Samples were collected onboard the R/V Isabu from July 5 to 22, 2017, along the 67°E transect, and from April 6 to 25, 2018, along the 60°E and 67°E transects ( Figure 1 ). Hydrographic parameters, such as temperature and salinity, were obtained using a conductivity-temperature-depth (CTD) mounted on a Rosette sampler. Seawater and particulate samples were collected using a 12 L Niskin sampler. For chlorophyll-a (Chl-a) analysis, 2 L of seawater samples were filtered through a pre-combusted (4 h, 450°C) glass fiber filter (Whatman, 0.7 μm pore size) and stored in a cryogenic freezer (–80°C) before measurement. The 15 mL filtrated seawater was collected and stored in a refrigerator (4°C) for dissolved inorganic nitrogen (DIN) analysis. To measure dissolved ²³⁴Th (²³⁴Th_dis), 4 L of seawater samples were immediately filtered through a silver membrane filter (Sterlitech, 1.2 µm pore-size, 25 mm diameter) and acidified with concentrated HNO₃ to adjust to pH ~2. For particulate ²³⁴Th (²³⁴Th_part), an additional 2 L of seawater was filtered through the silver filter after filtering 4 L of seawater for ²³⁴Th_dis samples. POC samples were collected by filtering 4 L of seawater through the silver membrane filter, sonicating with ethanol and acetone, and washing with deionized water before filtration to reduce residual POC.

Chl-a concentrations were measured using a fluorescence sensor (WET Labs ECO-AFL/FL), and onboard calibration was performed using a Turner Designs 10-AU fluorometer. Onboard measurements of DIN concentrations were conducted using an automatic nutrient analyzer (New QuAAtro39, SEAL Analytical).

To measure the ²³⁴Th_dis activity, an internal standard (²³⁰Th, 6.5 dpm) was added to 4 L of the sample, and the pH was raised to ~8 by adding NH₄OH after isotopic equilibration (~12 h). KMnO₄ and MnCl₂ were then added and allowed to stand for over 6 h to form MnO₂ precipitates while heating above 80°C (Cai et al., 2006). After the Mn precipitates were formed, they were filtered through the silver filter and covered with two layers of aluminum foil. Onboard measurements were conducted using a low-level beta counter (RISØ National Laboratories, Denmark). Procedural blanks (n = 3) accounted for less than 5% of the average sample activity. The measurements were conducted five times to confirm ²³⁴Th activity. After beta counting, ²³⁰Th recovery was measured by adding a ²²⁹Th spike and further separating Th using an anion-exchange column (Bio-Rad Laboratories, Hercules, CA). Both ²²⁹Th and ²³⁰Th were measured using a magnetic sector field inductively coupled plasma mass spectrometer (Element 2, Thermo Scientific) (Pike et al., 2005). On-board measurements of ²³⁴Th_part were also conducted using a low-level beta counter. After that, in the land-based laboratory, the measurements were conducted five times as well. Total ²³⁴Th (²³⁴Th_total) activity was determined by summing the activities of ²³⁴Th_dis and ²³⁴Th_part. ²³⁴Th flux was calculated using ²³⁸U/²³⁴Th disequilibria following the one-dimensional particle flux model given (Bacon et al., 1996; Owens et al., 2015):

where P T h @ z is the particulate flux of ²³⁴Th integrated with depth z.

To measure the POC concentration, the sample was decarbonated by HCl fumigation in a desiccator, and measurements were conducted using an elemental analyzer (EA 2400 CHNS/O Series II, PerkinElmer, USA) (Knap et al., 1996). The procedural blanks (n = 10) were below the detection limit of the elemental analyzer (0.03 µM).

In 2017, potential temperature and salinity in the upper 200m of the study area ranged from 12 to 30°C and 34.73 to 36.43, respectively ( Figure 2 ). Relatively higher salinity was observed in the surface layer along 5°N – 10°S in 2017, and this was also observed in the subsurface layer (~200 m) in 2018. The surface mixed layer (SML) defined by the vertical distribution of the potential temperature was approximately 50 m. The potential temperature range in 2018 was similar to that observed in 2017 ( Figure 2 ). The salinity ranged from 33.83 to 35.70, slightly lower than the values observed in 2017. In the surface layer, the low-salinity water was observed in the entire region of the 60°E line and along 5°S – 20°S of the 67°E line. The SML in both 60°E and 67°E lines ranged from 25 to 50 m. On the 67°E line, the northern regions (5°S – 15°S) had a relatively shallower SML than the southern regions (15°S – 25°S). The SCTR regions in 2017 and 2018, defined as a 20°C isotherm depth with a shallow thermocline, were observed at 3°S – 12°S and 4°S – 13°S both in 60°E and 67°E lines, respectively.

Chl-a concentrations in 2017 and 2018 ranged from 0.00 to 1.62 µg L^–1 and 0.00 to 1.87 µg L^–1, respectively ( Figure 2 ). In 2017, a subsurface chlorophyll maximum (SCM) was observed between 50 to 91 m, with no significant differences between 5°N – 14°S. However, Chl-a concentrations in the surface layer (< 50 m) of the SCTR regions were higher than those in the non-SCTR regions. In 2018, SCM depths were similar to those observed in 2017 but were much shallower in the SCTR regions than in the non-SCTR regions (15°S – 24°S). The euphotic zone, defined as the depth at which photosynthetically active radiation (PAR) reaches 1% of the surface layer, ranged from 43 to 161 m in 2017 (103 ± 46 m; mean ± standard deviation), where PAR values were available ( Supplementary Table S1 ). In 2018, the euphotic zone ranged from 31 to 135 m, with an average of 73 ± 27 m. No clear differences were observed between the 2017 and 2018 observations.

DIN concentrations in 2017 and 2018 ranged from 0.01 to 38.39 µM and 0.01 to 37.98 µM, respectively ( Figure 2 ). In 2017, the vertical distribution of DIN showed a depleted concentration in the surface layer, which gradually increased with depth. No clear differences were observed between the SCTR and non-SCTR regions. The vertical distribution of DIN in 2018 was similar to that observed in 2017. However, in the SCTR regions, the DIN concentrations in the subsurface layer (75 – 100 m) were one order of magnitude higher than those in non-SCTR regions.

In 2017 and 2018, ²³⁸U activity ranged from 2.41 to 2.48 dpm L^–1 and 2.36 to 2.49 dpm L^–1, respectively, based on salinity calculations (Owens et al., 2011). In 2017, ²³⁴Th_total and ²³⁴Th_part activities were 1.63 – 2.88 dpm L^–1 and 0.06 – 0.48 dpm L^–1, respectively ( Figure 3 ). A deficiency in ²³⁴Th_total relative to ²³⁸U was observed mainly in the surface layer (0 – 100 m). In the surface layer of the SCTR regions, ²³⁴Th_total deficiencies and ²³⁴Th_part concentrations were higher than those observed in the non-SCTR regions. In 2018, ²³⁴Th_total and ²³⁴Th_part concentrations were 0.34 – 2.78 dpm L^–1 and 0.05 – 0.66 dpm L^–1, respectively ( Figure 3 ). The ²³⁴Th deficiency in the SCTR regions was higher than that observed in the non-SCTR regions. Below the surface layer, ²³⁴Th_total activities were close to the equilibrium value in both years, indicating that particle settling mainly occurred within the 100 m layer. The excess ²³⁴Th_total in the subsurface layer could be due to the release of ²³⁴Th following remineralization of organic matter.

The POC concentrations in 2017 and 2018 ranged from 0.01 to 0.48 µM and 0.01 to 3.51 µM, respectively ( Figure 3 ), which were within the typical range (0.40 – 1.09 µM) observed in the equivalent region (Subha Anand et al., 2017; Subha Anand et al., 2018), except for the 60°E transect in 2018. The vertical distribution of POC in both years decreased gradually with increasing depth from the surface layer. In 2017, no clear latitudinal differences in POC concentration except 5°N region were observed; however, in 2018, POC concentrations in the SCTR regions were one to two orders of magnitude higher than those observed in the non-SCTR regions.

We calculated ²³⁴Th flux at a depth of 100 m to examine the POC flux as recommended by Buesseler et al. (2006). The integrated ²³⁴Th fluxes at a depth of 100 m in 2017 and 2018 varied from 237 to 991 dpm m^–2 d^–1 and 270 to 3191 dpm m^–2 d^–1, respectively ( Table 1 ). A comparison of stations at similar latitudes (10°S and 15°S) in both years revealed no significant differences, except for Station 4 (5°S) in 2018. In 2017, no clear differences were observed between the SCTR (3°S – 12°S) and non-SCTR regions. However, in 2018, ²³⁴Th fluxes in the SCTR regions (4°S – 13°S), 1402 ± 1097 dpm m^–2 d^–1, were several times higher than those observed in the non-SCTR regions (15°S – 24°S), 482 ± 189 dpm m^–2 d^–1. The range of ²³⁴Th flux in the study regions, except for Station 3 in 2018 (3191 ± 167 dpm m^–2 d^–1), was similar to that observed in the western Indian Ocean (88 to 2645 dpm m^–2 d^–1; Subha Anand et al., 2017; Subha Anand et al., 2018).

The POC flux through the 100 m layer was calculated by multiplying the ratio of POC to ²³⁴Th_part by ²³⁴Th flux. In 2017, the POC flux ranged from 0.49 to 3.15 mmol C m^–2 d^–1, with no evident differences between the SCTR and non-SCTR regions ( Table 1 ). However, in 2018, POC flux increased significantly in the SCTR regions (1.1 – 18.8 mmol C m^–2 d^–1) compared with that in the non-SCTR regions (0.6 – 0.7 mmol C m^–2 d^–1). The measured POC fluxes in the study region, except at Stations 3 and 4 in 2018, fell within the typical values for equivalent regions (0.1 – 9.0 mmol C m^–2 d^–1; Subha Anand et al., 2017; Subha Anand et al., 2018). The POC export fluxes at Stations 3 and 4 in 2018 (> 15 mmol C m^–2 d^–1) were comparable to those in highly productive regions, such as the Equatorial Atlantic (15 ± 10 mmol C m^–2 d^–1; Thomalla et al., 2008) and the Arctic Atlantic (20 ± 10 mmol C m^–2 d^–1; Le Moigne et al., 2013). Based on the distributions of salinity and DIN at 5°S in 2018, a strong upwelling of cold water with high concentrations of nutrients was observed ( Figures 2E, H ) and could enhance the POC flux. Similarly, in the SCTR regions, high primary productivity was usually observed due to the upwelling process (Dilmahamod et al., 2016). However, this regional phenomenon showed high variations in the magnitude of upwelling in the subsurface layer of the SCTR regions, which depend on air-sea interactions. Therefore, our results imply that the impact of the upwelling process in the SCTR on the POC flux depends on its magnitude and should be cautioned to examine the POC flux in these regions.

The efficiency of BCP in the surface layer can be examined by the e-ratio (%), which represents the POC export flux divided by the net primary production (NPP). The NPP at each station was estimated using a carbon-based productivity model (CbPM; Westberry et al., 2008) based on Moderate Resolution Imaging Spectroradiometer satellite data provided by Oregon State University Ocean Productivity (http://www.science.oregonstate.edu/ocean.productivity/). Compared to earlier Chl-based models, the CbPM can provide more accurate NPP values since it includes information on the influence of biological and physiological states (Behrenfeld et al., 2005; Westberry et al., 2008). NPP in 2017 and 2018 ranged from 21 to 45 mmol C m^–2 d^–1 and from 13 to 49 mmol C m^–2 d^–1, respectively ( Table 1 ). In both years, higher NPP values were observed with increasing latitude; however, there was no significant difference in the orders of NPP. The e-ratios in 2017 and 2018 were 2% – 15% and 3% – 68%, respectively ( Table 1 , Figure 4 ). The range of e-ratio was similar to that observed in the Indian Ocean (0.3 – 32.4%; Subha Anand et al., 2017; Subha Anand et al., 2018), except in Stations 3 and 4 in 2018. Generally, the e-ratio in the open ocean under non-bloom conditions was below 10% (Buesseler & Boyd, 2009; Black et al., 2018). However, under bloom conditions, the e-ratio could reach 50%, owing to the extensive scavenging of ²³⁴Th by organic particles (Buesseler et al., 1992). For Stations 3 and 4 in 2018, the high Th and POC fluxes suggested efficient scavenging in the SCTR regions, mainly due to the production of large particles in 2018. High primary production by smaller size of plankton such as picoplankton contributes less POC flux from the surface to the deeper layer due to their smaller size, slow sinking, and fast consumption in the microbial loop. Whereas larger size micro (diatoms) and nanophytoplankton contribute more POC flux to the upper layer of the ocean than to primary production (Richardson and Jackson, 2007). A positive Chl-a anomaly, especially for the winter season (June – August), was observed in 2018 based on the multi-sensor ocean-color Chl-a products (Ma et al., 2022), indicating the vigorous PP in the study area. Additionally, ²³⁴Th_part concentrations were lower in SCTR regions than those observed in non-SCTR regions, indicating an increase in Th ligands in solution or the production by the plankton communities with larger size (Buesseler et al., 2006). Thus, the relatively high export flux of POC observed in the SCTR regions in 2018 (> 45%), coupled with the relatively high NPP values and DIN concentrations, could be due to the production of fresh organic particles by PP in the surface layer. Therefore, our results suggest that the nutrient input through upwelling in the SCTR regions can enhance POC flux with efficient export comparable to bloom conditions. Our results implied that these persistent physical phenomena play a significant role in estimating the efficiency of carbon sequestration in the global oceans. However, because various conditions, such as light, biomass, and nutrients, are also important for POC flux in the surface layer, caution should be exercised in interpreting POC flux, and further consistent observations with high resolution are necessary.