.5% of all cases, with the remaining 10.5% representing neutral conditions of both ENSO and IOD. The four situations also represent the common features of all samples, not the exceptional cases. For example, when a strong pIOD occurs with a weak El Nino, the IO effect may dominate the total ITF change, making up a driving rather than buffering effect (such as November 2019). Summary and implications Mooring measurements in the Makassar Strait have documented pronounced variability in the ITF strength during the past decade of 2013–2022, showing extremes of 1.9 Sv in December 2015 and 16.6 Sv in September 2017. While the PO dynamics associated with ENSO cannot fully explain these changes, the role of IO dynamics remains largely uncertain. Here, with a novel attempt based on a series of high-resolution HYCOM experiments, we quantitatively reveal a diverse role of the IO dynamics in regulating the interannual variability and extremes of the ITF. The IO dynamics, primarily Kelvin waves driven by equatorial IO winds, can either buffer or drive the ITF variability, with both effects contributing to ITF extremes. The buffering effect takes place more frequently than the driving effect, 56% versus 44% in operating time, respectively. This diversity of the IO’s role stems from the complexity of the ENSO-IOD relationships. The IO tends to buffer the PO-driven ITF changes when ENSO events occur alone or in phase with IOD events, while its driving effect arises primarily from IOD events that occur independent of ENSO. Our finding underpins the essence of the IO on the ITF variability. The IO dynamics act as the primary agent for the IO climate to affect the ITF, providing implications for the prediction of the ITF and its impacts on surrounding regions. Modeling studies have shown that the ITF can affect the ENSO dynamics 65,66 . By causing changes in the ITF transport, the IOD affects the warm water volume of the western Pacific and thereby modulates the evolution of ENSO in the following year 33,67,68 . This study further reveals that the IO dynamics that cause the ITF changes can arise from not only the IOD but also the ENSO’s teleconnection. This likely implies that the ITF and IO are vitally involved in the “self-regulating” regime of the ENSO cycle. For example, during El Niño conditions, the upwelling Kelvin waves in the IO, arising from either a pIOD or the El Niño’s teleconnection, buffer the weakened ITF and thereby contribute to the discharge of the tropical PO. This process is favorable for the quick demise of the El Niño. Alternatively, an nIOD acts to further weaken the ITF and allow heat to further accumulate in the PO 8 – a process that prolongs the El Niño. In this sense, the diverse role of the IO dynamics serves as a potential source of ENSO complexity 69 . The occurrence of extreme ENSO and IOD events is projected to increase in a warming climate 70,71 , which makes us wonder whether the ITF variability will amplify. For the first time, our work calls attention to the increasingly observed ITF extremes and reveals the underlying dynamical complexity. Changes in the ITF strength substantially perturb the heat budget of the Maritime Continent 72 and the southeast IO 10–12 . These ITF extremes energize regional climate extremes such as marine heatwaves, amplifying the stress of climate change on the vulnerable marine ecosystems in these regions 73,74 . This calls for an urgent investigation into whether the ITF extremes will increase and lead to more regional climate extremes in surrounding regions. Meanwhile, climate models consistently project a reduction up to 3.4 Sv in the ITF transport in response to future greenhouse warming, which was attributed to the suppressed deep-layer upwelling in the PO 22,75 . Our work raises interesting questions regarding the influence of the IO dynamics, particularly those in response to the altering South Asian monsoon 76,77 , on the ITF centennial changes. It remains unclear how the weakening ITF with amplifying variability affects the ENSO and regional climate. However, at present, accurate simulation of the ITF remains a challenging task for climate models owing to unrealistic terrains and flawed parameterization schemes 23 . As the only oceanic conduit between the tropical Indo-Pacific Oceans, errors in the simulated ITF may propagate into the simulated ENSO and IOD. There are long-standing biases of the simulated ENSO and IOD in climate models 78,79,80 . For example, the strong IOD amplitude bias throughout successive generations of models is attributed to an overly active Bjerknes feedback in the southeastern tropical IO 81 . The successful simulation and mechanistic understanding of the ITF variability may excite more extensive investigations of the ITF and its simulation in climate models. This shall advance our understanding of inter-basin climate interaction significantly and shed light on the pathway forward for improving climate models. Methods Observed ITF volume transport in the Makassar Strait. There are a total of ~ 13.3 years of Acoustic Doppler Current Profilers (ADCP) measurements within the Labani Channel constriction (sill depth ~ 680 m) of the Makassar Strait, consisting of November 1996-July 1998 by the Arlindo program, January 2004-November 2006 by the INSTANT program (International Nusantara Stratification and Transport Program), and November 2006-August 2017 by MITF (Monitoring the ITF) program 5 . Two moorings (western and eastern) were deployed during Arlindo and INSTANT 6,49 , while the eastern mooring was not redeployed during MITF. To calculate the ITF volume transport, we use the along strait velocities 5 (ASVs) at the western mooring to represent the Makassar Strait throughflow across the Labani Channel. The downstream direction along the Labani Channel axis is 170° (referenced to true north). ASV is parallel to the Labani Channel axis. In practice, we assumed the velocity adjacent to the sidewalls as zero and applied linear interpolation from the western mooring (2°51.9′ S, 118°27.3′ E) to the sidewalls. Then, we integrated the ASVs across the section to obtain the volume transport 26 . Observation-based reanalysis datasets. The monthly ocean current data of the Global Ocean Reanalysis and Simulations product (GLORYS12V) 82 and sea level satellite altimeter data from the Copernicus Marine Environment Monitoring Service (CMEMS) 83 are analyzed. GLORYS12V1 was designed and implemented using the current real-time global forecasting CMEMS (Copernicus Marine Environment Monitoring Service) system and driven by the NEMO3.1 ocean/sea-ice general circulation model, covering the 1993–2020 period and with a 1/12° horizontal resolution and 50 vertical levels. For GLORYS12V1, the ITF transport (Sv) in the Makassar Strait is computed as the integration of ASVs in the upper 700 m at 3°S, $$\:\text{I}\text{T}\text{F}=-{\int\:}_{700\:m}^{0\:m}{\int\:}_{{x}_{W}}^{{x}_{E}}\text{A}\text{S}\text{V}(x,z,t)dxdz$$ 1 , where x E = 116.3°E and x W = 118.8°E are the longitudes of the western and eastern boundaries of the strait, x , z , and t are longitude, depth, and time, respectively. Positive ITF transport indicates the flow from the PO toward IO. We also analyzed the 0.25° surface wind data of the fifth generation of ECMWF reanalysis (ERA5) 84 . The Niño-3.4 index, defined as the average SST anomaly over 170°W–120°W, 5°S–5°N, and the Dipole Mode Index (DMI), defined as the SST anomaly difference between 50°E − 70°E, 10°S–10°N and 90°E − 110°E, 10°S–0°, are downloaded from the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center. HYCOM and regional forcing experiments. Here, we utilize the Hybrid Coordinate Ocean Model (HYCOM) version 2.3.01 to simulate the ITF variability and achieve insights into underlying mechanisms through regional forcing experiments. HYCOM combines isopycnal, sigma (terrain following), and z -level coordinates to optimize the representation of oceanic processe 85 and has been successfully utilized to simulate the ITF 44,45 . In this study, the HYCOM is configured to a quasi-global domain of 75°S-75°N with 0.1°×0.1° horizontal resolutions and 50 vertical layers. The layer thickness gradually enlarges from 3 m at the surface to about 500 m in the abyssal ocean. The model topography is based on Earth's Topography and Bathymetry (ETOPO1) which is interpolated onto the model grid and smoothed with a 1° x 1° window throughout the global ocean to remove steep features. Then, the topography of the Maritime Continent is tuned toward the original ETOPO1 data by extensive manual editing, which is necessary to ensure relatively realistic flow pathways of the ITF (Fig. 1 b) since the ETOPO1 data better depicts the terrain and flow in the Indonesian seas than the smoothed ETOPO5 data. There are sponge layers of 5° placed in the southern and northern boundaries, where the simulated temperature and salinity are relaxed to the monthly climatology of World Ocean Atlas 2013 86,87 . The model adopts hourly fields of 10-m wind speed, wind stress, surface shortwave and longwave radiations, precipitation rate, 2-m air temperature and humidity, and river discharge of ERA5 as surface forcing. The model is spun up for 20 years under repeated hourly atmospheric forcing of 1979. Restarting from the already spun-up solution, HYCOM is integrated forward using hourly mean ERA5 forcing fields from January 1979 to September 2022 to form the control simulation (CTRL). The CTRL is used as the reference and compared with observations to evaluate the model performance. The ITF transport in HYCOM is computed in the same manner as for GLORYS12V1. Given that 2013 was a neutral year for both ENSO and IOD ( Extended Data Fig. S5 ) and the ITF was also close to its mean state (Fig. 2 b), we use 2013 as the baseline state in the sensitivity experiments for the recent decade. The wind experiment (WND) is forced with original hourly wind stress (as in CTRL) from January 2014 through September 2022 along with repeated 2013 hourly fields for all the other forcing factors such as wind speed, radiation, precipitation, air temperature, and humidity. In our HYCOM configuration, wind stress controls all the ocean dynamical processes (circulation, waves, and mixing), while wind speed affects the evaporation rate and surface latent and sensible heat fluxes 11,12,88 . As such, WND represents the dynamical response of the ITF to wind variabilities of both the PO and IO. In the Pacific wind experiment (PWND), original hourly wind stress is only retained in the tropical Pacific Ocean (140°E-90°W, 25°S-25°N; Extended Data Fig. S1 ), while wind stress in other regions and other forcing fields are all fixed to repeated 2013 fields. Similarly, the Indian wind experiment (IWND) adopts the 2014–2022 hourly wind stress in the tropical Indian Ocean (45°E-115°E, 20°S-20°N). To avoid abrupt changes in wind forcing, in both PWND and IWND, we apply a 5° transition belt surrounding the region, where winds alter gradually from the 2014–2022 fields to the repeated 2013 fields. As such, the ITF transport anomalies (with the monthly climatology removed) in the PWND and IWND, written as ITF PWND and ITF IWND , represent the effects of wind-driven dynamical processes in the PO and IO, respectively. Construction of the SLA-based ITF proxy. To explore the impacts of ENSO and IOD on the ITF, it is instructive to obtain longer records of ITF PWND and ITF IWND than only the 2014–2022 period. The sea level anomaly (SLA) gradient between the western PO and the eastern IO has been proposed as a useful proxy for the ITF strength 30,62,63 . Regions of high correlations with the ITF transport 63 are usually adopted as the key regions for the SLA proxies. However, the correlation between SLA and the ITF in observations may be dictated by other processes and does not necessarily reflect a true dynamical linkage between each other. In particular, due to the strong influence of the PO SLA on the total ITF anomaly, the correlation with the total ITF anomaly on the IO side is largely determined by the correlation with the PO SLA ( Extended Data Fig. S3a ). Our regional forcing experiments with HYCOM, i.e., PWND and IWND, serve as useful tools to distinguish the influences of the PO and IO and accurately detect the regions suitable for SLA proxies. Practically, we calculated the SLA-ITF correlation of 2014–2022 separately in the PWND and IWND. The region with the maximum correlation (~ 0.7 or higher) in PWND, i.e., 130°E-160°E, 5°N-15°N, is selected as the key region on the PO side (Fig. 5 a); similarly, the region with the maximum correlation in IWND (~ 0.5 or higher), a diamond-shape region enveloping the Sumatra-Java island chain, is defined as the key region on the IO side (Fig. 5 b). Note that the key region on the IO side differs from all existing studies that use the total ITF anomaly to seek key regions 30,63,64 . This indicates that SLAs in this region are dynamically linked to the ITF, because the influence from the PO has been precluded in IWND. Then, the proxy for the total ITF anomaly ( ITF total ) can be calculated as the sum of the proxy ITF anomalies driven by the PO and IO dynamics, ITF P and ITF I , $$\:{ITF}_{\:\text{t}\text{o}\text{t}\text{a}\text{l}}={ITF}_{\text{P}}+{ITF}_{\text{I}}=\alpha\:\:SL{A}_{\text{P}}+\beta\:\:SL{A}_{\text{I}}$$ 2 , where SLA P and SLA I are the average SLAs in the key regions on the PO and IO sides, respectively, α = 0.27 Sv cm − 1 and β = -0.37 Sv cm − 1 are coefficients obtained through linear least-square fitting using PWND and IWND results, respectively. The correlation between the ITF P proxy and ITF PWND is 0.94, while that between ITF I and ITF IWND is 0.81 (Fig. 5 c). Then, the total ITF anomaly proxy of 1979–2022 can be obtained by substituting SLA P and SLA I of CTRL into the above equation. Declarations Data availability All observation, reanalysis and model data that support the findings of this study are available as follows. The INSTANT mooring data: http://www.marine.csiro.au/~cow074/data/instantdata_download.html; MITF mooring data: http://ocp.ldeo.columbia.edu/res/div/ocp/projects/MITF/cm_data/; the GLORYS12V1 reanalysis: https://resources.marine.copernicus.eu/productdetail/GLOBAL_MULTIYEAR_PHY_001_030; ERA5 data are available at Complete ERA5 global atmospheric reanalysis (copernicus.eu); HYCOM simulation: http://msdc.qdio.ac.cn ; Niño-3.4 index: https://www.cpc.ncep.noaa.gov/data/indices/; DMI: https://psl.noaa.gov/gcos_wgsp/Timeseries/DMI/; Satellite altimeter data of the Copernicus Marine Environment Monitoring Service (CMEMS) are obtained from https://resources.marine.copernicus.eu/products. Code availability: MATLAB codes for data analysis and graphing are available upon request. The HYCOM version 2.3.01 source code is available at https://github.com/HYCOM/HYCOM-src/releases . Acknowledgements: This research is jointly supported by the National Key R&D Program of China (2019YFA0606702), the Strategic Priority Research Program of Chinese Academy of Sciences (XDB42000000), and the Laoshan Laboratory (LSKJ202202601). Janet Sprintall acknowledges funding support from the US National Science Foundation award OCE-1851316. Author contributions: Y.Li and F.W. designed the study. R.L. performed the analysis. R.L. and Y.Li drafted the paper. Y.Lyu and R.L. conducted and evaluated the model experiments. J.S. provided valuable feedback and suggestions to enhance the quality of this article. All the authors contributed to the interpretation of the results and refinement of the manuscript. Competing interests: The authors declare no competing interests. Correspondence and requests for materials should be addressed to Y.Li and F.W. References Gordon, A. L. Interocean exchange of thermocline water. J. Geophys. Res. 91, 5037–5046 (1986). Wyrtki, K. Indonesian through flow and the associated pressure gradient. J. Geophys. Res. 92, 12941–12946 (1987). Gordon, A. Oceanography of the Indonesian Seas and Their Throughflow. oceanog 18, 14–27 (2005). Broecker, W. S. The great ocean conveyor. AIP Conf. Proc., 247, 129–161(1992). Gordon, A. L. et al. Makassar Strait Throughflow Seasonal and Interannual Variability: An Overview. JGR Oceans 124, 3724–3736 (2019). Gordon, A. L. et al. Makassar Strait throughflow, 2004 to 2006. Geophysical Research Letters 35, 2008GL036372 (2008). Pujiana, K., McPhaden, M. J., Gordon, A. L. & Napitu, A. M. Unprecedented Response of Indonesian Throughflow to Anomalous Indo‐Pacific Climatic Forcing in 2016. JGR Oceans 124, 3737–3754 (2019). Mayer, M., Alonso Balmaseda, M. & Haimberger, L. Unprecedented 2015/2016 Indo‐Pacific Heat Transfer Speeds Up Tropical Pacific Heat Recharge. Geophysical Research Letters 45, 3274–3284 (2018). Sprintall, J. et al. The Indonesian seas and their role in the coupled ocean–climate system. Nature Geosci 7, 487–492 (2014). Lee, S.-K. et al. Pacific origin of the abrupt increase in Indian Ocean heat content during the warming hiatus. Nature Geosci 8, 445–449 (2015). Li, Y., Han, W. & Zhang, L. Enhanced Decadal Warming of the Southeast Indian Ocean During the Recent Global Surface Warming Slowdown. Geophysical Research Letters 44, 9876–9884 (2017). Li, Y., Han, W., Wang, F., Zhang, L. & Duan, J. Vertical Structure of the Upper–Indian Ocean Thermal Variability. Journal of Climate 33, 7233–7253 (2020). Jyoti, J., Swapna, P., Krishnan, R. & Naidu, C. V. Pacific modulation of accelerated south Indian Ocean sea level rise during the early 21st Century. Clim. Dyn. 53, 4413–4432 (2019). Ummenhofer, C. C., Murty, S. A., Sprintall, J., Lee, T. & Abram, N. J. Heat and freshwater changes in the Indian Ocean region. Nat Rev Earth Environ 2, 525–541 (2021). Feng, M., McPhaden, M. J., Xie, S.-P. & Hafner, J. La Niña forces unprecedented Leeuwin Current warming in 2011. Sci Rep 3, 1277 (2013). Feng, M., Benthuysen, J., Zhang, N. & Slawinski, D. Freshening anomalies in the Indonesian throughflow and impacts on the Leeuwin Current during 2010-2011: FRESHENING OF THE INDONESIAN THROUGHFLOW. Geophys. Res. Lett. 42, 8555–8562 (2015). Schwarzkopf, F. U. & Böning, C. W. Contribution of Pacific wind stress to multi-decadal variations in upper-ocean heat content and sea level in the tropical south Indian Ocean: INDIAN OCEAN HEAT CONTENT AND SEA LEVEL. Geophys. Res. Lett. 38, n/a-n/a (2011). Volkov, D. L., Lee, S.-K., Gordon, A. L. & Rudko, M. Unprecedented reduction and quick recovery of the South Indian Ocean heat content and sea level in 2014–2018. Sci. Adv. 6, eabc1151 (2020). Ayers, J. M., Strutton, P. G., Coles, V. J., Hood, R. R. & Matear, R. J. Indonesian throughflow nutrient fluxes and their potential impact on Indian Ocean productivity. Geophysical Research Letters 41, 5060–5067 (2014). Meyers, G. Variation of Indonesian throughflow and the El Niño‐Southern Oscillation. J. Geophys. Res. 101, 12255–12263 (1996). England, M. H. & Huang, F. On the Interannual Variability of the Indonesian Throughflow and Its Linkage with ENSO. Journal of Climate 18, 1435–1444 (2005). Sen Gupta, A. et al. Future changes to the Indonesian Throughflow and Pacific circulation: The differing role of wind and deep circulation changes. Geophysical Research Letters 43, 1669–1678 (2016). Santoso, A., England, M. H., Kajtar, J. B. & Cai, W. Indonesian Throughflow Variability and Linkage to ENSO and IOD in an Ensemble of CMIP5 Models. Journal of Climate 35, 3161–3178 (2022). Wijffels, S. & Meyers, G. An Intersection of Oceanic Waveguides: Variability in the Indonesian Throughflow Region. J. Phys. Oceanogr. 34, 1232–1253 (2004). Liu, Q., Feng, M., Wang, D. & Wijffels, S. Interannual variability of the I ndonesian T hroughflow transport: A revisit based on 30 year expendable bathythermograph data. JGR Oceans 120, 8270–8282 (2015). Susanto, R. D., Ffield, A., Gordon, A. L. & Adi, T. R. Variability of Indonesian throughflow within Makassar Strait, 2004–2009. J. Geophys. Res. 117, 2012JC008096 (2012). Sprintall, J., Wijffels, S. E., Molcard, R. & Jaya, I. Direct estimates of the Indonesian Throughflow entering the Indian Ocean: 2004–2006. J. Geophys. Res. 114, 2008JC005257 (2009). Hu, X. et al. Interannual Variability of the Sulawesi Sea Circulation Forced by Indo‐Pacific Planetary Waves. JGR Oceans 124, 1616–1633 (2019). Oppo, D. W. & Rosenthal, Y. The Great Indo-Pacific Communicator. Science 328, 1492–1494 (2010). Tillinger, D. & Gordon, A. L. Fifty Years of the Indonesian Throughflow*. Journal of Climate 22, 6342–6355 (2009). Potemra, J. T. & Schneider, N. Interannual variations of the Indonesian throughflow. J. Geophys. Res. 112, 2006JC003808 (2007). Yuan, D. et al. Forcing of the Indian Ocean Dipole on the Interannual Variations of the Tropical Pacific Ocean: Roles of the Indonesian Throughflow. Journal of Climate 24, 3593–3608 (2011). Yuan, D., Zhou, H. & Zhao, X. Interannual Climate Variability over the Tropical Pacific Ocean Induced by the Indian Ocean Dipole through the Indonesian Throughflow. Journal of Climate 26, 2845–2861 (2013). Sprintall, J. & Révelard, A. The Indonesian Throughflow response to Indo‐Pacific climate variability. JGR Oceans 119, 1161–1175 (2014). Xie, S.-P., Annamalai, H., Schott, F. A. & McCreary, J. P. Structure and Mechanisms of South Indian Ocean Climate Variability*. J. Climate 15, 864–878 (2002). Zhang, W., Wang, Y., Jin, F.-F., Stuecker, M. F. & Turner, A. G. Impact of different El Niño types on the El Niño/IOD relationship: Relation of IOD with Two-Type El Niño. Geophys. Res. Lett. 42, 8570–8576 (2015). Luo, J.-J. et al. Interaction between El Niño and Extreme Indian Ocean Dipole. Journal of Climate 23, 726–742 (2010). Cai, W. et al. Pantropical climate interactions. Science 363, eaav4236 (2019). Murtugudde, R., Busalacchi, A. J. & Beauchamp, J. Seasonal‐to‐interannual effects of the Indonesian throughflow on the tropical Indo‐Pacific Basin. J. Geophys. Res. 103, 21425–21441 (1998). Klein, S. A., Soden, B. J. & Lau, N.-C. Remote Sea Surface Temperature Variations during ENSO: Evidence for a Tropical Atmospheric Bridge. J. Climate 12, 917–932 (1999). Behera, S. K. et al. A CGCM Study on the Interaction between IOD and ENSO. Journal of Climate 19, 1688–1705 (2006). Huang, Z., Zhang, W., Liu, C. & Stuecker, M. F. Extreme Indian Ocean dipole events associated with El Niño and Madden–Julian oscillation. Clim Dyn 59, 1953–1968 (2022). Guo, Y., Li, Y. & Wang, F. Destinations and Pathways of the Indonesian Throughflow Water in the Indian Ocean. Journal of Climate 36, 3717–3735 (2023). Metzger, E. J. et al. Simulated and observed circulation in the Indonesian Seas: 1/12° global HYCOM and the INSTANT observations. Dynamics of Atmospheres and Oceans 50, 275–300 (2010). Shinoda, T., Han, W., Metzger, E. J. & Hurlburt, H. E. Seasonal Variation of the Indonesian Throughflow in Makassar Strait. Journal of Physical Oceanography 42, 1099–1123 (2012). Shinoda, T. et al. Impact of the Madden–Julian Oscillation on the Indonesian Throughflow in the Makassar Strait during the CINDY/DYNAMO Field Campaign. Journal of Climate 29, 6085–6108 (2016). Liang, L., Xue, H. & Shu, Y. The Indonesian Throughflow and the Circulation in the Banda Sea: A Modeling Study. JGR Oceans 124, 3089–3106 (2019). Jiang, G., Wei, J., Malanotte‐Rizzoli, P., Li, M. & Gordon, A. L. Seasonal and Interannual Variability of the Subsurface Velocity Profile of the Indonesian Throughflow at Makassar Strait. JGR Oceans 124, 9644–9657 (2019). Susanto, R. D. & Gordon, A. L. Velocity and transport of the Makassar Strait throughflow. J. Geophys. Res. 110, 2004JC002425 (2005). Gordon, A. L. et al. South China Sea throughflow impact on the Indonesian throughflow. Geophysical Research Letters 39, 2012GL052021 (2012). Hu, S. & Sprintall, J. Interannual variability of the Indonesian Throughflow: The salinity effect. JGR Oceans 121, 2596–2615 (2016). Hu, S. & Sprintall, J. Observed strengthening of interbasin exchange via the Indonesian seas due to rainfall intensification. Geophysical Research Letters 44, 1448–1456 (2017). Lee, T., Fournier, S., Gordon, A. L. & Sprintall, J. Maritime Continent water cycle regulates low-latitude chokepoint of global ocean circulation. Nat Commun 10, 2103 (2019). Du, Y. et al. Decadal trends of the upper ocean salinity in the tropical Indo-Pacific since mid-1990s. Sci Rep 5, 16050 (2015). Li, J., Li, Y., Guo, Y., Li, G. & Wang, F. Decadal variability of sea surface salinity in the Southeastern Indian Ocean: Roles of local rainfall and the Indonesian throughflow. Front. Mar. Sci. 9, 1097634 (2023). Bindoff, N. L. & Mcdougall, T. J. Diagnosing Climate Change and Ocean Ventilation Using Hydrographic Data. J. Phys. Oceanogr. 24, 1137–1152 (1994). Shee, A., Sil, S. & Gangopadhyay, A. Recent changes in the upper oceanic water masses over the Indian Ocean using Argo data. Sci Rep 13, 20252 (2023). Gordon, A. L., Susanto, R. D. & Vranes, K. Cool Indonesian throughflow as a consequence of restricted surface layer flow. Nature 425, 824–828 (2003). Susanto, R. D., Gordon, A. L. & Sprintall, J. Observations and proxies of the surface layer throughflow in Lombok Strait. J. Geophys. Res. 112, 2006JC003790 (2007). Zhu, Q. & Wang, C. Contributions of Indo‐Pacific Forcings to Interannual Variability of the Indonesian Throughflow in the Upper and Lower Layers. JGR Oceans 129, e2023JC020306 (2024). Hu, X., Xue, H. & Liang, L. Impact of ENSO on the Entrance of the Indonesian Throughflow: The Oceanic Wave Propagation. JGR Oceans 127, e2022JC018782 (2022). Du, Y. & Qu, T. Three inflow pathways of the Indonesian throughflow as seen from the simple ocean data assimilation. Dynamics of Atmospheres and Oceans 50, 233–256 (2010). Susanto, R. D. & Song, Y. T. Indonesian throughflow proxy from satellite altimeters and gravimeters. JGR Oceans 120, 2844–2855 (2015). Potemra, J. Indonesian Throughflow Transport Variability Estimated from Satellite Altimetry. oceanog 18, 98–107 (2005). Lee, T., Fukumori, I., Menemenlis, D., Xing, Z. & Fu, L.-L. Effects of the Indonesian Throughflow on the Pacific and Indian Oceans. J. Phys. Oceanogr. 32, 1404–1429 (2002). Santoso, A., Cai, W., England, M. H. & Phipps, S. J. The Role of the Indonesian Throughflow on ENSO Dynamics in a Coupled Climate Model. Journal of Climate 24, 585–601 (2011). Duan, J., Li, Y., Zhang, L. & Wang, F. Impacts of the Indian Ocean Dipole on Sea Level and Gyre Circulation of the Western Tropical Pacific Ocean. Journal of Climate 33, 4207–4228 (2020). Wang, J., Zhang, S., Jiang, H. & Yuan, D. Effects of 2019 subsurface Indian Ocean initialization on the forecast of the 2020/2021 La Niña event. Clim Dyn 60, 2419–2435 (2023). Timmermann, A. et al. El Niño–Southern Oscillation complexity. Nature 559, 535–545 (2018). Cai, W. et al. Increased frequency of extreme Indian Ocean Dipole events due to greenhouse warming. Nature 510, 254–258 (2014a). Cai, W. et al. Increasing frequency of extreme El Niño events due to greenhouse warming. Nature Clim Change 4, 111–116 (2014b). Jin, Y. et al. Ocean Heat Content Increase of the Maritime Continent Since the 1990s. Geophysical Research Letters 51, e2023GL107526 (2024). Pearce, A. F. & Feng, M. The rise and fall of the “marine heat wave” off Western Australia during the summer of 2010/2011. Journal of Marine Systems 111–112, 139–156 (2013). Wernberg, T. et al. An extreme climatic event alters marine ecosystem structure in a global biodiversity hotspot. Nature Clim Change 3, 78–82 (2013). Feng, M., Zhang, X., Sloyan, B. & Chamberlain, M. Contribution of the deep ocean to the centennial changes of the Indonesian Throughflow. Geophysical Research Letters 44, 2859–2867 (2017). Krishnan, R. et al. Deciphering the desiccation trend of the South Asian monsoon hydroclimate in a warming world. Clim Dyn 47, 1007–1027 (2016). Bollasina, M. A., Ming, Y. & Ramaswamy, V. Anthropogenic Aerosols and the Weakening of the South Asian Summer Monsoon. Science 334, 502–505 (2011) Cai, W., Sullivan, A. & Cowan, T. Interactions of ENSO, the IOD, and the SAM in CMIP3 Models. Journal of Climate 24, 1688–1704 (2011). Taschetto, A. S. et al. Cold Tongue and Warm Pool ENSO Events in CMIP5: Mean State and Future Projections. Journal of Climate 27, 2861–2885 (2014). McKenna, S., Santoso, A., Gupta, A. S., Taschetto, A. S. & Cai, W. Indian Ocean Dipole in CMIP5 and CMIP6: characteristics, biases, and links to ENSO. Sci Rep 10, 11500 (2020). Cai, W. & Cowan, T. Why is the amplitude of the Indian Ocean Dipole overly large in CMIP3 and CMIP5 climate models? Geophysical Research Letters 40, 1200–1205 (2013). Lellouche, J.-M. et al. Recent updates to the Copernicus Marine Service global ocean monitoring and forecasting real-time 1∕12° high-resolution system. Ocean Sci. 14, 1093–1126 (2018). Le Traon, P. Y., Nadal, F. & Ducet, N. An Improved Mapping Method of Multisatellite Altimeter Data. J. Atmos. Oceanic Technol. 15, 522–534 (1998). Hersbach, H. et al. The ERA5 global reanalysis. Quart J Royal Meteoro Soc 146, 1999–2049 (2020). Bleck, R. An oceanic general circulation model framed in hybrid isopycnic-Cartesian coordinates. Ocean Modelling 4, 55–88 (2002). Locarnini, Ricardo A. et al. World ocean atlas 2013. Volume 1, Temperature. http://doi.org/10.7289/V55X26VD (2013). Zweng, Melissa M. et al. World ocean atlas 2013. Volume 2, Salinity. http://doi.org/10.7289/V5251G4D (2013). Li, Y. & Han, W. Decadal Sea Level Variations in the Indian Ocean Investigated with HYCOM: Roles of Climate Modes, Ocean Internal Variability, and Stochastic Wind Forcing*. Journal of Climate 28, 9143–9165 (2015). Additional Declarations There is NO Competing Interest. Supplementary Files ExtendedData.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4745867","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":328103313,"identity":"ec658694-e491-4bef-9ee9-d9aeed4ed04a","order_by":0,"name":"Yuanlong Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4UlEQVRIiWNgGAWjYBACPmYwZQHEzAcYeAyI0MIG0SIBYiYQqYUBrgWonocYh7Gx85hJ/NwhIWfOv+bjjTcFNgz87QcYPxfgdRiPmWTvGQljyxlvN1vOMUhjkDiTwCw9g4AWCd42icQNN85uk+YxOMzAcAMkSMiWv2AtZ56BtcgTo0UabMv5HjawFgPCWtiKrWXbJIwNbrAZg/zCY3gmsVkanxZ+/sMbb75ts5EzOH/44Y03f2zk5I4fPvgZf2hzQKNPIgESOwwMjA14NTAwsD+A2ncArGUUjIJRMApGAQYAACncPpHVpLO1AAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0002-7239-5756","institution":"Institute of Oceanology","correspondingAuthor":true,"prefix":"","firstName":"Yuanlong","middleName":"","lastName":"Li","suffix":""},{"id":328103314,"identity":"86d8e81f-92a2-43bc-872d-604f31930051","order_by":1,"name":"Rui Li","email":"","orcid":"","institution":"Institute of Oceanology","correspondingAuthor":false,"prefix":"","firstName":"Rui","middleName":"","lastName":"Li","suffix":""},{"id":328103315,"identity":"df5547ff-7f33-4fb5-812a-c68156b15d55","order_by":2,"name":"Yilong Lyu","email":"","orcid":"","institution":"Institute of Oceanology","correspondingAuthor":false,"prefix":"","firstName":"Yilong","middleName":"","lastName":"Lyu","suffix":""},{"id":328103316,"identity":"2ab60192-076c-47f1-9ece-4842ce83d7a5","order_by":3,"name":"Janet Sprintall","email":"","orcid":"https://orcid.org/0000-0002-7428-7580","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Janet","middleName":"","lastName":"Sprintall","suffix":""},{"id":328103317,"identity":"36b2880d-34dc-4b72-96c4-5f5c79ee12d8","order_by":4,"name":"Fan Wang","email":"","orcid":"https://orcid.org/0000-0001-5932-7567","institution":"Institute of Oceanology, Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Fan","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2024-07-15 23:15:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4745867/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4745867/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":62591945,"identity":"d0c5fb7f-4f98-45e1-ba36-e402e2ef3940","added_by":"auto","created_at":"2024-08-16 08:10:07","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":734706,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe simulated ITF in the HYCOM.\u003c/strong\u003e \u003cstrong\u003ea,\u003c/strong\u003e Surface current magnitude (m s\u003csup\u003e-1\u003c/sup\u003e) in the global ocean. \u003cstrong\u003eb,\u003c/strong\u003e Ocean currents averaged over 0-700 m in the Indonesian Seas (location marked in \u003cstrong\u003ea\u003c/strong\u003e), with color shading denoting the magnitude. The grey contour indicates the 700 m isobath.\u003cstrong\u003e c,\u003c/strong\u003e Mean along-strait velocity at 2.5°S - 3°S in Makassar Strait (location marked in \u003cstrong\u003eb\u003c/strong\u003e), with positive velocities denoting eastward and southward flows. All results are based on the 1979-2022 annual climatology of the HYCOM CTRL simulation.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/055d0484c8a1570c6104fa0f.png"},{"id":62591108,"identity":"14719609-7ad8-4dfa-a927-761d5da7d29e","added_by":"auto","created_at":"2024-08-16 08:02:07","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":321881,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVariability of the ITF transport in the Makassar Strait.\u003c/strong\u003e \u003cstrong\u003ea,\u003c/strong\u003e Monthly ITF transport (Sv) in the Makassar Strait during 1992-2022, derived from mooring observations (black), GLORYS12V1 data (red) and HYCOM CTRL (blue). Positive values denote transports toward the IO. Red and blue rectangles highlight the peak months of extreme ITF positive and negative fluctuation, respectively. The dashed lines represent the range of mean ITF ± one standard deviation from CTRL. \u003cstrong\u003eb,\u003c/strong\u003e as in \u003cstrong\u003ea\u003c/strong\u003e, but for interannual anomalies with the monthly climatology removed. \u003cstrong\u003ec,\u003c/strong\u003e Monthly ITF transport of 2014-2022 derived from CTRL (black), WND (green), PWND (red) and IWND (blue) simulations of HYCOM. \u003cstrong\u003ed,\u003c/strong\u003e as in \u003cstrong\u003ec\u003c/strong\u003e, but for interannual transport anomalies with the monthly climatology removed. The original monthly anomalies and the 13-month low-passed anomalies are plotted as thin and thick curves, respectively.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/b88c6a103fdbd40ba920528c.png"},{"id":62591110,"identity":"d419c4f0-0a39-4535-99be-34e0d1be327e","added_by":"auto","created_at":"2024-08-16 08:02:07","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":958650,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEffects of PO and IO dynamics in two extreme events.\u003c/strong\u003e \u003cstrong\u003ea-f, \u003c/strong\u003eThe 0 - 700m average currents (m s\u003csup\u003e-1\u003c/sup\u003e) (color shading denotes the magnitude) in December 2015 (\u003cstrong\u003ea, c, e\u003c/strong\u003e) and September 2017 (\u003cstrong\u003eb, d, f\u003c/strong\u003e), derived from CTRL (\u003cstrong\u003ea, b\u003c/strong\u003e), PWND (\u003cstrong\u003ec, d\u003c/strong\u003e), and IWND (\u003cstrong\u003ee, f\u003c/strong\u003e) simulations of HYCOM. The 700 m isobath is plotted as grey contours.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/1cfd2536d46dec1927ea891a.png"},{"id":62592624,"identity":"5798436e-73f4-44b9-aa63-168213a6cf53","added_by":"auto","created_at":"2024-08-16 08:18:07","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":481169,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEffect of the IO dynamics. a,\u003c/strong\u003e Scatter plot of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e (monthly ITF transport anomaly in PWND) versus \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e (monthly ITF transport anomaly in IWND) for 2014-2022. Gray and white dots indicate positive and negative values of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eWND\u003c/sub\u003e, respectively. The orange-bordered dots represent the ITF extremes, defined as \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eCTRL\u003c/sub\u003e exceeding ±1 standard deviation range for at least three months. Blue and red shading areas denote the periods coinciding with driving and buffering effects of the IO dynamics, respectively. White areas denote \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e \u0026lt; 0.5 Sv or \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e \u0026lt; 0.5 Sv, which are not considered in our analysis. Starting from the upper-right corner and moving counterclockwise, we identify quadrants 1 to 4. \u003cstrong\u003eb, c,\u003c/strong\u003e Composite wind stress anomaly (vectors; in Pa) and SLAs (color shading; in cm) for the driving (\u003cstrong\u003eb\u003c/strong\u003e) and buffering (\u003cstrong\u003ec\u003c/strong\u003e) effects of the IO dynamics, derived from WND. The composite is calculated as the average difference between the enhanced-ITF periods minus attenuated-ITF periods.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/d7328f0591215df0345a6949.png"},{"id":62591947,"identity":"e43c4beb-cf73-45b5-8e19-babe1687881e","added_by":"auto","created_at":"2024-08-16 08:10:07","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":310791,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSLA-based proxies for the ITF.\u003c/strong\u003e \u003cstrong\u003ea, b,\u003c/strong\u003e Linear correlation coefficient between the SLA and the ITF anomaly during 2014-2022 in PWND (\u003cstrong\u003ea\u003c/strong\u003e) and IWND (\u003cstrong\u003eb\u003c/strong\u003e). Black contours indicate the high-correlation regions (\u0026gt;0.7 in PWND and \u0026gt;0.5 in IWND), used to define the key regions for the construction of SLA proxies. Dots in panels (\u003cstrong\u003ea\u003c/strong\u003e) and (\u003cstrong\u003eb\u003c/strong\u003e) indicate regression coefficients that are statistically significant at the 95% confidence level. \u003cstrong\u003ec,\u003c/strong\u003e Monthly ITFPWND and ITFIWND and their SLA-based proxies, ITFP and ITFI. Red and blue colors refer to PO and IO components. \u003cstrong\u003ed,\u003c/strong\u003eMonthly SLA-based proxy for the total ITF transport anomaly, that is, the ITFP + ITFI (black), compared with the ITF transport anomaly in CTRL (grey) during 1979-2022. See Methods for the construction of the SLA-based proxy.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/8b54f1603b42e644809bca9c.png"},{"id":62591111,"identity":"12ecd3d1-2c7c-4e66-aa57-84212a7461bc","added_by":"auto","created_at":"2024-08-16 08:02:07","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":547353,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRole of the IO dynamics determined by the ENSO-IOD relationship. a,\u003c/strong\u003e Scatter plot of the Niño-3.4 versus the DMI during 1979-2022. The color of the dots denotes the \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e proxy, and the blacked-bordered (frameless) points refer to driving (buffering) cases. Neutral conditions of ENSO and IOD are defined within ±0.5°C range of Niño-3.4 and ±0.13 °C of DMI and marked by dashed lines. Conditions of El Niño (EN), La Niña (LN), positive IOD (pIOD), and negative IOD (nIOD) and their co-occurrence (such as EN + pIOD) are marked, along with their occurrence probability (the percentage of time for this specific condition relative to all conditions). Starting from the upper-right corner and moving counterclockwise, we identify quadrants 1 to 4. \u003cstrong\u003eb - g,\u003c/strong\u003e Composite of zonal wind stress anomaly (contours; in Pa) and SLA (color shading; in cm) for the conditions of El Niño plus pIOD (\u003cstrong\u003eb\u003c/strong\u003e), La Niña plus nIOD (\u003cstrong\u003ec\u003c/strong\u003e), independent El Niño (\u003cstrong\u003ed\u003c/strong\u003e), independent La Niña (\u003cstrong\u003ee\u003c/strong\u003e), independent nIOD (\u003cstrong\u003ef\u003c/strong\u003e), and independent pIOD (\u003cstrong\u003eg\u003c/strong\u003e), derived from CTRL. All the analysis is conducted for ITF\u003csub\u003eI \u003c/sub\u003eand ITF\u003csub\u003eI \u003c/sub\u003e+ ITF\u003csub\u003eP\u003c/sub\u003e samples exceeding the ±0.4 standard deviation range (± 0.4 Sv and 0.7 Sv, respectively).\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/7a6654e0387f18740f8044f8.png"},{"id":70795123,"identity":"2217c88b-0a30-47ba-ad1e-65c0904e1d1b","added_by":"auto","created_at":"2024-12-06 20:43:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3853254,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/f36aa991-5686-463c-9b9b-cc7d787fefe1.pdf"},{"id":62591106,"identity":"0780527c-5f8f-4769-b3b5-e2179c6f3aee","added_by":"auto","created_at":"2024-08-16 08:02:07","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1725999,"visible":true,"origin":"","legend":"","description":"","filename":"ExtendedData.docx","url":"https://assets-eu.researchsquare.com/files/rs-4745867/v1/658fa45eb574bcab59820d1f.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Role of the Indian Ocean dynamics in the Indonesian Throughflow variability and extremes","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe Indonesian Throughflow (ITF) refers to a complex ocean current system flowing from the Pacific (PO) to the Indian Ocean (IO) through seas and channels within the Maritime Continent\u003csup\u003e1\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e3\u003c/sup\u003e. As a vital component of the global conveyor belt\u003csup\u003e1\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e4\u003c/sup\u003e, the ITF serves as the primary conduit for the inter-basin exchange between the PO and IO. In situ measurements in the Makassar Strait\u003csup\u003e5\u003c/sup\u003e \u0026ndash; the primary inflow passage of the ITF with a mean volume transport of ~12 Sv\u003csup\u003e6\u003c/sup\u003e \u0026ndash; have documented pronounced variability in the ITF strength during the past decade, such as an abrupt drop during 2015 and 2016\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e8\u003c/sup\u003e and a surge to nearly 20 Sv in 2017\u003csup\u003e5\u003c/sup\u003e. These changes, probably accompanied by property changes of the ITF water\u003csup\u003e9\u003c/sup\u003e, have led to dramatic heat and freshwater redistributions over the Indo-Pacific Oceans\u003csup\u003e8\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e10\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e14\u003c/sup\u003e and exert regional impacts on climate extremes\u003csup\u003e15\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e16\u003c/sup\u003e, sea level\u003csup\u003e13\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e17\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e18\u003c/sup\u003e, and biogeochemical cycles\u003csup\u003e19\u003c/sup\u003e. Yet, mechanisms governing the ITF variability and extremes have not been fully established, inhibiting accurate prediction of the ITF variability and its impacts.\u003c/p\u003e\n\u003cp\u003eYear-to-year variability of the ITF has been linked to El Ni\u0026ntilde;o/Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD), with both modes causing oceanic wave adjustment in the Indo-Pacific Ocean circulations and thereby affecting the ITF\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e20\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e23\u003c/sup\u003e. Changes in Pacific trade winds associated with ENSO modulate the Indo-Pacific sea-level gradient and the ITF by evoking oceanic Rossby waves, rendering a weaker ITF in El Ni\u0026ntilde;o condition and a stronger ITF in La Ni\u0026ntilde;a conditions\u003csup\u003e24\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e26\u003c/sup\u003e. IOD events can generate equatorial Kelvin waves that propagate to the Sumatra‐Java coasts and then into the Indonesian Seas\u003csup\u003e24\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e27\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e28\u003c/sup\u003e, which also perturb the ITF strength\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e29\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e34\u003c/sup\u003e. Notably, a developing El Ni\u0026ntilde;o can frequently, though not always, trigger a positive IOD through atmospheric teleconnection\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e34\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e38\u003c/sup\u003e, which then drives upwelling Kelvin waves that act to alleviate the ITF weakening induced by the El Ni\u0026ntilde;o\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e25\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e34\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e39\u003c/sup\u003e. As such, the effect of the IO greatly complicates the relationship between ENSO and the ITF and brings difficulties to the prediction of the ITF variability.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eClarifying the role of the IO in the ITF variability is challenging. Climates of the tropical Indo-Pacific Oceans are intrinsically coupled\u003csup\u003e38\u003c/sup\u003e, particularly in the vital interaction between ENSO and IOD\u003csup\u003e37\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e40\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e42\u003c/sup\u003e. The observed correlation of ~0.3-0.5 between the ITF and IOD stems largely from the impacts of ENSO on both\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e25\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e43\u003c/sup\u003e. As such, it is difficult to cleanly isolate the IO effect by analyzing short observational records. Moreover, resolutions of the Coupled Model Intercomparison Project Phase 6 (CMIP6) models are typically ~100 km and so do not resolve the narrow passages of the ITF. With unrealistic topography in the Maritime Continent, these models cannot faithfully simulate the ITF structure in climatology\u003csup\u003e22\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e26\u003c/sup\u003e. In comparison, high-resolution (e.g., ~10 km or finer) ocean model simulations achieve a more favorable representation of the ITF and wave propagation into the Indonesian Seas\u003csup\u003e28\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e44\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e48\u003c/sup\u003e. These issues highlight the need to reproduce the ITF transport and its changes consistent with observations through high-resolution model simulations and explore the underlying mechanisms.\u003c/p\u003e\n\u003cp\u003eEncouraged by the success in reproducing observed ITF variations since the 1990s, here we use a quasi-global, eddy-resolving ocean model constrained by realistic topography in the Maritime Continent (\u003cstrong\u003eMethods\u003c/strong\u003e) to better understand the competing PO and IO mechanisms governing the ITF variability. Specifically, we carried out a series of regional forcing experiments to isolate the effect of wind-driven dynamical processes in the IO on the ITF. Our results reveal a diverse role of the IO in the ITF variability, emerging as a buffering effect for ~56% and a driving effect for ~44% of the time. Both effects are considerable in modulating the ITF transport and its extremes. The buffering effect tends to emerge during strong ENSO events, while the driving effect is usually exerted by IOD events that occur independent of ENSO. These findings will improve the predictability of the ITF variability and its broad impacts, such as regional climate extremes surrounding the Maritime Continent. Our work also highlights the inter-basin climate interaction between the Indo-Pacific Oceans through the ITF, which has potential implications for climate model simulations.\u0026nbsp;\u003c/p\u003e"},{"header":"Results","content":"\u003ch2\u003eInterannual variability and extremes of the ITF\u003c/h2\u003e\n\u003cp\u003eHere, we utilize a quasi-global (75\u0026deg;S-75\u0026deg;N; \u003cstrong\u003eFig. 1a\u003c/strong\u003e) hindcast of the Hybrid Coordinate Ocean Model (HYCOM) to simulate the ITF variability and understand the underlying processes (\u003cstrong\u003eMethods\u003c/strong\u003e). By adopting eddy-resolving horizontal resolutions of 0.1\u0026deg;\u0026times;0.1\u0026deg; and applying topography correction for the Maritime Continent, the control run (CTRL) of HYCOM has produced the realistic flow pathways and structure of the ITF (\u003cstrong\u003eFig. 1b, c\u003c/strong\u003e), as referenced to existing observation-based knowledge\u003csup\u003e5\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e6\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e26\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e27\u003c/sup\u003e. The simulated mean transport and standard deviation in CTRL agree with available mooring observations in key channels of the ITF (Makassar, Lombok, Ombai, and Timor Straits; \u003cstrong\u003eExtended data table S1\u003c/strong\u003e), particularly in the Makassar Strait (11.3 \u0026plusmn; 3.71 Sv in HYCOM versus 11.1 \u0026plusmn; 3.29 Sv in mooring observation). The simulated Makassar Strait transport achieves a correlation of \u003cem\u003er\u003c/em\u003e = 0.70 with the 13.3 years of mooring observations since 1997, which is significant at a 99% confidence level, and is also consistent with that produced by the Global Ocean Reanalysis and Simulations product (GLORYS12V1) reanalysis data (\u003cem\u003er\u003c/em\u003e = 0.73) since 1993 (\u003cstrong\u003eFig. 2a, b\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eMooring observations in the Makassar Strait have experienced pronounced ITF variability during the past decade. The volume transport abruptly dropped to ~2.0 Sv in the boreal winters (\u0026ldquo;boreal\u0026rdquo; omitted hereafter) of 2014-2015 and 2015-2016 and consequently rebounded to \u0026gt;16.0 Sv in the summers of 2016 and 2017\u0026nbsp;(\u003cstrong\u003eFig. 2a\u003c/strong\u003e). These extremes were formed by the superimposition of interannual variations upon the seasonal cycle that features a stronger Makassar Strait transport in summer than in winter\u003csup\u003e48\u003c/sup\u003e\u003csup\u003e-\u003c/sup\u003e\u003csup\u003e50\u003c/sup\u003e. The\u0026nbsp;notable impacts of these extremes on inter-basin heat and freshwater redistributions and the regional ecosystems have been documented previously\u003csup\u003e8\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e10\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e18\u003c/sup\u003e. The CTRL run of HYCOM has agreeably reproduced these variabilities and extremes, although the simulated extremes are slightly stronger in amplitude than in observations or GLORYS12V1. We define extremes when the CTRL transport exceeds the one standard deviation range (8.3-15.3 Sv; dashed lines in \u003cstrong\u003eFig. 2a\u003c/strong\u003e) for at least three months. During the 31-year period of 1992-2022, the ITF experienced 17 extremes (5 strong and 12 weak extremes), with 7 of them (1 strong and 6 weak extremes) occurring in the past decade (2013-2022). There were 29 months of extreme ITF transports in all during 2013-2022, whereas during the preceding decade of 2001-2010, there were only 13 months of extreme events. In December 2015, the simulated ITF transport was close to zero, indicating the Makassar Strait throughflow had nearly disappeared, and then subsequently strengthened to as large as 17.3 Sv (enhanced by 53%) in September 2017.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAfter the removal of the seasonal cycle, interannual variations stand out with a standard deviation of 2.5 Sv and anomalies up to ~6.0 Sv (\u003cstrong\u003eFig. 2b\u003c/strong\u003e),\u0026nbsp;confirming the key role of interannual variability in the occurrence of extremes. To explore the dynamical processes underlying the ITF variability and extremes, sensitivity experiments of HYCOM were carried out, with WND, PWND and IWND experiments representing the dynamical response of the ITF to global, tropical PO and tropical IO wind variabilities, respectively (\u003cstrong\u003eMethods\u003c/strong\u003e). Existing studies have suggested that the salinity and temperature components account for ~30% and ~70% of the interannual variability of the ITF geostrophic transport, respectively\u003csup\u003e43\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e51\u003c/sup\u003e\u003csup\u003e\u0026nbsp;\u003c/sup\u003eand emphasized the impact of precipitation-induced salinity changes\u003csup\u003e52\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e53\u003c/sup\u003e. It is important to recognize that both the salinity and temperature components are modulated by wind stress to some extent. Wind stress, through its influence on large-scale circulation, dictates the advection and leads to salinity variations of the southeastern IO\u003csup\u003e28\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e54\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e55\u003c/sup\u003e. Meanwhile, the heaving process driven by Ekman pumping\u003csup\u003e56\u003c/sup\u003e dominates upper-ocean temperature variability in the Indian Ocean\u003csup\u003e57\u003c/sup\u003e. Hence, it is not surprising to see the dominance of wind forcing in the ITF variability in our HYCOM experiments (\u003cstrong\u003eFig. 2c\u003c/strong\u003e). The ITF transport anomaly in WND (\u003cem\u003eITF\u003c/em\u003e\u003csub\u003eWND\u003c/sub\u003e) greatly resembles that of the CTRL (\u003cem\u003eITF\u003c/em\u003e\u003csub\u003eCTRL\u003c/sub\u003e), and their correlation coefficient is up to 0.95.\u003c/p\u003e\n\u003cp\u003eWijffels and Meyers (2004) found that about 60\u0026ndash;90% of sea level variability within the Indonesian Seas and southeast IO can be explained by free Kelvin and Rossby waves generated by remote zonal winds along the equator of the Indian and Pacific Oceans. In addition, certain studies have emphasized the role of local winds in the Maritime Continent\u003csup\u003e58\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e59\u003c/sup\u003e, though their primary significance lies within the realm of seasonal and higher-frequency timescales\u003csup\u003e53\u003c/sup\u003e. Here, we use PWND and IWND to isolate the roles of PO and IO wind forcing, respectively (\u003cstrong\u003eMethods; Extended Data Fig. S1\u003c/strong\u003e). The results confirm that the remote wind forcing from the equatorial PO and IO largely accounts for the ITF variability, as evidenced by a correlation of 0.85 between \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e+\u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e (the sum of ITF transport anomalies in PWND and IWND) and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eWND\u003c/sub\u003e. The missing variance suggests the possible role of local winds, but that is beyond the scope of our study. \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e surpasses\u003cem\u003e\u0026nbsp;ITF\u003c/em\u003e\u003csub\u003eCTRL\u003c/sub\u003e in magnitude in most cases (\u003cstrong\u003eFig. 2c, d\u003c/strong\u003e), indicating an overall counteracting effect of the IO dynamics. Despite an overall stronger effect of the PO winds, the IO\u0026rsquo;s influence cannot be disregarded\u003csup\u003e25\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e31\u003c/sup\u003e, and it can even assume a dominant role during certain periods\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e9\u003c/sup\u003e. In addition to the overall transport, our model also characterizes the vertical structure of the ITF\u0026rsquo;s variability in the Makassar Strait (\u003cstrong\u003eExtended Data Fig. S2\u003c/strong\u003e). Both observations\u003csup\u003e5\u003c/sup\u003e and numerical models\u003csup\u003e23\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e31\u003c/sup\u003e have suggested that the ITF may exhibit opposing transport anomalies between the upper and the deeper layers, which are linked to different dynamic processes\u003csup\u003e23\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e31\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e60\u003c/sup\u003e.\u0026nbsp;This is also seen in our HYCOM simulations which show a clear baroclinic structure of the ITF anomalies and distinct features in PWND and IWND (\u003cstrong\u003eExtended Data Fig. S2\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eRole of the IO dynamics\u003c/h2\u003e\n\u003cp\u003eNext, we attempt to understand the role of IO dynamics in modulating the ITF variability. This is first pursued by looking at two extreme cases of the Makassar Strait transport (\u003cstrong\u003eFig. 3\u003c/strong\u003e).\u0026nbsp;In December 2015, the ITF transport in CTRL was drastically weakened to \u0026lt;1.0 Sv in the Makassar Strait (\u003cstrong\u003eFig. 3a\u003c/strong\u003e). In PWND, the weakening of the ITF is even more striking. Owing to the strong PO influence associated with the 2015-2016 super El Ni\u0026ntilde;o, the ITF almost disappeared in the Makassar Strait and near the outlet regions (\u003cstrong\u003eFig. 3c\u003c/strong\u003e). By contrast, in IWND, the southward flow in the Makassar Strait was stronger than normal in December 2015 (by 16% above average in volume transport), indicating\u0026nbsp;a buffering effect of the IO winds for the weakening of the ITF (\u003cstrong\u003eFig. 3e\u003c/strong\u003e). In September 2017, the ITF was strengthened in both PWND and IWND (\u003cstrong\u003eFig. 3d, f\u003c/strong\u003e). In this case, the IO and PO dynamics thus act mutually to drive an enhancement of the ITF, in which the contributions of the IO and PO were 52% and 48%, respectively.\u003c/p\u003e\n\u003cp\u003eThe analysis of extremes presented above indicates that the IO dynamics may either buffer or drive the ITF variability. The relationship between the IO and PO effects can be visualized by a scatter plot of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e versus \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e (\u003cstrong\u003eFig. 4a\u003c/strong\u003e). Most monthly data points fall in the quadrants 2 and 4, indicating the opposite effects of the PO and IO on the ITF change. When \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e is weaker than \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e in magnitude, the IO acts to \u0026ldquo;buffer\u0026rdquo; the ITF change dominated by the PO effect. This situation accounts for 56.3% of the time during 2014-2022, indicated by red areas in \u003cstrong\u003eFig. 4a\u003c/strong\u003e. Notably, there are many weak extremes (10 months) located in the quadrant 2 with a stronger negative \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e and a weaker positive \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e. They mainly represent the buffering effect of the IO dynamics on the weakened ITF during strong El Ni\u0026ntilde;o events, as in the 2015-2016 winter (\u003cstrong\u003eFig. 3a-c\u003c/strong\u003e). By contrast, there are only three months of strong ITF extreme in quadrant 4. In that case, the PO enhanced the ITF transport owing to the La Ni\u0026ntilde;a event, which was buffered by the IO. On average, the buffering effect of the IO on the ITF change is 1.4 Sv, as quantified by the standard deviation of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e in the buffering situation. This accounts for ~41% of the corresponding ITF change driven by the PO (3.4 Sv as the standard deviation of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e). When the IO\u0026rsquo;s buffering effect operates, the total ITF anomaly (\u003cem\u003eITF\u003c/em\u003e\u003csub\u003eCTRL\u003c/sub\u003e) is 1.8 Sv in standard deviation and significantly weaker than \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e.\u003c/p\u003e\n\u003cp\u003eWhen \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e is stronger in magnitude than the opposing \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e (in quadrants 2 and 4) or of the same sign as \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e (in quadrants 1 and 3), the IO contributes constructively to the total ITF change, which is regarded as a \u0026ldquo;driving\u0026rdquo; effect. This situation constitutes 43.6% of the time during 2014-2022, marked as blue areas in \u003cstrong\u003eFig. 4a\u003c/strong\u003e. In this situation, the standard deviation of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e is 1.1 Sv, weaker than the corresponding \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e (2.2 Sv in standard deviation). The standard deviation of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eCTRL\u003c/sub\u003e is 1.6 Sv, which is smaller than the sum of the two effects above. This is because \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e does not always have the same sign as \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e in driving cases. Interestingly, the joint strengthening of the ITF by the PO and IO (quadrant 1) occurs most frequently \u0026ndash; a dominant regime for the occurrence of strong ITF extremes. By contrast, a joint weakening (quadrant 3) is seldom observed. This is linked to the complexity of the relationship between ENSO and IOD, which will be discussed in the following subsection.\u003c/p\u003e\n\u003cp\u003eOn interannual timescales, changes in the ITF strength are mainly controlled by the interbasin sea-level gradient between the PO and IO\u003csup\u003e2\u003c/sup\u003e, which is in turn perturbed by wind-forced planetary waves\u003csup\u003e24\u003c/sup\u003e. The diverse effects of the IO dynamics can also be understood in this framework. A composite analysis suggests that the driving effect on the ITF is linked to sea-level anomalies (SLAs) in the eastern tropical IO (\u003cstrong\u003eFig. 4b\u003c/strong\u003e) that alter the inter-basin sea-level gradient and the ITF strength. These SLAs near the exit region of the ITF are established by equatorial winds through eastward-propagating Kelvin waves\u003csup\u003e7\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e61\u003c/sup\u003e. A buffering scenario is usually linked to a pattern of in-phase SLAs in the entrance and exit regions (the western PO and the eastern IO, respectively) that are driven by opposing equatorial wind anomalies in the two basins. In \u003cstrong\u003eFig. 4c\u003c/strong\u003e for example, the positive SLAs on the IO side act to dampen the enhancements of the sea-level gradient and the ITF induced by the stronger positive SLAs in the PO side.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eThe relationship between the ITF with ENSO and IOD events\u003c/h2\u003e\n\u003cp\u003eNext, we explore what determines the buffering or driving role of the IO dynamics. The above analysis has highlighted the key role of SLA in the ITF variability arising from IO and PO. Various SLA-based proxies for the ITF strength have been proposed by existing studies to understand how ENSO and IOD give rise to ITF changes\u003csup\u003e30\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e45\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e62\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e63\u003c/sup\u003e. In these studies, SLAs in key regions on the IO and PO sides, usually identified by the high correlations between the ITF transport and SLA, are used to construct the inter-basin SLA gradient to represent the ITF strength. In this study, this approach is applied separately to PWND and IWND experimental output (\u003cstrong\u003eMethods\u003c/strong\u003e). We identified the key regions with maximum correlation coefficients (\u003cstrong\u003eFig. 5a, b\u003c/strong\u003e). In PWND, a rectangle region of 130\u0026deg;E-160\u0026deg;E, 5\u0026deg;N-15\u0026deg;N stands out with a maximum correlation of \u0026gt;0.7, whereas in IWND, a diamond-shape region surrounding the Sumatra-Java island chain is identified as the key region (\u003cem\u003er\u003c/em\u003e \u0026gt; 0.5) on the IO side of the ITF (\u003cstrong\u003eFig. 5b\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNote that the key region on the IO side differs from all existing studies that use the total ITF anomaly to seek key regions\u003csup\u003e30\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e63\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e64\u003c/sup\u003e. This can be understood by its correlations with the ITF in different experiments (\u003cstrong\u003eExtended Data Fig. S3\u003c/strong\u003e): In IWND, the SLA in the IO key region shows a high correlation with \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e, but in CTRL, mimicking the observed ocean, its correlation with \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eCTRL\u003c/sub\u003e is minimal because of the dominance of the PO dynamics in the ITF variability. As such, it is difficult to identify the \u0026ldquo;true\u0026rdquo; key region on the IO side through correlation analysis of observational data. By contrast, the key region on the PO side shows a high correlation with the ITF in both PWND and CTRL (\u003cstrong\u003eExtended Data Fig. S3\u003c/strong\u003e). Using the SLAs of the two key regions, we construct proxies of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e through least-square fitting in PWND and IWND, respectively. Further, by applying SLAs of CTRL to the proxy algorithm, we obtain proxies of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e for the entire 1979-2022 period, denoted as \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eP\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e, respectively (\u003cstrong\u003eFig. 5c\u003c/strong\u003e). Note that altimetric SLAs since 1992 can also be used in this algorithm to generate observation-based proxies, which are consistent with the model-based proxies (\u003cstrong\u003eExtended Data Fig. S4\u003c/strong\u003e). The sum of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eP\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e compares favorably well with \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eCTRL\u003c/sub\u003e during 1979-2022, showing a correlation of 0.82 (\u003cstrong\u003eFig. 5d\u003c/strong\u003e). This indicates that the proxies have captured the primary mechanisms governing the interannual variability of the ITF.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGiven that the ENSO and IOD are the two primary origins of interannual wind variabilities over the equatorial Indo-Pacific Oceans, we hypothesize that the complexity of the ENSO-IOD relationship is deterministic in the diverse role of IO dynamics. Compared to \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e, the lengthened \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e allows us to explore the diverse role of IO dynamics under the impacts of ENSO and IOD events more robustly. Among the 525 months of 1979-2022, there were 211 months with significant IO-driven ITF changes (defined as exceeding the \u0026plusmn;0.4 standard deviation: \u0026plusmn;0.4 Sv for \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e and \u0026plusmn;0.7 Sv for \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e + \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eP\u003c/sub\u003e) and used for our analysis. The results suggest that the IO\u0026rsquo;s effect on the ITF, represented by \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e, is modulated by both ENSO and IOD (\u003cstrong\u003eFig. 6a\u003c/strong\u003e). Generally, El Ni\u0026ntilde;o and positive IOD (pIOD) conditions render a positive \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e (the IO enhancing of the ITF), while La Ni\u0026ntilde;a and negative IOD (nIOD) conditions favor a negative \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e. It is interesting to note that in quadrant 4, even in nIOD condition, there are positive \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e values dictated by the El Ni\u0026ntilde;o. This clearly points to the essence of ENSO, in addition to the IOD, in driving the IO dynamics.\u003c/p\u003e\n\u003cp\u003eA buffering effect of the IO mainly operates in two situations. The first is during the co-occurrence of in-phase ENSO and IOD conditions, e.g., an El Ni\u0026ntilde;o accompanied by a pIOD (pIOD) (or a La Nina plus a nIOD), which accounts for 35.0% of all cases. The positive and negative scenarios are nearly symmetric in occurrence frequency, 17.0% versus 18.0% of all cases (in quadrants 1 and 3 of \u003cstrong\u003eFig. 6a\u003c/strong\u003e), respectively. The composites show that westerly winds in the equatorial PO associated with El Ni\u0026ntilde;o attenuate the ITF resulting in a strong sea-level decrease on the PO side, while the easterly winds of the co-occurring pIOD act to enhance the ITF by causing weaker sea-level falling on the IO side\u003cstrong\u003e\u0026nbsp;(Fig. 6c)\u003c/strong\u003e.\u0026nbsp;The stronger impact of the El Nino results in a IO buffering effect that counteracts the weakening of the ITF.\u0026nbsp;This scenario was commonly observed during strong El Ni\u0026ntilde;o events that are more apt to trigger the pIOD, such as 2015-2016 one (\u003cstrong\u003eFig. 2a-c\u003c/strong\u003e). The negative scenario of La Ni\u0026ntilde;a plus nIOD leads to a similar pattern of the opposite sign (\u003cstrong\u003eFig. 6f\u003c/strong\u003e). The other situation is the ENSO events occurring in a neutral condition of IOD (\u003cstrong\u003eFig. 6d, g\u003c/strong\u003e). This scenario was also commonly observed, accounting for 22.5% of all cases (8.5% for El Ni\u0026ntilde;o plus 14.0% for La Ni\u0026ntilde;a; \u003cstrong\u003eFig. 6a\u003c/strong\u003e). An ENSO event gives rise to wind anomalies in the eastern equatorial IO through atmospheric teleconnections\u003csup\u003e35\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e\u003csup\u003e40\u003c/sup\u003e, causing SLAs on the IO side (\u003cstrong\u003eFig. 6d, g\u003c/strong\u003e) and buffers the ENSO-induced ITF change.\u003c/p\u003e\n\u003cp\u003eA driving effect of the IO arises primarily from IOD events\u0026nbsp;that occur \u0026ldquo;independent\u0026rdquo; of ENSO, which makes up 24.4% of all cases. Westerly equatorial winds in the IO associated with an independent nIOD event attenuate the ITF by evoking downwelling Kelvin waves (\u003cstrong\u003eFig. 6e\u003c/strong\u003e), whereas a pIOD involves easterly winds and drives a strengthening of the ITF (\u003cstrong\u003eFig. 6b\u003c/strong\u003e). Note that the independent nIOD is more frequently observed than the independent pIOD (18.0% versus 6.4% in all cases, respectively). Another situation for the driving effect is for out-of-phase ENSO and IOD events, representing 7.6% of all cases, including La Ni\u0026ntilde;a plus pIOD (4.2%) and El Ni\u0026ntilde;o plus nIOD (3.4%). In this situation, ENSO and IOD operate mutually to drive the ITF change, with the strong ITF extreme in 2017 summer as an example. This situation has much fewer samples than the in-phase scenarios (35.0%), reflecting the overall positive ENSO-IOD correlation, and therefore their composites are weak and insignificant (not shown). This also determines that the buffering effect of the Indian Ocean prevails over its driving effect. We should state that the four situations described above are the main rather than the whole story.\u0026nbsp;They add up to 89.5% of all cases, with the remaining 10.5% representing neutral conditions of both ENSO and IOD. The four situations also represent the common features of all samples, not the exceptional cases. For example, when a strong pIOD occurs with a weak El Nino, the IO effect may dominate the total ITF change, making up a driving rather than buffering effect (such as November 2019).\u003c/p\u003e"},{"header":"Summary and implications","content":"\u003cp\u003eMooring measurements in the Makassar Strait have documented pronounced variability in the ITF strength during the past decade of 2013\u0026ndash;2022, showing extremes of 1.9 Sv in December 2015 and 16.6 Sv in September 2017. While the PO dynamics associated with ENSO cannot fully explain these changes, the role of IO dynamics remains largely uncertain. Here, with a novel attempt based on a series of high-resolution HYCOM experiments, we quantitatively reveal a diverse role of the IO dynamics in regulating the interannual variability and extremes of the ITF. The IO dynamics, primarily Kelvin waves driven by equatorial IO winds, can either buffer or drive the ITF variability, with both effects contributing to ITF extremes. The buffering effect takes place more frequently than the driving effect, 56% versus 44% in operating time, respectively. This diversity of the IO\u0026rsquo;s role stems from the complexity of the ENSO-IOD relationships. The IO tends to buffer the PO-driven ITF changes when ENSO events occur alone or in phase with IOD events, while its driving effect arises primarily from IOD events that occur independent of ENSO. Our finding underpins the essence of the IO on the ITF variability. The IO dynamics act as the primary agent for the IO climate to affect the ITF, providing implications for the prediction of the ITF and its impacts on surrounding regions.\u003c/p\u003e \u003cp\u003eModeling studies have shown that the ITF can affect the ENSO dynamics\u003csup\u003e65,66\u003c/sup\u003e. By causing changes in the ITF transport, the IOD affects the warm water volume of the western Pacific and thereby modulates the evolution of ENSO in the following year\u003csup\u003e33,67,68\u003c/sup\u003e. This study further reveals that the IO dynamics that cause the ITF changes can arise from not only the IOD but also the ENSO\u0026rsquo;s teleconnection. This likely implies that the ITF and IO are vitally involved in the \u0026ldquo;self-regulating\u0026rdquo; regime of the ENSO cycle. For example, during El Ni\u0026ntilde;o conditions, the upwelling Kelvin waves in the IO, arising from either a pIOD or the El Ni\u0026ntilde;o\u0026rsquo;s teleconnection, buffer the weakened ITF and thereby contribute to the discharge of the tropical PO. This process is favorable for the quick demise of the El Ni\u0026ntilde;o. Alternatively, an nIOD acts to further weaken the ITF and allow heat to further accumulate in the PO\u003csup\u003e8\u003c/sup\u003e \u0026ndash; a process that prolongs the El Ni\u0026ntilde;o. In this sense, the diverse role of the IO dynamics serves as a potential source of ENSO complexity\u003csup\u003e69\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe occurrence of extreme ENSO and IOD events is projected to increase in a warming climate\u003csup\u003e70,71\u003c/sup\u003e, which makes us wonder whether the ITF variability will amplify. For the first time, our work calls attention to the increasingly observed ITF extremes and reveals the underlying dynamical complexity. Changes in the ITF strength substantially perturb the heat budget of the Maritime Continent\u003csup\u003e72\u003c/sup\u003e and the southeast IO\u003csup\u003e10\u0026ndash;12\u003c/sup\u003e. These ITF extremes energize regional climate extremes such as marine heatwaves, amplifying the stress of climate change on the vulnerable marine ecosystems in these regions\u003csup\u003e73,74\u003c/sup\u003e. This calls for an urgent investigation into whether the ITF extremes will increase and lead to more regional climate extremes in surrounding regions.\u003c/p\u003e \u003cp\u003eMeanwhile, climate models consistently project a reduction up to 3.4 Sv in the ITF transport in response to future greenhouse warming, which was attributed to the suppressed deep-layer upwelling in the PO\u003csup\u003e22,75\u003c/sup\u003e. Our work raises interesting questions regarding the influence of the IO dynamics, particularly those in response to the altering South Asian monsoon\u003csup\u003e76,77\u003c/sup\u003e, on the ITF centennial changes. It remains unclear how the weakening ITF with amplifying variability affects the ENSO and regional climate.\u003c/p\u003e \u003cp\u003eHowever, at present, accurate simulation of the ITF remains a challenging task for climate models owing to unrealistic terrains and flawed parameterization schemes\u003csup\u003e23\u003c/sup\u003e. As the only oceanic conduit between the tropical Indo-Pacific Oceans, errors in the simulated ITF may propagate into the simulated ENSO and IOD. There are long-standing biases of the simulated ENSO and IOD in climate models\u003csup\u003e78,79,80\u003c/sup\u003e. For example, the strong IOD amplitude bias throughout successive generations of models is attributed to an overly active Bjerknes feedback in the southeastern tropical IO\u003csup\u003e81\u003c/sup\u003e. The successful simulation and mechanistic understanding of the ITF variability may excite more extensive investigations of the ITF and its simulation in climate models. This shall advance our understanding of inter-basin climate interaction significantly and shed light on the pathway forward for improving climate models.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eObserved ITF volume transport in the Makassar Strait.\u003c/strong\u003e There are a total of ~\u0026thinsp;13.3 years of Acoustic Doppler Current Profilers (ADCP) measurements within the Labani Channel constriction (sill depth\u0026thinsp;~\u0026thinsp;680 m) of the Makassar Strait, consisting of November 1996-July 1998 by the Arlindo program, January 2004-November 2006 by the INSTANT program (International Nusantara Stratification and Transport Program), and November 2006-August 2017 by MITF (Monitoring the ITF) program\u003csup\u003e5\u003c/sup\u003e. Two moorings (western and eastern) were deployed during Arlindo and INSTANT\u003csup\u003e6,49\u003c/sup\u003e, while the eastern mooring was not redeployed during MITF. To calculate the ITF volume transport, we use the along strait velocities\u003csup\u003e5\u003c/sup\u003e (ASVs) at the western mooring to represent the Makassar Strait throughflow across the Labani Channel. The downstream direction along the Labani Channel axis is 170\u0026deg; (referenced to true north). ASV is parallel to the Labani Channel axis. In practice, we assumed the velocity adjacent to the sidewalls as zero and applied linear interpolation from the western mooring (2\u0026deg;51.9\u0026prime; S, 118\u0026deg;27.3\u0026prime; E) to the sidewalls. Then, we integrated the ASVs across the section to obtain the volume transport\u003csup\u003e26\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eObservation-based reanalysis datasets.\u003c/strong\u003e The monthly ocean current data of the Global Ocean Reanalysis and Simulations product (GLORYS12V)\u003csup\u003e82\u003c/sup\u003e and sea level satellite altimeter data from the Copernicus Marine Environment Monitoring Service (CMEMS)\u003csup\u003e83\u003c/sup\u003e are analyzed. GLORYS12V1 was designed and implemented using the current real-time global forecasting CMEMS (Copernicus Marine Environment Monitoring Service) system and driven by the NEMO3.1 ocean/sea-ice general circulation model, covering the 1993\u0026ndash;2020 period and with a 1/12\u0026deg; horizontal resolution and 50 vertical levels. For GLORYS12V1, the ITF transport (Sv) in the Makassar Strait is computed as the integration of ASVs in the upper 700 m at 3\u0026deg;S,\u003c/p\u003e\n\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e$$\\:\\text{I}\\text{T}\\text{F}=-{\\int\\:}_{700\\:m}^{0\\:m}{\\int\\:}_{{x}_{W}}^{{x}_{E}}\\text{A}\\text{S}\\text{V}(x,z,t)dxdz$$\u003c/div\u003e\n \u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\n\u003c/div\u003e\n\u003cp\u003e,\u003c/p\u003e\n\u003cp\u003ewhere \u003cem\u003ex\u003c/em\u003e\u003csub\u003eE\u003c/sub\u003e = 116.3\u0026deg;E and \u003cem\u003ex\u003c/em\u003e\u003csub\u003eW\u003c/sub\u003e = 118.8\u0026deg;E are the longitudes of the western and eastern boundaries of the strait, \u003cem\u003ex\u003c/em\u003e, \u003cem\u003ez\u003c/em\u003e, and \u003cem\u003et\u003c/em\u003e are longitude, depth, and time, respectively. Positive ITF transport indicates the flow from the PO toward IO. We also analyzed the 0.25\u0026deg; surface wind data of the fifth generation of ECMWF reanalysis (ERA5)\u003csup\u003e84\u003c/sup\u003e. The Ni\u0026ntilde;o-3.4 index, defined as the average SST anomaly over 170\u0026deg;W\u0026ndash;120\u0026deg;W, 5\u0026deg;S\u0026ndash;5\u0026deg;N, and the Dipole Mode Index (DMI), defined as the SST anomaly difference between 50\u0026deg;E\u0026thinsp;\u0026minus;\u0026thinsp;70\u0026deg;E, 10\u0026deg;S\u0026ndash;10\u0026deg;N and 90\u0026deg;E\u0026thinsp;\u0026minus;\u0026thinsp;110\u0026deg;E, 10\u0026deg;S\u0026ndash;0\u0026deg;, are downloaded from the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHYCOM and regional forcing experiments.\u003c/strong\u003e Here, we utilize the Hybrid Coordinate Ocean Model (HYCOM) version 2.3.01 to simulate the ITF variability and achieve insights into underlying mechanisms through regional forcing experiments. HYCOM combines isopycnal, sigma (terrain following), and \u003cem\u003ez\u003c/em\u003e-level coordinates to optimize the representation of oceanic processe\u003csup\u003e85\u003c/sup\u003e and has been successfully utilized to simulate the ITF\u003csup\u003e44,45\u003c/sup\u003e. In this study, the HYCOM is configured to a quasi-global domain of 75\u0026deg;S-75\u0026deg;N with 0.1\u0026deg;\u0026times;0.1\u0026deg; horizontal resolutions and 50 vertical layers. The layer thickness gradually enlarges from 3 m at the surface to about 500 m in the abyssal ocean. The model topography is based on Earth\u0026apos;s Topography and Bathymetry (ETOPO1) which is interpolated onto the model grid and smoothed with a 1\u0026deg; x 1\u0026deg; window throughout the global ocean to remove steep features. Then, the topography of the Maritime Continent is tuned toward the original ETOPO1 data by extensive manual editing, which is necessary to ensure relatively realistic flow pathways of the ITF (Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003eb) since the ETOPO1 data better depicts the terrain and flow in the Indonesian seas than the smoothed ETOPO5 data. There are sponge layers of 5\u0026deg; placed in the southern and northern boundaries, where the simulated temperature and salinity are relaxed to the monthly climatology of World Ocean Atlas 2013\u003csup\u003e86,87\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe model adopts hourly fields of 10-m wind speed, wind stress, surface shortwave and longwave radiations, precipitation rate, 2-m air temperature and humidity, and river discharge of ERA5 as surface forcing. The model is spun up for 20 years under repeated hourly atmospheric forcing of 1979. Restarting from the already spun-up solution, HYCOM is integrated forward using hourly mean ERA5 forcing fields from January 1979 to September 2022 to form the control simulation (CTRL). The CTRL is used as the reference and compared with observations to evaluate the model performance. The ITF transport in HYCOM is computed in the same manner as for GLORYS12V1. Given that 2013 was a neutral year for both ENSO and IOD (\u003cstrong\u003eExtended Data Fig. S5\u003c/strong\u003e) and the ITF was also close to its mean state (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eb), we use 2013 as the baseline state in the sensitivity experiments for the recent decade.\u003c/p\u003e\n\u003cp\u003eThe wind experiment (WND) is forced with original hourly wind stress (as in CTRL) from January 2014 through September 2022 along with repeated 2013 hourly fields for all the other forcing factors such as wind speed, radiation, precipitation, air temperature, and humidity. In our HYCOM configuration, wind stress controls all the ocean dynamical processes (circulation, waves, and mixing), while wind speed affects the evaporation rate and surface latent and sensible heat fluxes\u003csup\u003e11,12,88\u003c/sup\u003e. As such, WND represents the dynamical response of the ITF to wind variabilities of both the PO and IO. In the Pacific wind experiment (PWND), original hourly wind stress is only retained in the tropical Pacific Ocean (140\u0026deg;E-90\u0026deg;W, 25\u0026deg;S-25\u0026deg;N; \u003cstrong\u003eExtended Data Fig. S1\u003c/strong\u003e), while wind stress in other regions and other forcing fields are all fixed to repeated 2013 fields. Similarly, the Indian wind experiment (IWND) adopts the 2014\u0026ndash;2022 hourly wind stress in the tropical Indian Ocean (45\u0026deg;E-115\u0026deg;E, 20\u0026deg;S-20\u0026deg;N). To avoid abrupt changes in wind forcing, in both PWND and IWND, we apply a 5\u0026deg; transition belt surrounding the region, where winds alter gradually from the 2014\u0026ndash;2022 fields to the repeated 2013 fields. As such, the ITF transport anomalies (with the monthly climatology removed) in the PWND and IWND, written as \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e, represent the effects of wind-driven dynamical processes in the PO and IO, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConstruction of the SLA-based ITF proxy.\u003c/strong\u003e To explore the impacts of ENSO and IOD on the ITF, it is instructive to obtain longer records of \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e than only the 2014\u0026ndash;2022 period. The sea level anomaly (SLA) gradient between the western PO and the eastern IO has been proposed as a useful proxy for the ITF strength\u003csup\u003e30,62,63\u003c/sup\u003e. Regions of high correlations with the ITF transport\u003csup\u003e63\u003c/sup\u003e are usually adopted as the key regions for the SLA proxies. However, the correlation between SLA and the ITF in observations may be dictated by other processes and does not necessarily reflect a true dynamical linkage between each other. In particular, due to the strong influence of the PO SLA on the total ITF anomaly, the correlation with the total ITF anomaly on the IO side is largely determined by the correlation with the PO SLA (\u003cstrong\u003eExtended Data Fig. S3a\u003c/strong\u003e). Our regional forcing experiments with HYCOM, i.e., PWND and IWND, serve as useful tools to distinguish the influences of the PO and IO and accurately detect the regions suitable for SLA proxies.\u003c/p\u003e\n\u003cp\u003ePractically, we calculated the SLA-ITF correlation of 2014\u0026ndash;2022 separately in the PWND and IWND. The region with the maximum correlation (~\u0026thinsp;0.7 or higher) in PWND, i.e., 130\u0026deg;E-160\u0026deg;E, 5\u0026deg;N-15\u0026deg;N, is selected as the key region on the PO side (Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ea); similarly, the region with the maximum correlation in IWND (~\u0026thinsp;0.5 or higher), a diamond-shape region enveloping the Sumatra-Java island chain, is defined as the key region on the IO side (Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003eb). Note that the key region on the IO side differs from all existing studies that use the total ITF anomaly to seek key regions\u003csup\u003e30,63,64\u003c/sup\u003e. This indicates that SLAs in this region are dynamically linked to the ITF, because the influence from the PO has been precluded in IWND. Then, the proxy for the total ITF anomaly (\u003cem\u003eITF\u003c/em\u003e\u003csub\u003etotal\u003c/sub\u003e) can be calculated as the sum of the proxy ITF anomalies driven by the PO and IO dynamics, \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eP\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e,\u003c/p\u003e\n\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\n \u003cdiv class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e$$\\:{ITF}_{\\:\\text{t}\\text{o}\\text{t}\\text{a}\\text{l}}={ITF}_{\\text{P}}+{ITF}_{\\text{I}}=\\alpha\\:\\:SL{A}_{\\text{P}}+\\beta\\:\\:SL{A}_{\\text{I}}$$\u003c/div\u003e\n \u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\n\u003c/div\u003e\n\u003cp\u003e,\u003c/p\u003e\n\u003cp\u003ewhere \u003cem\u003eSLA\u003c/em\u003e\u003csub\u003eP\u003c/sub\u003e and \u003cem\u003eSLA\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e are the average SLAs in the key regions on the PO and IO sides, respectively, \u003cem\u003e\u0026alpha;\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.27 Sv cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e and \u003cem\u003e\u0026beta;\u003c/em\u003e = -0.37 Sv cm\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e are coefficients obtained through linear least-square fitting using PWND and IWND results, respectively. The correlation between the \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eP\u003c/sub\u003e proxy and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003ePWND\u003c/sub\u003e is 0.94, while that between \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e and \u003cem\u003eITF\u003c/em\u003e\u003csub\u003eIWND\u003c/sub\u003e is 0.81 (Fig. \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003ec). Then, the total ITF anomaly proxy of 1979\u0026ndash;2022 can be obtained by substituting \u003cem\u003eSLA\u003c/em\u003e\u003csub\u003eP\u003c/sub\u003e and \u003cem\u003eSLA\u003c/em\u003e\u003csub\u003eI\u003c/sub\u003e of CTRL into the above equation.\u003c/p\u003e\n\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003cbr\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll observation, reanalysis and model data that support the findings of this study are available as follows. The INSTANT mooring data: http://www.marine.csiro.au/~cow074/data/instantdata_download.html; MITF mooring data: http://ocp.ldeo.columbia.edu/res/div/ocp/projects/MITF/cm_data/; the GLORYS12V1 reanalysis: https://resources.marine.copernicus.eu/productdetail/GLOBAL_MULTIYEAR_PHY_001_030; ERA5 data are available at Complete ERA5 global atmospheric reanalysis (copernicus.eu); HYCOM simulation: \u003cu\u003ehttp://msdc.qdio.ac.cn\u003c/u\u003e; Ni\u0026ntilde;o-3.4 index: https://www.cpc.ncep.noaa.gov/data/indices/; DMI: https://psl.noaa.gov/gcos_wgsp/Timeseries/DMI/; Satellite altimeter data of the Copernicus Marine Environment Monitoring Service (CMEMS) are obtained from https://resources.marine.copernicus.eu/products.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability: \u003c/strong\u003eMATLAB codes for data analysis and graphing are available upon request. The HYCOM version 2.3.01 source code is available at \u003cu\u003ehttps://github.com/HYCOM/HYCOM-src/releases\u003c/u\u003e. \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003eThis research is jointly supported by the National Key R\u0026amp;D Program of China (2019YFA0606702), the Strategic Priority Research Program of Chinese Academy of Sciences (XDB42000000), and the Laoshan Laboratory (LSKJ202202601). Janet Sprintall acknowledges funding support from the US National Science Foundation award OCE-1851316.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions: \u003c/strong\u003eY.Li and F.W. designed the study. R.L. performed the analysis. R.L. and Y.Li drafted the paper. Y.Lyu and R.L. conducted and evaluated the model experiments. J.S. provided valuable feedback and suggestions to enhance the quality of this article. All the authors contributed to the interpretation of the results and refinement of the manuscript. \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests: \u003c/strong\u003eThe authors declare no competing interests. \u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCorrespondence and requests for materials\u003c/strong\u003e should be addressed to Y.Li and F.W.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eGordon, A. L. Interocean exchange of thermocline water. J. Geophys. Res. 91, 5037\u0026ndash;5046 (1986).\u003c/li\u003e\n \u003cli\u003eWyrtki, K. Indonesian through flow and the associated pressure gradient. J. Geophys. Res. 92, 12941\u0026ndash;12946 (1987).\u003c/li\u003e\n \u003cli\u003eGordon, A. Oceanography of the Indonesian Seas and Their Throughflow. oceanog 18, 14\u0026ndash;27 (2005).\u003c/li\u003e\n \u003cli\u003eBroecker, W. S. The great ocean conveyor. AIP Conf. Proc., 247, 129\u0026ndash;161(1992).\u003c/li\u003e\n \u003cli\u003eGordon, A. L. et al. Makassar Strait Throughflow Seasonal and Interannual Variability: An Overview. JGR Oceans 124, 3724\u0026ndash;3736 (2019).\u003c/li\u003e\n \u003cli\u003eGordon, A. L. et al. Makassar Strait throughflow, 2004 to 2006. Geophysical Research Letters 35, 2008GL036372 (2008).\u003c/li\u003e\n \u003cli\u003ePujiana, K., McPhaden, M. J., Gordon, A. L. \u0026amp; Napitu, A. M. Unprecedented Response of Indonesian Throughflow to Anomalous Indo‐Pacific Climatic Forcing in 2016. JGR Oceans 124, 3737\u0026ndash;3754 (2019).\u003c/li\u003e\n \u003cli\u003eMayer, M., Alonso Balmaseda, M. \u0026amp; Haimberger, L. Unprecedented 2015/2016 Indo‐Pacific Heat Transfer Speeds Up Tropical Pacific Heat Recharge. Geophysical Research Letters 45, 3274\u0026ndash;3284 (2018).\u003c/li\u003e\n \u003cli\u003eSprintall, J. et al. The Indonesian seas and their role in the coupled ocean\u0026ndash;climate system. Nature Geosci 7, 487\u0026ndash;492 (2014).\u003c/li\u003e\n \u003cli\u003eLee, S.-K. et al. Pacific origin of the abrupt increase in Indian Ocean heat content during the warming hiatus. Nature Geosci 8, 445\u0026ndash;449 (2015).\u003c/li\u003e\n \u003cli\u003eLi, Y., Han, W. \u0026amp; Zhang, L. Enhanced Decadal Warming of the Southeast Indian Ocean During the Recent Global Surface Warming Slowdown. Geophysical Research Letters 44, 9876\u0026ndash;9884 (2017).\u003c/li\u003e\n \u003cli\u003eLi, Y., Han, W., Wang, F., Zhang, L. \u0026amp; Duan, J. Vertical Structure of the Upper\u0026ndash;Indian Ocean Thermal Variability. Journal of Climate 33, 7233\u0026ndash;7253 (2020).\u003c/li\u003e\n \u003cli\u003eJyoti, J., Swapna, P., Krishnan, R. \u0026amp; Naidu, C. V. Pacific modulation of accelerated south Indian Ocean sea level rise during the early 21st Century. Clim. Dyn. 53, 4413\u0026ndash;4432 (2019).\u003c/li\u003e\n \u003cli\u003eUmmenhofer, C. C., Murty, S. A., Sprintall, J., Lee, T. \u0026amp; Abram, N. J. Heat and freshwater changes in the Indian Ocean region. Nat Rev Earth Environ 2, 525\u0026ndash;541 (2021).\u003c/li\u003e\n \u003cli\u003eFeng, M., McPhaden, M. J., Xie, S.-P. \u0026amp; Hafner, J. La Ni\u0026ntilde;a forces unprecedented Leeuwin Current warming in 2011. Sci Rep 3, 1277 (2013).\u003c/li\u003e\n \u003cli\u003eFeng, M., Benthuysen, J., Zhang, N. \u0026amp; Slawinski, D. Freshening anomalies in the Indonesian throughflow and impacts on the Leeuwin Current during 2010-2011: FRESHENING OF THE INDONESIAN THROUGHFLOW. Geophys. Res. Lett. 42, 8555\u0026ndash;8562 (2015).\u003c/li\u003e\n \u003cli\u003eSchwarzkopf, F. U. \u0026amp; B\u0026ouml;ning, C. W. Contribution of Pacific wind stress to multi-decadal variations in upper-ocean heat content and sea level in the tropical south Indian Ocean: INDIAN OCEAN HEAT CONTENT AND SEA LEVEL. Geophys. Res. Lett. 38, n/a-n/a (2011).\u003c/li\u003e\n \u003cli\u003eVolkov, D. L., Lee, S.-K., Gordon, A. L. \u0026amp; Rudko, M. Unprecedented reduction and quick recovery of the South Indian Ocean heat content and sea level in 2014\u0026ndash;2018. Sci. Adv. 6, eabc1151 (2020).\u003c/li\u003e\n \u003cli\u003eAyers, J. M., Strutton, P. G., Coles, V. J., Hood, R. R. \u0026amp; Matear, R. J. Indonesian throughflow nutrient fluxes and their potential impact on Indian Ocean productivity. Geophysical Research Letters 41, 5060\u0026ndash;5067 (2014).\u003c/li\u003e\n \u003cli\u003eMeyers, G. Variation of Indonesian throughflow and the El Ni\u0026ntilde;o‐Southern Oscillation. J. Geophys. Res. 101, 12255\u0026ndash;12263 (1996).\u003c/li\u003e\n \u003cli\u003eEngland, M. H. \u0026amp; Huang, F. On the Interannual Variability of the Indonesian Throughflow and Its Linkage with ENSO. Journal of Climate 18, 1435\u0026ndash;1444 (2005).\u003c/li\u003e\n \u003cli\u003eSen Gupta, A. et al. Future changes to the Indonesian Throughflow and Pacific circulation: The differing role of wind and deep circulation changes. Geophysical Research Letters 43, 1669\u0026ndash;1678 (2016).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eSantoso, A., England, M. H., Kajtar, J. B. \u0026amp; Cai, W. Indonesian Throughflow Variability and Linkage to ENSO and IOD in an Ensemble of CMIP5 Models. Journal of Climate 35, 3161\u0026ndash;3178 (2022).\u003c/li\u003e\n \u003cli\u003eWijffels, S. \u0026amp; Meyers, G. An Intersection of Oceanic Waveguides: Variability in the Indonesian Throughflow Region. J. Phys. Oceanogr. 34, 1232\u0026ndash;1253 (2004).\u003c/li\u003e\n \u003cli\u003eLiu, Q., Feng, M., Wang, D. \u0026amp; Wijffels, S. Interannual variability of the I ndonesian T hroughflow transport: A revisit based on 30 year expendable bathythermograph data. JGR Oceans 120, 8270\u0026ndash;8282 (2015).\u003c/li\u003e\n \u003cli\u003eSusanto, R. D., Ffield, A., Gordon, A. L. \u0026amp; Adi, T. R. Variability of Indonesian throughflow within Makassar Strait, 2004\u0026ndash;2009. J. Geophys. Res. 117, 2012JC008096 (2012).\u003c/li\u003e\n \u003cli\u003eSprintall, J., Wijffels, S. E., Molcard, R. \u0026amp; Jaya, I. Direct estimates of the Indonesian Throughflow entering the Indian Ocean: 2004\u0026ndash;2006. J. Geophys. Res. 114, 2008JC005257 (2009).\u003c/li\u003e\n \u003cli\u003eHu, X. et al. Interannual Variability of the Sulawesi Sea Circulation Forced by Indo‐Pacific Planetary Waves. JGR Oceans 124, 1616\u0026ndash;1633 (2019).\u003c/li\u003e\n \u003cli\u003eOppo, D. W. \u0026amp; Rosenthal, Y. The Great Indo-Pacific Communicator. Science 328, 1492\u0026ndash;1494 (2010).\u003c/li\u003e\n \u003cli\u003eTillinger, D. \u0026amp; Gordon, A. L. Fifty Years of the Indonesian Throughflow*. Journal of Climate 22, 6342\u0026ndash;6355 (2009).\u003c/li\u003e\n \u003cli\u003ePotemra, J. T. \u0026amp; Schneider, N. Interannual variations of the Indonesian throughflow. J. Geophys. Res. 112, 2006JC003808 (2007).\u003c/li\u003e\n \u003cli\u003eYuan, D. et al. Forcing of the Indian Ocean Dipole on the Interannual Variations of the Tropical Pacific Ocean: Roles of the Indonesian Throughflow. Journal of Climate 24, 3593\u0026ndash;3608 (2011).\u003c/li\u003e\n \u003cli\u003eYuan, D., Zhou, H. \u0026amp; Zhao, X. Interannual Climate Variability over the Tropical Pacific Ocean Induced by the Indian Ocean Dipole through the Indonesian Throughflow. Journal of Climate 26, 2845\u0026ndash;2861 (2013).\u003c/li\u003e\n \u003cli\u003eSprintall, J. \u0026amp; R\u0026eacute;velard, A. The Indonesian Throughflow response to Indo‐Pacific climate variability. JGR Oceans 119, 1161\u0026ndash;1175 (2014).\u003c/li\u003e\n \u003cli\u003eXie, S.-P., Annamalai, H., Schott, F. A. \u0026amp; McCreary, J. P. Structure and Mechanisms of South Indian Ocean Climate Variability*. J. Climate 15, 864\u0026ndash;878 (2002).\u003c/li\u003e\n \u003cli\u003eZhang, W., Wang, Y., Jin, F.-F., Stuecker, M. F. \u0026amp; Turner, A. G. Impact of different El Ni\u0026ntilde;o types on the El Ni\u0026ntilde;o/IOD relationship: Relation of IOD with Two-Type El Ni\u0026ntilde;o. Geophys. Res. Lett. 42, 8570\u0026ndash;8576 (2015).\u003c/li\u003e\n \u003cli\u003eLuo, J.-J. et al. Interaction between El Ni\u0026ntilde;o and Extreme Indian Ocean Dipole. Journal of Climate 23, 726\u0026ndash;742 (2010).\u003c/li\u003e\n \u003cli\u003eCai, W. et al. Pantropical climate interactions. Science 363, eaav4236 (2019).\u003c/li\u003e\n \u003cli\u003eMurtugudde, R., Busalacchi, A. J. \u0026amp; Beauchamp, J. Seasonal‐to‐interannual effects of the Indonesian throughflow on the tropical Indo‐Pacific Basin. J. Geophys. Res. 103, 21425\u0026ndash;21441 (1998).\u003c/li\u003e\n \u003cli\u003eKlein, S. A., Soden, B. J. \u0026amp; Lau, N.-C. Remote Sea Surface Temperature Variations during ENSO: Evidence for a Tropical Atmospheric Bridge. J. Climate 12, 917\u0026ndash;932 (1999).\u003c/li\u003e\n \u003cli\u003eBehera, S. K. et al. A CGCM Study on the Interaction between IOD and ENSO. Journal of Climate 19, 1688\u0026ndash;1705 (2006).\u003c/li\u003e\n \u003cli\u003eHuang, Z., Zhang, W., Liu, C. \u0026amp; Stuecker, M. F. Extreme Indian Ocean dipole events associated with El Ni\u0026ntilde;o and Madden\u0026ndash;Julian oscillation. Clim Dyn 59, 1953\u0026ndash;1968 (2022).\u003c/li\u003e\n \u003cli\u003eGuo, Y., Li, Y. \u0026amp; Wang, F. Destinations and Pathways of the Indonesian Throughflow Water in the Indian Ocean. Journal of Climate 36, 3717\u0026ndash;3735 (2023).\u003c/li\u003e\n \u003cli\u003eMetzger, E. J. et al. Simulated and observed circulation in the Indonesian Seas: 1/12\u0026deg; global HYCOM and the INSTANT observations. Dynamics of Atmospheres and Oceans 50, 275\u0026ndash;300 (2010).\u003c/li\u003e\n \u003cli\u003eShinoda, T., Han, W., Metzger, E. J. \u0026amp; Hurlburt, H. E. Seasonal Variation of the Indonesian Throughflow in Makassar Strait. Journal of Physical Oceanography 42, 1099\u0026ndash;1123 (2012).\u003c/li\u003e\n \u003cli\u003eShinoda, T. et al. Impact of the Madden\u0026ndash;Julian Oscillation on the Indonesian Throughflow in the Makassar Strait during the CINDY/DYNAMO Field Campaign. Journal of Climate 29, 6085\u0026ndash;6108 (2016).\u003c/li\u003e\n \u003cli\u003eLiang, L., Xue, H. \u0026amp; Shu, Y. The Indonesian Throughflow and the Circulation in the Banda Sea: A Modeling Study. JGR Oceans 124, 3089\u0026ndash;3106 (2019).\u003c/li\u003e\n \u003cli\u003eJiang, G., Wei, J., Malanotte‐Rizzoli, P., Li, M. \u0026amp; Gordon, A. L. Seasonal and Interannual Variability of the Subsurface Velocity Profile of the Indonesian Throughflow at Makassar Strait. JGR Oceans 124, 9644\u0026ndash;9657 (2019).\u003c/li\u003e\n \u003cli\u003eSusanto, R. D. \u0026amp; Gordon, A. L. Velocity and transport of the Makassar Strait throughflow. J. Geophys. Res. 110, 2004JC002425 (2005).\u003c/li\u003e\n \u003cli\u003eGordon, A. L. et al. South China Sea throughflow impact on the Indonesian throughflow. Geophysical Research Letters 39, 2012GL052021 (2012).\u003c/li\u003e\n \u003cli\u003eHu, S. \u0026amp; Sprintall, J. Interannual variability of the Indonesian Throughflow: The salinity effect. JGR Oceans 121, 2596\u0026ndash;2615 (2016).\u003c/li\u003e\n \u003cli\u003eHu, S. \u0026amp; Sprintall, J. Observed strengthening of interbasin exchange via the Indonesian seas due to rainfall intensification. Geophysical Research Letters 44, 1448\u0026ndash;1456 (2017).\u003c/li\u003e\n \u003cli\u003eLee, T., Fournier, S., Gordon, A. L. \u0026amp; Sprintall, J. Maritime Continent water cycle regulates low-latitude chokepoint of global ocean circulation. Nat Commun 10, 2103 (2019).\u003c/li\u003e\n \u003cli\u003eDu, Y. et al. Decadal trends of the upper ocean salinity in the tropical Indo-Pacific since mid-1990s. Sci Rep 5, 16050 (2015).\u003c/li\u003e\n \u003cli\u003eLi, J., Li, Y., Guo, Y., Li, G. \u0026amp; Wang, F. Decadal variability of sea surface salinity in the Southeastern Indian Ocean: Roles of local rainfall and the Indonesian throughflow. Front. Mar. Sci. 9, 1097634 (2023).\u003c/li\u003e\n \u003cli\u003eBindoff, N. L. \u0026amp; Mcdougall, T. J. Diagnosing Climate Change and Ocean Ventilation Using Hydrographic Data. J. Phys. Oceanogr. 24, 1137\u0026ndash;1152 (1994).\u003c/li\u003e\n \u003cli\u003eShee, A., Sil, S. \u0026amp; Gangopadhyay, A. Recent changes in the upper oceanic water masses over the Indian Ocean using Argo data. Sci Rep 13, 20252 (2023).\u003c/li\u003e\n \u003cli\u003eGordon, A. L., Susanto, R. D. \u0026amp; Vranes, K. Cool Indonesian throughflow as a consequence of restricted surface layer flow. Nature 425, 824\u0026ndash;828 (2003).\u003c/li\u003e\n \u003cli\u003eSusanto, R. D., Gordon, A. L. \u0026amp; Sprintall, J. Observations and proxies of the surface layer throughflow in Lombok Strait. J. Geophys. Res. 112, 2006JC003790 (2007).\u003c/li\u003e\n \u003cli\u003eZhu, Q. \u0026amp; Wang, C. Contributions of Indo‐Pacific Forcings to Interannual Variability of the Indonesian Throughflow in the Upper and Lower Layers. JGR Oceans 129, e2023JC020306 (2024).\u003c/li\u003e\n \u003cli\u003eHu, X., Xue, H. \u0026amp; Liang, L. Impact of ENSO on the Entrance of the Indonesian Throughflow: The Oceanic Wave Propagation. JGR Oceans 127, e2022JC018782 (2022).\u003c/li\u003e\n \u003cli\u003eDu, Y. \u0026amp; Qu, T. Three inflow pathways of the Indonesian throughflow as seen from the simple ocean data assimilation. Dynamics of Atmospheres and Oceans 50, 233\u0026ndash;256 (2010).\u003c/li\u003e\n \u003cli\u003eSusanto, R. D. \u0026amp; Song, Y. T. Indonesian throughflow proxy from satellite altimeters and gravimeters. JGR Oceans 120, 2844\u0026ndash;2855 (2015).\u003c/li\u003e\n \u003cli\u003ePotemra, J. Indonesian Throughflow Transport Variability Estimated from Satellite Altimetry. oceanog 18, 98\u0026ndash;107 (2005).\u003c/li\u003e\n \u003cli\u003eLee, T., Fukumori, I., Menemenlis, D., Xing, Z. \u0026amp; Fu, L.-L. Effects of the Indonesian Throughflow on the Pacific and Indian Oceans. J. Phys. Oceanogr. 32, 1404\u0026ndash;1429 (2002).\u003c/li\u003e\n \u003cli\u003eSantoso, A., Cai, W., England, M. H. \u0026amp; Phipps, S. J. The Role of the Indonesian Throughflow on ENSO Dynamics in a Coupled Climate Model. Journal of Climate 24, 585\u0026ndash;601 (2011).\u003c/li\u003e\n \u003cli\u003eDuan, J., Li, Y., Zhang, L. \u0026amp; Wang, F. Impacts of the Indian Ocean Dipole on Sea Level and Gyre Circulation of the Western Tropical Pacific Ocean. Journal of Climate 33, 4207\u0026ndash;4228 (2020).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWang, J., Zhang, S., Jiang, H. \u0026amp; Yuan, D. Effects of 2019 subsurface Indian Ocean initialization on the forecast of the 2020/2021 La Ni\u0026ntilde;a event. Clim Dyn 60, 2419\u0026ndash;2435 (2023).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eTimmermann, A. et al. El Ni\u0026ntilde;o\u0026ndash;Southern Oscillation complexity. Nature 559, 535\u0026ndash;545 (2018).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eCai, W. et al. Increased frequency of extreme Indian Ocean Dipole events due to greenhouse warming. Nature 510, 254\u0026ndash;258 (2014a).\u003c/li\u003e\n \u003cli\u003eCai, W. et al. Increasing frequency of extreme El Ni\u0026ntilde;o events due to greenhouse warming. Nature Clim Change 4, 111\u0026ndash;116 (2014b).\u003c/li\u003e\n \u003cli\u003eJin, Y. et al. Ocean Heat Content Increase of the Maritime Continent Since the 1990s. Geophysical Research Letters 51, e2023GL107526 (2024).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePearce, A. F. \u0026amp; Feng, M. The rise and fall of the \u0026ldquo;marine heat wave\u0026rdquo; off Western Australia during the summer of 2010/2011. Journal of Marine Systems 111\u0026ndash;112, 139\u0026ndash;156 (2013).\u003c/li\u003e\n \u003cli\u003eWernberg, T. et al. An extreme climatic event alters marine ecosystem structure in a global biodiversity hotspot. Nature Clim Change 3, 78\u0026ndash;82 (2013).\u003c/li\u003e\n \u003cli\u003eFeng, M., Zhang, X., Sloyan, B. \u0026amp; Chamberlain, M. Contribution of the deep ocean to the centennial changes of the Indonesian Throughflow. Geophysical Research Letters 44, 2859\u0026ndash;2867 (2017).\u003c/li\u003e\n \u003cli\u003eKrishnan, R. et al. Deciphering the desiccation trend of the South Asian monsoon hydroclimate in a warming world. Clim Dyn 47, 1007\u0026ndash;1027 (2016).\u003c/li\u003e\n \u003cli\u003eBollasina, M. A., Ming, Y. \u0026amp; Ramaswamy, V. Anthropogenic Aerosols and the Weakening of the South Asian Summer Monsoon. Science 334, 502\u0026ndash;505 (2011)\u003c/li\u003e\n \u003cli\u003eCai, W., Sullivan, A. \u0026amp; Cowan, T. Interactions of ENSO, the IOD, and the SAM in CMIP3 Models. Journal of Climate 24, 1688\u0026ndash;1704 (2011).\u003c/li\u003e\n \u003cli\u003eTaschetto, A. S. et al. Cold Tongue and Warm Pool ENSO Events in CMIP5: Mean State and Future Projections. Journal of Climate 27, 2861\u0026ndash;2885 (2014).\u003c/li\u003e\n \u003cli\u003eMcKenna, S., Santoso, A., Gupta, A. S., Taschetto, A. S. \u0026amp; Cai, W. Indian Ocean Dipole in CMIP5 and CMIP6: characteristics, biases, and links to ENSO. Sci Rep 10, 11500 (2020).\u003c/li\u003e\n \u003cli\u003eCai, W. \u0026amp; Cowan, T. Why is the amplitude of the Indian Ocean Dipole overly large in CMIP3 and CMIP5 climate models? Geophysical Research Letters 40, 1200\u0026ndash;1205 (2013).\u003c/li\u003e\n \u003cli\u003eLellouche, J.-M. et al. Recent updates to the Copernicus Marine Service global ocean monitoring and forecasting real-time 1∕12\u0026deg;\u0026nbsp;high-resolution system. Ocean Sci. 14, 1093\u0026ndash;1126 (2018).\u003c/li\u003e\n \u003cli\u003eLe Traon, P. Y., Nadal, F. \u0026amp; Ducet, N. An Improved Mapping Method of Multisatellite Altimeter Data. J. Atmos. Oceanic Technol. 15, 522\u0026ndash;534 (1998).\u003c/li\u003e\n \u003cli\u003eHersbach, H. et al. The ERA5 global reanalysis. Quart J Royal Meteoro Soc 146, 1999\u0026ndash;2049 (2020).\u003c/li\u003e\n \u003cli\u003eBleck, R. An oceanic general circulation model framed in hybrid isopycnic-Cartesian coordinates. Ocean Modelling 4, 55\u0026ndash;88 (2002).\u003c/li\u003e\n \u003cli\u003eLocarnini, Ricardo A. et al. World ocean atlas 2013. Volume 1, Temperature. http://doi.org/10.7289/V55X26VD \u0026nbsp;(2013).\u003c/li\u003e\n \u003cli\u003eZweng, Melissa M. et al. World ocean atlas 2013. Volume 2, Salinity. http://doi.org/10.7289/V5251G4D (2013).\u003c/li\u003e\n \u003cli\u003eLi, Y. \u0026amp; Han, W. Decadal Sea Level Variations in the Indian Ocean Investigated with HYCOM: Roles of Climate Modes, Ocean Internal Variability, and Stochastic Wind Forcing*. Journal of Climate 28, 9143\u0026ndash;9165 (2015).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-4745867/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4745867/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe Indonesian throughflow (ITF) regulates heat and freshwater distributions of the Indo-Pacific Oceans and fundamentally affects the climate. The past decade has witnessed acute interannual variations in the Makassar Strait – the main ITF inflow passage, reaching monthly extremes of 1.9 Sv (1 Sv ≡ 10\u003csup\u003e6\u003c/sup\u003e m\u003csup\u003e3\u003c/sup\u003e s\u003csup\u003e-1\u003c/sup\u003e) in 2015 and 16.6 Sv in 2017, compared with a mean transport of ~12 Sv. The Pacific Ocean dynamics dictated by El Niño/Southern Oscillation (ENSO) cannot fully explain these variations and the role of the Indian Ocean (IO) dynamics remains uncertain. Here, we use a 0.1°, quasi-global ocean model to cleanly isolate the impact of the IO dynamics on the ITF. The wind-driven IO dynamics are found to play a significant role in either buffering or driving ITF variability. The buffering effect is commonly seen during strong ENSO events, while the driving effect arises from Indian Ocean dipole (IOD) events independent of ENSO. Notably, the IO dynamics buffered the weak ITF extreme of 2015 by ~35% and contributed to the strong ITF extreme of 2017 by ~23%. Our study aids in the prediction of regional climate extremes under the intensifying ENSO and IOD scenarios expected in the future.\u003c/p\u003e","manuscriptTitle":"Role of the Indian Ocean dynamics in the Indonesian Throughflow variability and extremes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-16 08:02:02","doi":"10.21203/rs.3.rs-4745867/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"432aa2e1-38f8-401f-b1b6-21ef5c8c2dc6","owner":[],"postedDate":"August 16th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":34725177,"name":"Earth and environmental sciences/Ocean sciences/Physical oceanography"},{"id":34725178,"name":"Earth and environmental sciences/Climate sciences/Ocean sciences/Physical oceanography"}],"tags":[],"updatedAt":"2024-12-06T20:35:43+00:00","versionOfRecord":[],"versionCreatedAt":"2024-08-16 08:02:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4745867","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4745867","identity":"rs-4745867","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}