The intricate bottom topography (i.e., widely distributed seamounts and guyots) of the West Pacific Ocean results in complicated distributions of different thermohaline currents. The deep water below about 3,500 m in the West Pacific Ocean is dominated by the Lower Circumpolar Deep Water (LCDW), which flows northward along the Chatham Rise in the western South Pacific Ocean (Chiswell et al., 2015) and enters the Central Pacific Ocean through the Samoan Passage, where it bifurcates into western, northern, and eastern branches (Figure 1) (Johnson and Toole, 1993; Kawabe et al., 2003; Kawabe and Fujio, 2010). The Upper Circumpolar Deep Water (UCDW), originated from the Antarctic Circumpolar Current (ACC), is transported by the anticyclonic flow at shallower depths through the South Pacific and ultimately enters the Philippine Sea via the Caroline Basin (Figure 1) (Kawabe et al., 2003). Upwelled LCDW in the Northeast Pacific Ocean is transformed into another upper deep water, the North Pacific Deep Water (NPDW), between 2,000 and 3,500 m, which is modified on its southward route by mixing with UCDW.

The LCDW fills the deep ocean below about 3,500 m mainly in the West Pacific Ocean (Supplementary Figure S1), producing a relatively young water mass age (500–700 years) in the Central Pacific Basin and older water mass age in the Caroline, Philippine, and North Pacific basins (700–1,000 years) between 3,500 m and 4,000 m, coinciding with the decreasing carbonate ion ([CO₃ ²⁻]) from the Central Pacific Ocean to the Northwest Pacific Ocean as a result of accumulated respiration of organic matter (Supplementary Figures S2, S4). Ages of the water mass and [CO₃ ²⁻] at a depth between 2,500 m and 3,000 m, where UCDW dominates, show weaker inter-basin contrast, but aging of the water mass (decreasing [CO₃ ²⁻]) from the South to the North can still be found in our study region (Supplementary Figures S3, S4).

Data used in this study include surficial CaCO₃ content in sediments (_wtCaCO₃%), carbonate ion concentration in the deep ocean ([CO₃ ^2-]), PIC rain (F_B), and non-carbonate flux (F_M).

Surficial sedimentary _wtCaCO₃% data used in this article are compiled from previously published references, with details appended in Supplementary Table S1. Sediment cores from the West Pacific Ocean are separated into four geographic groups (e.g., the Northwest Pacific Ocean, the Philippine Sea, the Western Equatorial Pacific Ocean (including mostly the Caroline Basin), and the Central Pacific Ocean). The topographic complex in the deep ocean (e.g., the Ogasawara–Mariana–Palau Ridges, the Caroline Ridges, the Ontong Java Plateau, and the Mid-Pacific Mountains) sets the boundary between different regions (Figure 1), and measured carbonate chemistry (GLODAPv2.2020; Olsen et al., 2020) and simulated deep water-mass ages by the Norwegian Earth System Model (NorESM; Tjiputra et al., 2013) differ among the four regions (Supplementary Figures S2, S3). Carbonate ion concentration ([CO₃ ^2-]) data used in this study were calculated from titration alkalinity, total dissolved inorganic carbon, salinity, temperature, and pressure (water depth) data obtained from the Global Ocean Data Analysis Project version 2.2020 database (GLODAPv2.2020; Millero, 1995; Olsen et al., 2020). A basin-wide distribution of the deep sea [CO₃ ²⁻] below 1,000 m is illustrated in Supplementary Figure S4, with average values between 2,000 and 3,000 m for different ocean regions and the in situ variance summarized in Table 1.

Rain (flux) of particulate inorganic carbon (PIC; F_B) to the surficial sediment consists mostly settling CaCO₃ produced in the surface ocean by marine calcifiers. To calculate the PIC rain, the net primary production was first estimated (Supplementary Figure S5A) using the monthly Moderate Resolution Imaging Spectroradiometer (MODIS) datasets based on a carbon-based production model (CBPM; Behrenfeld et al., 2005), which is then translated into the export flux of particulate organic carbon (POC; Supplementary Figure S5B) using algorithms developed by Laws et al. (2000).The export flux of PIC is then derived based on the PIC:POC export ratio (Sarmiento et al., 2002), and the PIC rain (Table 1) is assigned to be half of the export flux of PIC due to carbonate dissolution in the water column as a result of POC remineralization (Feely et al., 2004). Dilution of marine CaCO₃ in the pelagic ocean is largely determined by non-carbonate components, such as lithogenic materials and biogenic opal, despite dominant fluvial input to ocean margins. As the major source for lithogenic input (F_M), eolian dust is estimated by various means (Kienast et al., 2016), and the results are shown in Supplementary Table S2. In particular, model simulated dust deposition (Supplementary Figure S6; Jickells et al., 2005) suggests that the Northwest Pacific Ocean has the highest mean dust deposition at about 0.97 g m⁻² yr⁻¹, and dust deposition in other areas ranges from ∼0.2 g m⁻² yr⁻¹ (the Central Pacific Ocean and the Western Equatorial Pacific Ocean) to ∼0.84 g m⁻² yr⁻¹ (the Philippine Sea).

To better constrain the observed sedimentary _wtCaCO₃% distribution, a critical depth where CaCO₃ turns under-saturated beneath (the carbonate saturation depth; CSD) is first calculated from [CO₃ ²⁻] in the deep ocean between 2,500 m and 3,000 m in different basins, where the sea water turns under-saturated (Supplementary Figure S4), based on empirical relationships established by Boudreau et al. (2010). Another critical depth where the carbonate sedimentation rate balances its dissolution rate (the carbonate compensation depth; CCD) can be determined by finding the depth, below which no accumulation of CaCO₃ could be found. Observed locations of CCD can be used to derive values of the dissolution rate of CaCO₃ (k^*; Table 1) with the PIC rain (F_B), deep sea [CO₃ ^2-], and non-carbonate flux (F_M) (Li et al., 2021; Liu et al., 2022).

Ultimately, the sedimentary _wtCaCO₃% (B) at a given water depth between CSD and CCD can be calculated by: B = F B − k ∗ ( [ CO 3 2 − ] sat − [ CO 3 2 − ] ) ( F B − k ∗ ( [ CO 3 2 − ] sat − [ CO 3 2 − ] ) ) + ( ρ CaCO 3 F M / ρ M ) , (1) where ρ_CaCO3 is the mass density of CaCO₃ (2.5 × 10⁴ mol m⁻³; Boudreau, 2013), ρ_M is the density of non-carbonate sediment (2.5 × 10⁶ g m³; Archer, 1996), F_B is the PIC rain (mol m⁻² yr⁻¹), [CO₃ ²⁻] and [CO₃ ²⁻]_sat are deep-sea carbonate ion and deep-sea carbonate ion at saturation, respectively, and F_M is the non-carbonate flux (g m⁻² yr⁻¹), which can be calculated from observed constant sedimentary _wtCaCO₃% (B) above the CSD (Liu et al., 2022; Supplementary Table S2).

Locations of the samples and distributions of sedimentary CaCO₃ in different ocean regions are shown in Figure 2. The deposition depth of marine CaCO₃ is relatively shallow in the Northwest Pacific Ocean and the Philippine Sea. Pelagic sediments in the Northwest Pacific have 70%–80% _wtCaCO₃ at ∼2,500 m, below which _wtCaCO₃% decreases to zero at ∼4,000 m (black triangles in Figures 2A,B). In the pelagic Philippine Sea, sedimentary _wtCaCO₃% decreases with a water depth from ∼92% at ∼1,600 m, and almost no CaCO₃ can be found in sediments at depths below ∼4,100 m (filled black left triangles in Figures 2C,D), with most of sedimentary _wtCaCO₃% data between 2,000 m and 4,000 m from core-tops retrieved on the Ogasawara–Mariana and Kyushu–Palau ridges (unfilled left triangles in Figures 2C,D). The sedimentary CaCO₃ distribution in the Western Equatorial Pacific Ocean (diamonds) and the Central Pacific Ocean (circles) is presented in Figures 2E–H, with sedimentary _wtCaCO₃% decreasing below ∼3,000 m with depths to reach zero at ∼4,500 m and ∼5,200 m, respectively.

Features in observed CaCO₃-depth profiles can be largely reproduced by our model using parameters listed in Table 1. The predicted positions of CSD for different regions generally correspond to where the first systematic decrease of sedimentary _wtCaCO₃% can be observed (Figure 2). The CSD is the deepest in the Central Pacific Ocean (∼2,900 m) with a slight shoaling to ∼2,800 m in the Western Equatorial Pacific Ocean and ∼2,700 m in the Philippine Sea. The CSD is the shallowest (∼2,300 m) in the Northwest Pacific Ocean. Such results are based on measured deep-sea [CO₃ ²⁻] averaged over 300 m above the sea floor for depths between 2,500 m and 3,000 m, obtained from GLODAPv2-2020 data (Supplementary Figure S4; Olsen et al., 2020). The dissolution rate of CaCO₃ (k^*) calculated based on observed values of CCD is on the same order of magnitude among sediments from different regions of the West Pacific Ocean (Table 1), with a relatively low k^* (∼2 m yr⁻¹) for the Central Pacific Ocean, a high k^* (∼5 m yr⁻¹) for the Western Equatorial Pacific Ocean, and medium k^* (∼3 m yr⁻¹) for other regions.

The predicted CaCO₃-depth profiles (solid lines in Figure 2) from our model can reproduce the general patterns of observed sedimentary CaCO₃ distribution across different domains, despite some data points deviated from the depth-CaCO₃ profiles and uncertainties associated with the sedimentary _wtCaCO₃% data that will be addressed later. A comparison between the measured _wtCaCO₃% data and model outputs (control runs) suggests that model simulations are statistically consistent with the actual data (Supplementary Figure S7).

Intrusion of the LCDW and UCDW originated from the Southern Ocean into the Central Pacific Basin results in relatively higher dissolved [CO₃ ²⁻] (∼75 μmol kg⁻¹) below 2,000 m than other regions in the West Pacific Ocean (Supplementary Figures S3, S4). Therefore, the deep water in the Central Pacific Basin is relatively carbonate saturated so that CaCO₃ can be better preserved at depths. Aging of the UCDW and a branch of the LCDW that enters the Western Equatorial Pacific Ocean do not integrate enough with effect of POC remineralization with time in the Western Equatorial Pacific Ocean so that the regional deep-sea [CO₃ ^2-] in the deep Western Equatorial Pacific Ocean (∼74 μmol kg⁻¹) remains close to that in the Central Pacific Ocean. The deep water in the Philippine Sea is largely controlled by the UCDW since direct intrusion of the younger and deeper LCDW is restricted by the island chains, which leads to lower deep-sea [CO₃ ²⁻] (∼71 μmol kg⁻¹) than that in the deep Western Equatorial Pacific Ocean. Further north, weakening of LCDW with a volume transport of 12–14 Sv (10⁶ m³ s⁻¹) in the Central Pacific Ocean to 6 Sv in the Northwest Pacific Ocean (Kawabe and Fujio, 2010), as well as enhanced control of the North Pacific Deep Water (NPDW) in the Northwest Pacific Ocean, is in coincidence with a distinct decline of [CO₃ ²⁻] (∼68 μmol kg⁻¹) compared to other regions and shallow saturation depth of sedimentary CaCO₃ (2,420 m; Figure 2; Table 1).

To counteract the increasing T_CO2 resulted from accumulated organic matter respiration with time and compensate for the loss of [CO₃ ²⁻], carbonate dissolution takes place at shallower ocean depths as the deep water ages (Figures 2; Supplementary Figure S4). Therefore, the contrasting features of sedimentary CaCO₃ between the equatorial regions (e.g., the Western Equatorial Pacific Ocean and the Central Pacific Ocean) and the northwest regions (the Philippine Sea and the Northwest Pacific Ocean) in our study region can be attributed to patterns of the abyssal circulation. Below the carbonate saturation depth, accumulation of CaCO₃ in marine sediments is determined by the rate of carbonate dissolution (k^*) (Boudreau et al., 2020). The approach we take can help to estimate the basin-scale-averaged carbonate dissolution rate (k^*) in the West Pacific Ocean, which seems to be in concert with k^* based on water-side control (Table 1; Sulpis et al., 2018), consistent with our previous estimation of k^* in the Southwestern Atlantic Ocean and marginal seas in the western Pacific Ocean (Li et al., 2021; Liu et al., 2022). Therefore, our calculation is in support of the assertion that carbonate dissolution is mainly controlled by diffusion through the diffusion boundary layer (Boudreau et al., 2020).

With the lowest k^* (∼2 m yr⁻¹) among different regions in the West Pacific Ocean, CaCO₃ dissolves slow in under-saturated deep waters in the Central Pacific Ocean so that more CaCO₃ can get accumulated in sediment at a deeper depth. In contrast, the high k^* (∼5 m yr⁻¹) in the Western Equatorial Pacific Ocean infers fast CaCO₃ dissolution, and CaCO₃ can get accumulated in sediment at a much shallower depth compared to that in the Central Pacific Ocean, though the PIC rain in the Western Equatorial Pacific Ocean is doubling that in the Central Pacific Ocean. The rate of carbonate dissolution (k^*) due to water-side control is determined by the rate of deep current, and our derived high k^* in the Western Equatorial Pacific Ocean and low k^* in the Central Pacific Ocean are in line with high bottom current speed in the Western Equatorial Pacific Ocean and low bottom current speed in the Central Pacific Ocean (Sulpis et al., 2018).

It is worth mentioning that the uncertainty in regional [CO₃ ^2-] or observed location of CCD cannot fully account for the variability of basinal sedimentary _wtCaCO₃% distribution (Figures 2, 3). The sedimentary _wtCaCO₃% above the carbonate saturation depth in both the Western Equatorial Pacific Ocean and the Central Pacific Ocean shows strong regional variability (Figure 3), which could either be caused by different in situ PIC rain (F_B) or varied dilution intensity of the non-carbonate flux (F_M). The estimated PIC rain (F_B; Table 1) in the Western Equatorial Pacific Ocean is more than doubling that in the Central Pacific Ocean. However, the sedimentary _wtCaCO₃% above the CSD in the Western Equatorial Pacific Ocean is not higher than that in the Central Pacific Ocean, suggesting non-carbonate materials in sediments is playing an important role in determining _wtCaCO₃%.

Calculated fluxes of non-carbonate materials (F_M) are high in the Western Equatorial Pacific Ocean (∼2.4 g m⁻² yr⁻¹), which is generally in agreement with the high flux of non-carbonate materials recorded by sediment traps but remarkably higher than simulated dust deposition rates and Th-derived lithogenic flux (Supplementary Figure S6; Supplementary Table S2). Such a comparison indicates that lithogenic input alone is not sufficient to account for the carbonate dilution in the Western Equatorial Pacific Ocean. Alternatively, high fluxes of biogenic opal with spatial disparity in the Western Equatorial Pacific Ocean, and perhaps in the equatorial upwelling region in the Central Pacific Ocean as well, could have contributed to the dilution of CaCO₃ in sediment.

Lack of sedimentary _wtCaCO₃% data above CSD in both the Northwest Pacific Ocean and the Philippine Sea hinders our constraint on the flux of non-carbonate materials in both basins. We thus take the dust deposition rates directly as lithogenic flux (F_M) in the model, and our model predictions of depth-profiles of sedimentary _wtCaCO₃% can reasonably reproduce the scarce sedimentary _wtCaCO₃% data from the two regions, despite lower sedimentary _wtCaCO₃% data between 2,000 m and 4,000 m from core-tops retrieved mostly on the Ogasawara–Mariana and Kyushu–Palau ridges (unfilled symbols in Figure 2). It is worth mentioning that Th-derived lithogenic flux in the Northwest Pacific Ocean is 1–4 times higher than the rate of dust deposition (Supplementary Table S2), and such a magnitude of increase of F_M suffices to reproduce the low sedimentary _wtCaCO₃% in the pelagic Northwest Pacific Ocean between 2,500 m and 3,600 m (Figure 4).

Sediment-trap-recorded flux of non-carbonate materials in the Northwest Pacific Ocean is 6–20 times higher than the rate of dust deposition (Supplementary Table S2) due to contribution of biogenic opal (Kawahata et al., 1998). Therefore, the spatial variance in the flux of biogenic opal in the Northwest Pacific Ocean could potentially explain some of the low _wtCaCO₃% found in sediment on the Emperor Seamount Chain. However, trap-recorded flux in the oligotrophic Philippine Sea is only 1–2 times higher than the dust deposition rate (Supplementary Table S2; Kawahata et al., 1998), and such uncertainty could not explain the diluted sedimentary _wtCaCO₃% in that region (Figure 4C). To reproduce the low sedimentary _wtCaCO₃% from the Ogasawara–Mariana and Kyushu–Palau ridges, one needs 2–40 times higher flux of non-carbonate materials than the dust deposition rate, which could only be attributed to detritus derived from volcanic activities in this area.