The field survey was conducted during the Jala Citra 3–2023 Flores Expedition, which passed through the Pantar Strait on 4–5 April and May 18, 2023. This expedition was a marine research activity with the theme "Uncovering Ocean Floor Features and Characteristics of the Flores Sea Water Column for Navigation Safety, Defense, and Sustainable Marine Development." The research cruise was organized by the Hydro-Oceanography Center, Indonesian Navy, in collaboration with researchers from relevant Ministries/Agencies, Universities, and professionals, and private parties engaged in marine surveying and mapping. One of the targeted areas of the Jala Citra Expedition was Pantar Strait Figure 1.

To investigate the origin of the EUE cold-water mass, we used the MBES Kongsberg EM 302 and CTD Valeport Midas SVX2 onboard the Research Vessel (RV) KRI SPICA 934 operated by the Center for Hydro-Oceanography Indonesian Navy. For navigation, the KRI SPICA 934 was equipped with a GNSS fUGRO SeaSTAR 3610 for accurate georeferencing.

With a 30 kHz frequency and 0.5° to 4° beam width, the depth range (swath width) of MBES EM 302 is from 10 m to 7,000 m (up to 5.5 times water depth / 8 km). The MBES 302 system has up to 288 beams / 432 soundings per swath with pointing angles that can automatically be adjusted according to the achievable coverage. In the present study, the equiangular beam spacing of the MBES EM 302 was set with a beam width and angular coverage of 1 and 12–70, respectively. The vessel speed was maintained at 4–7 knots during the measurement from 4–5 April 2023. The MBES paths in the Pantar Strait are parallel to the coastline, as shown in Figure 1, with a total path length of 105,030 m. The cross path was not conducted for safety reasons because the narrow strait and strong current in the Pantar Strait restricted vessel maneuvers.

The specifications of the CTD Midas SVX2 are listed in Table 1. The geographic location of the CTD deployment, with one cast per station, is shown in Figure 1 and Table 2. The tidal conditions of the CTD casts are shown in Figure 2. In addition, we used the CTD of Wirasatriya et al., (2023), obtained from the CTD Sea-Bird Electronic SBE 37SMP (Station 2).

The MBES data were processed using CARIS HIPS and SIPS software licensed to the Hydro-Oceanography Center of the Indonesian Navy. CARIS HIPS and SIPS are professional hydrographic data-processing software suites developed by Teledyne CARIS. This is widely considered the industry standard for processing sonar and lidar data collected during maritime surveys. The acronym HIPS (SIPS) stands for Hydrographic Information Processing System (Sonar Information Processing System) that focuses on bathymetry/depth data (seafloor imagery and backscatter). Data processing in CARIS HIPS and SHIPS follows the Combined Uncertainty and Bathymetry Estimator) (CUBE) workflow (Canadian Hydrographic Service, 2012). The MBES data analysis was divided into three steps: correction and calibration, data cleaning, and gridding and visualization Figure 3.

Vessel motion corrections were applied to the raw depth data from the MBES by adjusting the vessel pitch, roll, and yaw using motion sensors and GPS data. The results of the MBES EM 302 KRI Spica-934 patch test were obtained at depth of 124 m–3000 m Table 3. Subsequently, the data were calibrated to reduce the errors and ensure accuracy. The first calibration step was sound velocity correction, which is essential for accurate depth measurements. The speed of sound in water is influenced by the temperature, salinity, and pressure, and variations in these factors can lead to significant measurement errors. Four CTD casts were used to profile the water column and to correct for these variations. The second calibration was a tidal correction using tidal prediction, which was validated with a tide gauge at the Kalabahi Station to ensure that depth measurements were corrected relative to the mean lower low-water level. Subsequently, the horizontal (position) and vertical (depth) data were merged to produce georeferenced data.

The horizontal and vertical total propagated uncertainties (TPUs) were determined according to established procedures, including the EM 302 KRI Spica offset calculation and patch test. Analysis of the post TPU horizontal calculation values derived from the sounding results in this area indicates that the parameters influencing the sounding results are as follows: sonar 50%, roll 18%, pitch 17%, heading 9%, GPS 3%, IMU Timing 2%, and other 1%. Similarly, the post TPU vertical values from the sounding results suggest that the parameters affecting the results were sonar (angle) 38%, sonar (range) 22%, sound velocity 21%, roll 14%, tide 3%, and others 3%. The calibrated merged data were used to produce the CUBE surface. Quality control (QC) of the CUBE surface was a transducer performance test to correct any misalignment of the echosounder transducers to ensure accurate beam angles. The purpose of the performance test was to determine the best MBES swath angle for use in sounding activities. A performance test of the EM 302 KRI Spica-934 was carried out on a relatively flat bottom surface. The results of the performance tests at depths of 650–850 m with various beam angles are shown in Figure 4.

After data calibration, data cleaning was performed using a high-pass filter to remove high-frequency noise from the vessel movement and other sources. Furthermore, spurious data points such as spikes that do not represent actual seabed features were manually identified as outliers and removed by visual inspection of the swath data.

Finally, gridded data were obtained by interpolating the missing data points within the grid to create a continuous surface and merged data from multiple passes over the same area to improve coverage and accuracy. The final grid resolution of the bathymetry map was 30 m, and the bathymetry map was visualized in 2D and 3D formats. Volumetric and cross-sectional analyses were performed using QGIS 4.0.

Raw CTD data were processed using DataLog X2, a license-free application, to extract vertical sound velocity, temperature, conductivity, and salinity data. The vertical temperature profile was visualized using Ocean Data View (ODV) software. ODV is an open-source computer program used for interactive exploration, analysis, and visualization of oceanographic and other geo-referenced data. The vertical profiles of temperature data from the five CTD stations were interpolated using the Data-Interpolating Variational Analysis (DIVA) method embedded in ODV to obtain the vertical temperature distribution along the Pantar Strait.

Black circles in (a) denote the detection of four submarine ridges at the southern part of Pura Island, while the black circles in (c) are the five basins (i.e., Basins A, B, C, D, and E) along the channel that connects the Mulut Kumbang Strait to the deepest basin at the southern part of Pantar Strait. The dashed red line in (b) denotes the cross-section line of bathymetry as shown in Figure 5.

The bottom topography of the Pantar Strait was determined using bathymetry data obtained from the MBES survey on 4-5 April 2023, spanning from the Flores Sea to the southern tip of the Pantar Strait Figure 5a. The deepest part of the bathymetric survey was located in the Flores Sea, reaching a depth of 1,800 m. In the northern part of Pura Island in the Pantar Strait, the seabed is generally shallow, with a depth of less than 400 m. The southwestern part of Ternate Island features a seabed depth of up to 650 m, while the southern part of Pura Island is dominated by a deep bowl-like basin, with a wide plain reaching depths of approximately 800 m.

To investigate the source and propagation path of the cold-water mass of the EUE, the analysis focused on the middle of the Pantar Strait in the eastern part of Pura Island Figure 5c. Taking advantage of high-resolution bathymetry from the MBES, five basins ( A, B, C, D, and E) were observed along the channel that connects the Mulut Kumbang Strait to the deepest basin in the southern part of the Pantar Strait. Basins A–D have characteristic depths of 150, 350, 400, and 600 m. Basin E is the deepest and largest basin in the southern part of the Pantar Strait. Four oceanic ridges were identified at the northern tip of this basin, as indicated by the black circles in Fig. 5a. In contrast, at the southern tip, the deepest part of this basin was observed, reaching more than 1,000 m Figure 5b. Interestingly, a river-like meandering channel spanning from the continental slope and flowing southeastward to the deepest basin was observed. This indicated the existence of an ancient geological process that created the channel. However, this feature is beyond the scope of this study and is left for future research.

To obtain more detailed features of these basins, a three-dimensional (3D) model was generated along the channel, as shown in Figure 6. Figure 6 shows that the basins have coarse surfaces with ridges bordering each basin. However, a channel was clearly observed spanning from the Mulut Kumbang Strait to Basin E. From Basin E, this channel tapered to the Mulut Kumbang Strait. The tapered shape of the channel leading from Basin E to the Mulut Kumbang Strait suggests that it could act as a funnel, concentrating and directing the cold-water mass toward the strait. This narrowing of the channel may help explain the extreme nature of the upwelling event by facilitating the concentrated flow of cold water into the strait. Furthermore, a tall ridge is observed on the right side of Basin A in the western part of the mouth of the Mulut Kumbang Strait. This ridge has a steep slope at a height of 300 m from its base, which may hamper water masses spilling out of the channel.

The Observed Five Basins Along the Southern Pantar Strait Are Marked by A, B, C, D, and E. The high-resolution bathymetry map processed from MBES managed to reveal the area of Basin E, which exceeds 1,000 m in depth, covers 4.75 km² Figure 7a, and has a volume of 150.96 × 10⁶ m³. The cross-section of the bathymetry in the deepest basin shows that the deepest basin reaches 1,060 m Figure 7b.

Figure 7. (a) Areal Size of Pantar Strait with a Depth of More Than 1,000 m. The White Line Denotes the Cross-Section of Bathymetry in (b). (b) Cross-Section of Bathymetry as Shown by the White Line in (a) and the Red Dashed Line in Figure 3b.

The vertical distribution of temperature obtained from CTD measurement demonstrates that below 400 m, the temperature is lower than 10 °C. However, we can see that the 10 °C isotherm at 8.425 °S to 8.35 °S is shallower than that at 8.35 °S to 8.3 °S, indicating a shallow mixed layer depth in the southern part Figure 8. This may be due to differences in tidal conditions during CTD casting. CTD casting at stations 2 and 3 (4 and 5) was conducted during the ebb (flood) tide Figure 2. Shoaling at mixed layer depths is often associated with mechanisms such as tidal mixing and the dynamics of ocean currents. For instance, in regions where tidal forces are strong, such as off the Yangtze River Estuary, tidal mixing can induce upwelling by creating horizontal density gradients across tidal fronts. This occurs as part of a secondary circulation driven by cross-frontal density gradients, with topographic features playing a significant role in directing bottom currents upwards, thereby regulating the location and intensity of tidal fronts (Lü et al., 2006). In the case of the Pantar Strait, shoaling mixed-layer depths during flood tides also indicate an upwelling signal. The upwelling process in the Pantar Strait reduced the temperature of the surface layer. Thus, tides also influence variations in the surface temperature along the Pantar Strait. Nevertheless, the surface temperature drop is not comparable with the case of EUE at the Mulut Kumbang Strait as reported by Wirasatriya et al.,(2023), which can reduce surface temperature by ~10 °C.

A straightforward calculation can be employed to illustrate the propagation of cold-water masses from the deepest basin of the southern Pantar Strait to the Mulut Kumbang Strait during Easterly Upwelling Events (EUE). As shown in Figure 8a, the deepest point in the southern Pantar Strait reached a maximum depth of 1,060 m, whereas the horizontal distance separating this basin from the Mulut Kumbang Strait is approximately 19,425 m. Consequently, the diagonal distance accounting for both depth and horizontal displacement was calculated to be 19,454 m. Simulations conducted by Prasetyo et al. (2024) indicate that the maximum current velocity in the Pantar Strait during the spring tide varies between 2.8 m/s and 3.2 m/s. Using an average current velocity of 3 m/s, the cold water mass required roughly 1.8 h to traverse the diagonal distance from the deepest basin to the Mulut Kumbang Strait. This estimated duration aligns well with the observed EUEs, which typically last between 1 and 2 h (Wirasatriya et al., 2023, 2025a).

Using an MBES, we revealed the complex bottom topography of the southern Pantar Strait, providing new evidence for the cold water source of the EUE in the Mulut Kumbang Strait. Previously, the limitation in bathymetric coverage from SBES observations, as reported by Wirasatriya et al. (2023), identified only a single basin (Basin B) between Pura Island and the southwestern Alor subdistrict, which they posited as the source of the cold-water mass of the EUE. However, the more detailed bathymetric survey presented in this study revealed five distinct basins aligned along the channel, indicating a far more intricate bathymetric architecture than previously recognized. This complexity likely plays a crucial role in modulating cold-water mass dynamics by providing a pathway that influences water mass flow and mixing. Furthermore, while Astuti et al. (2025) proposed that the Savu Sea is the source of the cold-water mass, the deeper bathymetry in Basin E provides compelling evidence that the cold-water mass could have originated within the Pantar Strait. Their indication was based on the genetic material of bathypelagic fish obtained during the EUE in the Mulut Kumbang Strait. Because bathypelagic fishes live at a depth of 1,000–4000 m depth, the source of the cold-water mass during EUE should come from a water column deeper than 1,000 m. Based on the bathymetry map by Setiawan et al. (2020), the maximum depth of the Pantar Strait is less than 1,000 m. Thus, Astuti et al. (2025) assumed that the source of the cold-water mass came from the Savu Sea, which has a maximum depth of 3470 m (Kartaadiputra et al., 1982; Van Der Werff, 1995).

The new bathymetry presented in this study provides an alternative hypothesis that the source of the cold-water mass of EUE may come from the southernmost basin of the Pantar Strait, that is Basin E. Thus, the genetic material of bathypelagic fishes, that is Benthosema fibulatum, Dasyscopelus asper, Diaphus garmani, Diaphus suborbitalis, Diaphus watasei, and Howella cf. broidei obtained by Astuti et al. (2025) during EUE in the Mulut Kumbang Strait may come from Basin E, which has an area deeper than 1,000 m.

The finding of a tapered channel from Basin E to the Mulut Kumbang Strait exemplifies how channel geometry and bathymetry control tidal current acceleration and the associated upwelling phenomena. Narrow channels and sills have long been recognized as amplifying tidal currents, often generating localized internal mixing and secondary circulation that leads to upwelling along strait boundaries (Tassigny et al., 2026; Nakamura et al., 2000). In this context, flood tide advection into the strait brings subsurface cold, nutrient-rich water masses into shallower zones, while ebb tides flush these waters back into the open sea, producing a highly dynamic exchange driven by the tidal cycle (Kawaguchi et al., 2020).

Moreover, frictional effects on the seafloor during strong tidal currents, especially under spring tide conditions, can generate bottom Ekman transport and vertical motion, which contribute to upwelling. These bottom boundary layer dynamics are critical in straits with complex bathymetry, where the interaction of tidal currents with seafloor topography induces secondary circulation that is responsible for the vertical displacement of colder waters (Kawaguchi et al., 2020; Lieberthal et al., 2019). The repeated twice-daily pattern of upwelling events is in accordance with the semidiurnal tidal constituents commonly observed in narrow straits, where two flood and two ebb phases occur within each tidal day (Longhitano, 2010; Tsimplis et al., 1995).

Despite the relatively small tidal amplitude, tidal upwelling in similar microtidal or semi-closed basins can produce significant local enhancements in mixing and vertical nutrient transport, as tidal currents are spatially intensified through geometric contraction or resonance effects. These mechanisms explain the upwelling strictly confined to the Mulut Kumbang Strait, as opposed to a broader coastal setting where wind forcing dominates (Longhitano et al., 2010). This phenomenon starkly contrasts broad coastal upwelling systems, where wind forcing predominantly drives vertical transport. The distinct confinement of upwelling to the Mulut Kumbang Strait indicates that tidal dynamics, rather than wind forcing, are the principal drivers at this location, aligning with numerical simulations of tidal upwelling in narrow channels (Lin & Gan, 2024; Kim et al., 2023).

Although the present study provides an initial transport estimate, this simplified approach neglects the multivariate complexity of hydrodynamics, such as the effects of irregular bottom topography, vertical stratification, and internal wave interactions, which critically modulate cold-water propagation. Topographic features can induce localized tidal amplification, internal tide generation, flow separation, and mixing (Cucco et al., 2016; Bignami et al., 2025), whereas density stratification mediates internal wave dynamics and density-driven flows that influence water mass transformation and pathways (Evans et al., 2018; Li et al., 2020). These processes emphasize the need for further high-resolution hydrodynamic and numerical modelling to fully characterize cold-water transport mechanisms in this complex strait system.

Overall, the evidence indicates that the phenomenon in the Mulut Kumbang Strait is a distinctive tide-driven upwelling system that is cyclical in nature and is controlled by the interaction of spring tide dynamics with a narrow channel morphology. The fortnightly spring neap tidal variations regulate the timing and intensity, while the narrow geometry concentrates tidal currents sufficiently strongly to produce the observed cold-water ingress during flood tides and egress during ebb tides. Such tidal upwelling not only alters local thermal and nutrient conditions but can also have broader ecological impacts on productivity and biogeochemical cycling in the strait and adjacent waters (Susanto and Ray, 2022; Sharples, 2007).

In summary, integrating recent bathymetric and ecological data with established tidal dynamics frameworks highlights that tide-generated upwelling in narrow straits, such as Mulut Kumbang, arises from the amplification of semidiurnal tidal currents constrained by channel geometry. Spring neap tidal modulation imparts fortnightly variability, while bottom boundary layer processes foster vertical cold-water flux. The cyclic import and export of nutrient-rich cold water establishes a unique upwelling system with significant implications for regional oceanography and marine ecosystems, consistent with theoretical and observational insights into global strait-induced upwelling phenomena (Lin & Gan, 2024; Bignami et al., 2025; Kawaguchi et al., 2020).

Future research should focus on investigating the dynamics of cold-water mass propagation through a tapered channel connecting the basins to the Mulut Kumbang Strait. High-resolution temporal monitoring of temperature and salinity profiles along the channel during the EUE could provide insights into how cold-water masses are transported during EUEs. Additionally, numeri

cal modeling of the hydrodynamics in this region would help simulate the flow patterns and confirm the role of the channel in directing the cold-water mass. Realistic simulations require a three-dimensional baroclinic numerical model that incorporates detailed bathymetric data obtained from MBESs. Such modeling approaches capture nonlinear interactions between topography, stratification, tidal currents, and other forcing mechanisms, enabling a more accurate depiction of cold-water mass propagation during EUE. The significance of including such physical complexities in numerical studies has been well illustrated for the Southern Ocean and other straits, where three-dimensional coupled models reveal processes that are overlooked in simplified flow calculations (Evans et al., 2018; Cucco et al., 2016; Wells et al., 2025). Understanding this modeling task will provide a comprehensive understanding of water mass dynamics in the southern Pantar Strait. Furthermore, the tapered shape of the channel that connects the five basins to the Mulut Kumbang Strait likely concentrates and accelerates the flow of the cold-water mass, creating a funneling effect that enhances upwelling during extreme events. Further investigation of the hydrodynamics of this channel is essential to understand how its shape affects the propagation of cold water and contributes to the intensity of upwelling events. Because the cold-water mass could be periodically funneled toward the Mulut Kumbang Strait during specific tidal conditions, this potentially explains the timing and recurrence of EUEs. Thus, future studies should examine the correlation between the timing of spring tides, availability of cold water in the basins, and frequency of EUE to better understand the factors controlling these events.