Volume 49, Issue 21 e2022GL100638
Research Letter
Free Access

Carbon Sequestration of the Middle Miocene Sunda Shelf Facilitated Global Climate Change

Pengfei Ma

Corresponding Author

Pengfei Ma

State Key Laboratory of Marine Geology, Tongji University, Shanghai, China

Correspondence to:

P. Ma,

[email protected]

Contribution: Conceptualization, Methodology, ​Investigation, Writing - original draft, Writing - review & editing, Funding acquisition

Search for more papers by this author
Zhifei Liu

Zhifei Liu

State Key Laboratory of Marine Geology, Tongji University, Shanghai, China

Contribution: Conceptualization, ​Investigation, Writing - review & editing, Funding acquisition

Search for more papers by this author
Meichen Jiang

Meichen Jiang

School of Ecological and Environmental Sciences, East China Normal University, Shanghai, China

Contribution: Methodology, Software, ​Investigation, Writing - review & editing, Visualization

Search for more papers by this author
Han Cheng

Han Cheng

State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences, Beijing, China

Contribution: ​Investigation, Writing - review & editing

Search for more papers by this author
Lin Zhang

Lin Zhang

Haikou Marine Geological Survey Center, China Geological Survey, Haikou, China

Contribution: Resources, Writing - review & editing, Funding acquisition

Search for more papers by this author
Dizhu Cai

Dizhu Cai

Haikou Marine Geological Survey Center, China Geological Survey, Haikou, China

Contribution: Resources, Writing - review & editing, Funding acquisition

Search for more papers by this author
First published: 31 October 2022
Citations: 1

Abstract

Long-term organic carbon (OC) burial on continental shelves has been widely recognized for regulating atmospheric CO2 (pCO2) and the global climate. However, quantitatively assessing shelf OC burial's role and process in prominent climate transitions is challenging. Using 367 drilling sites, we evaluated the impact of OC burial on the Sunda Shelf, the world's largest tropical shelf, on the middle Miocene carbon perturbation and climate change. Comparing the Miocene Climatic Optimum (MCO) and the Middle Miocene Climate Transition (MMCT) results demonstrated that OC was buried faster during the MMCT with an increment of 0.07–0.14 × 103 GtC/Myr. This would cause an additional 34.77–69.16 ppm pCO2 sequestration, contributing at least one-sixth of the global pCO2 reduction. We found OC burial regulated the long-standing Monterey carbon isotope excursion via negative and positive feedbacks during the MCO and MMCT, respectively. Expanded terrestrial carbon reservoirs and enhanced burial efficiency were key steps in the latter positive feedback.

Key Points

  • Organic carbon (OC) burial on the Sunda Shelf is essential to the global carbon cycle during the middle Miocene

  • The OC burial rate of the Middle Miocene Climate Transition (MMCT) was faster than that of the Miocene Climatic Optimum

  • Expanded terrestrial carbon reservoirs and enhanced organic matter burial efficiency promoted the MMCT

Plain Language Summary

The organic carbon (OC) formed by the photosynthesis of marine and terrestrial organisms consumes atmospheric CO2 (pCO2). Its long-term burial in marine environments, especially on continental shelves, could influence the global carbon cycle and induce climate change. However, precisely assessing how much OC has been buried on shelves and quantifying how it affected climate change in the geological past is challenging. This is mainly due to the difficulty of obtaining complete spatial-temporal records buried deeply on shelves and the complexity of climate feedbacks involving OC burial. To fill this gap to a certain extent, we calculated the middle Miocene OC burial of the Sunda Shelf, the world's largest tropical shelf, using 367 drilling sites and evaluated its impact on the profound carbon perturbation and climate transition of this period. We found that more OC was buried during the greenhouse period, but a faster burial rate occurred after polar cooling. The accelerated OC burial on the Sunda Shelf would cause an additional 34.77–69.16 ppm pCO2 sequestration, accounting for at least one-sixth of the global pCO2 reduction of the middle Miocene. Efficient OC burial on the shelf was promoted by drainage system progradation and vegetation expansion arose along with sea level drop.

1 Introduction

The atmospheric CO2 (pCO2) concentration has coevolved with the global climate during the Cenozoic and is likely to induce future climate change (Rae et al., 2021). Throughout the geological timescales, pCO2 concentration has been a delicate balance between emissions from volcanic and metamorphic degassing and consumption by silicate weathering and organic carbon (OC) burial (Berner, 2003). Directly linked to the photosynthesis of marine and terrestrial organisms, OC burial can rapidly interact with pCO2 and contribute to long-term climate evolution via multiple feedbacks (e.g., France-Lanord & Derry, 1997; Galy et al., 2007). Moreover, enhanced OC burial has generally taken place after massive volcanic degassing and, together with it, caused dramatic carbon-cycle perturbations and profound climate transitions (e.g., Sosdian et al., 2020; Vincent & Berger, 1985).

Although continental shelves account for only 7%–10% of the global ocean area, they contribute about 80% of the oceanic OC burial and up to 50% of the OC delivered to the deep sea (Bauer et al., 2013). Therefore, quantifying continental shelves' OC burial is crucial for reconstructing the long-term carbon cycle and understanding the related complex climate feedbacks. Currently, the indirect method depending on seawater carbon isotope and mass-balance calculations has typically been employed to assess the global OC burial rates (Berner, 2003; Vincent & Berger, 1985). In comparison, relatively few studies have been done to directly calculate regional or global OC burial using stratigraphic and total organic carbon (TOC) data (Cartapanis et al., 2016; Collins et al., 2017). An essential reason for this is that accessing the complete spatial-temporal sedimentary records, especially those deeply buried in shelves, is challenging.

As the most extensive tropical shelf in the world, the Sunda Shelf is characterized by a gentle gradient and has been proved crucial for global climate evolution (Hanebuth et al., 2000; Windler et al., 2019, Figure 1). Previous works have mainly focused on its Quaternary flooding and exposure (e.g., DiNezio & Tierney, 2013; Zhong et al., 2021), as well as its potential effect on the long-term Neogene cooling trend (e.g., Collins et al., 2017; Park et al., 2020). However, precise estimates have yet to be conducted to quantify the Sunda Shelf's OC burial and its subsequent role throughout profound climate transitions because tremendous amounts of organic matter (OM) have accumulated on the shelf since the Oligocene (Chonchawalit, 1993; Leong, 2000). This shortcoming was exacerbated by the lack of scientific drilling and long continuous coring. Alternatively, an integrated estimate using commercial drilling sites widespread throughout the Sunda Shelf with acceptable age controls may be another choice compensating for the absence of high-resolution dated scientific cores.

Details are in the caption following the image

Geographic and bathymetric features of the Sunda Shelf. (a) Geographic map showing the position of the Sunda Shelf. Base images and bathymetric data are from GeoMapApp (www.geomapapp.org, Ryan et al., 2009). (b) Drilling sites (white circles, Table S1 in Supporting Information S1) distribution and their corresponding sedimentary polygons. Gray-shaded polygons are non-depositional areas distributed along the middle Miocene sedimentation boundary (Solid black line, Straume et al., 2019, Text S1 in Supporting Information S1). The middle Miocene carbonate platforms (blue brick fills, Leong, 2000; Kob & Ali, 2008; Yang et al., 2016) were expelled before polygon division.

In such circumstances, directly quantifying the middle Miocene OC burial of the Sunda Shelf and addressing its impact on global climate change would be an excellent starting point for various reasons. First, the middle Miocene is characterized by a significant transition from a prolonged greenhouse phase (the Miocene Climatic Optimum [MCO]) to an icehouse phase through an intermediate interval (the Middle Miocene Climate Transition [MMCT]) (Holbourn et al., 2014). Second, dramatic carbon cycle fluctuations during this transition, as indicated by carbon isotope and pCO2, have long been recognized as related to shelf OC burial (Vincent & Berger, 1985). Third, the Sunda Shelf has remained relatively stable since the earliest Miocene and accumulated massive but poorly quantified OC during the middle Miocene (Leong, 2000).

2 Geological Setting

The Sunda Shelf is located at the junction of the Pacific Ocean, Indian Ocean, Eurasia, and Australasia (Figure 1a). This area has acted as the continental core of Southeast Asia and has been relatively stable since the Early Mesozoic (Hall, 2012). A wide shelf similar to the present day had appeared in response to extensive transgression since the earliest Miocene after the opening of the southwestern South China Sea (Hall, 2012; Morley et al., 2016). From then on, its evolution has been closely linked to global change via eustatic sea-level fluctuations. On the one hand, sea-level ascents and descents induced by climate variation have resulted in periodical transgression-regression of the shelf (Hanebuth et al., 2000; Morley et al., 2021). On the other hand, the large-scale flooding and exposure of the Sunda Shelf have caused climate change at various scales by affecting oceanic circulation, moisture transport, silicate-carbonate weathering, photosynthetic carbon fixation, etc. (Abrams et al., 2018; DiNezio & Tierney, 2013; Sarr et al., 2019).

During the middle Miocene, the major shelf basins, including the Pattani, Malay, Penyu, West Natuna, Mekong/Cuu Long, Wan'an/Nam Con Son, and Sarawak, accumulated estuarine to marine sediments (Leong, 2000; Morley et al., 2016, Figure 1b). Mixed terrestrial and marine OM was deposited and formed thick source rocks of shales, mudstones, and coaly deposits (Collins et al., 2017; Leong, 2000). Thermal maturation indexes of OM demonstrate that these middle Miocene source rocks are predominantly immature to marginally mature (Chonchawalit, 1993; Leong, 2000). Thus, buried OC generally has not undergone large-scale migration and emission, making such strata perfect archives to preserve the original carbon storage and cycle signals during this crucial climate transition. Well-established sequence stratigraphic frameworks (Morley et al., 20112021, Figure S1 in Supporting Information S1) permit the correlation of the middle Miocene strata among different shelf basins. Two contrasting cycle assemblages dated at 16.9–14.5 and 14.5–13.5 Ma (Morley et al., 2021) correspond well with the MCO and MMCT periods, respectively (Text S2 in Supporting Information S1).

3 Materials and Methods

We reviewed all published literature regarding the middle Miocene deposition of the Sunda Shelf and compiled 367 drilling sites with specific coordinates (Figure 1). According to the stratigraphic correlation scheme (Figure S1 in Supporting Information S1), thickness features of the MMCT and MCO intervals of all drilling sites were summarized. Moreover, TOC (n = 1,355) and Rock-Eval pyrolyzable hydrocarbon (S2, n = 743) data were compiled to quantify the buried OC and classify the OM type (Table S2 in Supporting Information S1). At some but not all drilling sites, organic geochemical data were collected.

Following the regional petroleum exploration practices (e.g., Leong, 2000; Ramli, 1988), we conservatively defined the sedimentation boundary of the middle Miocene shelf with the 2 km thickness contour (Straume et al., 2019). After expelling the middle Miocene carbonate platforms (Kob & Ali, 2008; Leong, 2000; Yang et al., 2016), Thiessen polygons were generated to determine the depositional area controlled by each drilling site using XTools Pro 22 for Esri ArcGIS 10.5 (Figure 1). Here, a conservative estimate was further performed by assigning zero thicknesses to the middle Miocene strata of the boundary (Text S1 in Supporting Information S1). The mass sediment accumulation of the Sunda Shelf during the MCO or MMCT was determined by calculating each polygon's value and summarizing them via the following equation:
urn:x-wiley:00948276:media:grl65057:grl65057-math-0001
where ρsed is the average density (2,400 kg/m3) of sedimentary rocks, HX is the thickness of the MCO or MMCT strata within a polygon, and SX is the area of a specific polygon. Here, we used the average density value because the tested dry density data were unavailable, and the middle Miocene strata are generally buried deeper than 1,500 m with well-compacted sediments. The HX of each polygon was determined by the thickness values of the drilling site within it and its boundaries; the latter features were interpolated using data from adjacent drilling sites. Then, lower, median, and upper estimates were carried out using the 25%, 50%, and 75% quantiles of the HX data, respectively.
Similarly, the OC burial of the MCO and MMCT was separately estimated by calculating individual polygon values and summarizing all of them using the following equation:
urn:x-wiley:00948276:media:grl65057:grl65057-math-0002
where ρsh is the average density (2,540 kg/m3) of shale/mudstone (potential source rock), Psh is the shale/mudstone percentage, TOCi is the tested data at a certain depth, Fp is the recycled petrogenic OM fraction, Hi is the thickness controlled by the TOC data, and HXd is the total stratigraphic thickness at a specific drilling site. Psh values were compiled from drilling sites with detailed lithological features (Table S1 in Supporting Information S1). A block- or basin-scale Psh and shale/mudstone TOC data set was applied to the polygons without measured values. The 25%, 50%, and 75% quantiles of these two parameters were employed to perform the estimates and generate result ranges. (Table S3 in Supporting Information S1). The generation and redeposition of petrogenic OM always co-occur with the burial of biospheric OM in sediments (Blattmann et al., 2018; Galy et al., 2015). Therefore, the tested TOC is a mixture of the two components. However, the petrogenic OM does not lower the pCO2 of its redeposition period. As such, it is necessary to make appropriate corrections to the TOC data to exclude the effect of petrogenic OM. The Fp used here is the present value of the Sunda Shelf sediments (ca. 12%, Feng, 2022), as the middle Miocene properties are unavailable. This value represents an overestimate of the petrogenic OM contribution of the middle Miocene sediments, thus resulting in conservative estimates of biospheric OM. This is because the petrogenic OM yield is controlled by the physical erosion of river systems (Galy et al., 2015), and the largest river (the Mekong River) draining the Sunda Shelf was not fully established during the middle Miocene (Liu et al., 2017). Finally, the buried OC was directly converted into equivalent pCO2 (Text S3 in Supporting Information S1).

After determining the mass accumulation of sediments (in Gt) and OC (in GtC) for the MCO and MMCT, we calculated the corresponding rates (in Gt/Myr and GtC/Myr, respectively) with their durations (Table S3 in Supporting Information S1). Moreover, the mass and rate changes (Δ) of sediment accumulation and OC burial from the MCO to the MMCT were also evaluated (Figure 2). The OM was classified into three types, predominantly containing lacustrine algae (Type I), marine organisms (Type II), and terrestrial plants (Type III, Langford & Blanc-Valleron, 1990), based on the TOC and S2 data. In this step, coaly samples with abnormally high TOC were not considered because they were generally restricted in distribution and not representative of the primary shelf sediments.

Details are in the caption following the image

Changes in mass and rate of sediment accumulation (a and c) and organic carbon burial (b and d) of the shelf from the Miocene Climatic Optimum to the Middle Miocene Climate Transition. Features illustrated here are median estimates. For the lower and upper estimates, please see Figure S2 in Supporting Information S1.

4 Results

4.1 Sediment Accumulation

Mass sediment accumulation and the corresponding sediment accumulation rates of the MCO and MMCT periods for each polygon and the entire shelf were calculated (Figures 2 and 3a, Table S3 in Supporting Information S1). Our results demonstrated that 0–6.43 × 103 and 0–3.89 × 103 Gt sediments were accumulated in the individual polygons during the MCO and MMCT, respectively (Table S3 in Supporting Information S1). In total, more sediments were deposited during the MCO (205.36–281.72 × 103 Gt) than the MMCT (148.23–197.74 × 103 Gt). However, the opposite was true from a sediment accumulation rate perspective, with the MMCT interval buried 62.66–80.36 × 103 Gt/Myr faster than the MCO interval (Figures 2 and 3). The higher sediment accumulation rate for the colder stage of the middle Miocene was also revealed by the results calculated using seismic data sparsely distributed at the western and northern parts of the Sunda Shelf (Clift, 2006).

Details are in the caption following the image

Organic carbon (OC) burial quantification and reconstruction of the middle Miocene Sunda Shelf. (a) Mass and rates for sediment accumulation, OC burial, and pCO2 sequestration during the Miocene Climatic Optimum (MCO, shown in pink) and the Middle Miocene Climate Transition (MMCT, shown in blue). (b) Organic matter (OM) types classified by total OC and S2. Depositional model of the Sunda Shelf during (c) the MCO and (d) the MMCT. Terrestrial OM contribution significantly increased during the MMCT period. Similar results appeared in single basins (Figure S3 in Supporting Information S1).

4.2 Organic Carbon Burial

Lower, median, and upper estimates for the buried OC of the shelf were assessed (Figures 2 and 3a, Table S3 in Supporting Information S1). For single polygons, the maximum OC burial during the MCO and MMCT was 0.02–0.09 × 103 and 0.01–0.05 × 103 GtC, respectively (Table S3 in Supporting Information S1). Similar to the mass sediment accumulation, the total buried OC of the MCO (0.62–3.26 × 103 GtC, with a median estimate of 1.48 × 103 GtC) was much higher than that of the MMCT (0.33–1.50 × 103 GtC, with a median estimate of 0.73 × 103 GtC). According to Derry and France-Lanord (1996), the global net buried OC during the MCO and MMCT were about 28.37 × 103 and 11.39 × 103 GtC. Thus, the Sunda Shelf played a crucial role during this process and accounted for 2.19%–11.48% and 2.90%–13.17% of the global net OC burial during the MCO and MMCT, respectively. Nevertheless, regarding the OC burial rate, the MMCT period of the shelf once again came to dominate, allowing for an accelerated burial rate of 0.07–0.14 × 103 GtC/Myr (Figure 2 and Figure S2 in Supporting Information S1). This increment corresponds to an extra 34.77–69.16 ppm of pCO2 consumption during the MMCT (Figure 3).

4.3 Organic Matter Type

The MCO and MMCT samples with TOC and S2 data from four shelf basins (Pattani, Malay, Penyu, and Sarawak; Table S2 in Supporting Information S1) were plotted in discriminant diagrams as single basins and as a whole (Figure 3 and Figure S3 in Supporting Information S1). In all single basins, the middle Miocene OM was identified as Type II to Type III (Figure S3 in Supporting Information S1). But the slopes of the fitted lines for the MMCT samples were much lower, demonstrating more terrestrial component inputs during this period (Figure S3 in Supporting Information S1). After combining all data, it is clear that Type II and Type III OM, mainly from marine organisms and terrestrial plants, were mixed together in the middle Miocene shelf sediments. Nevertheless, the marine and terrestrial components occupied the dominant position in the mixed OM during the MCO and MMCT, respectively (Figure 3b). This result is consistent with the higher contents of alginite and bituminite closely related to marine organisms in the MCO samples of the Pattani Basin (Tanakwang, 1999).

5 Discussions

5.1 Controls on Enhanced Carbon Sequestration During the MMCT

Stratigraphy-preserved OC that modulated the climate change of the geological past accounts for only ca. 10% of the total OC delivered to sediments. Most OC was degraded and returned to the atmosphere before being efficiently buried (Middelburg, 2019). Therefore, aside from primary productivity, the sediment accumulation rate that influences the exposure time of OM to oxygen is also crucial for OC mass accumulation (Galy et al., 2007; Lee et al., 2019).

Though the TOC values of the MMCT samples were generally lower than those of the MCO samples, both at single drilling sites and at the basin scale (Table S2 in Supporting Information S1), we could not confirm that the primary productivity of the MMCT period was lower. Because higher sediment accumulation rates dilute the sediment OC contents, even if higher primary productivity occurred (Katz, 2005). Our results lead to confidence that the higher OC burial efficiency during the MMCT occurred contemporaneously with higher sediment accumulation rates and increased terrestrial OM input (Figure 3). The higher linear sedimentation rates of the Sunda Shelf were also confirmed during the transition and glacial periods of the late Quaternary (Jiwarungrueangkul et al., 2019; Zhao et al., 2017). Moreover, the spatial-temporal evolution of the terrestrial OM proportion was attributable to changes in the influence of source-area input (e.g., Feakins et al., 2020; Khim et al., 2020; Ramirez et al., 2016).

Here, we propose a conceptual model for OC burial on the Sunda Shelf. During the MCO, when the global sea level was relatively high in the greenhouse background, transgression occurred to the maximum, and the entire shelf was drowned (Figure 3c). Mixed marine and terrestrial OM deposited on the shelf. Though terrestrial OM prevailed in coastal areas, such as the most landward Pattani Basin (Figure S3 in Supporting Information S1), the marine OM was slightly predominant overall (Figure 3). Moreover, a large portion of them may have been oxidized before being efficiently buried because of the relatively low sediment accumulation rates. During the MMCT, the global sea level gradually dropped along with high-latitude cooling and then caused drainage system progradation and vegetation expansion (Figure 3d). Terrestrial carbon reservoirs distributed along low-latitude coasts, such as mangroves, have a high capacity to fix pCO2 (Collins et al., 2017; Sanders et al., 2014). The terrestrial OM could thus be more easily transferred to and more quickly buried in the shelf with higher sediment accumulation rates, allowing for the preservation of the dominant terrestrial OM in sediments and enhanced carbon sequestration during the MMCT (Figure 3).

5.2 Implications for Global Climate Change

A long-standing (ca. 3.5 Myr) positive Monterey carbon isotope excursion (MCIE) superimposed with nine isolated carbon isotope maxima (CM) events and pCO2 fluctuation on several hundred ppm magnitudes manifested a dramatic perturbation of the middle Miocene carbon cycle (Holbourn et al., 2007; Rae et al., 2021; Vincent & Berger, 1985). Regional studies and global compilations of benthic δ13C and δ18O confirmed that the MCO and MMCT were primarily encompassed in the broad MCIE and corresponded to the first seven and last two CM events, respectively (Holbourn et al., 2007; Westerhold et al., 2020, Figure 4).

Details are in the caption following the image

Proxies related to the global carbon cycle during the middle Miocene. (a) Eccentricity cycles and calculated low latitude insolation intensity (Laskar et al., 2004). (b) Global compilation for benthic δ13C and δ18O (Westerhold et al., 2020). (c) Global mean sea level reconstruction (Miller et al., 2020). (d) Planktic foraminifera (T. trilobus) B/Ca indicative of seawater pH evolution. (e) δ11B calculated pCO2 (Rae et al., 2021). (f) Lower, median, and upper estimates for pCO2 sequestration rates of the middle Miocene Sunda Shelf (this study). (g) Eruption volumes (Kasbohm & Schoene, 2018) and rates (Hooper et al., 2002) of the Columbia River Basalt Group.

Large-scale OC burial on continental shelves has long been recognized as crucial in causing the MCIE (Sosdian et al., 2020; Vincent & Berger, 1985). The original Monterey Hypothesis was raised as positive feedback composed of intensified high-latitude cooling, strengthening coastal upwelling, and lowering pCO2 by OC burial (Vincent & Berger, 1985). Extensive OC burial took place with the MCIE, as demonstrated by our results pertaining to the Sunda Shelf and other case studies of the East Pacific Ocean (Föllmi et al., 2005; Holbourn et al., 2014). However, global cooling did not occur immediately along with widespread OC burial, and the greenhouse MCO was maintained for ca. 2.5 Myr after the appearance of the MCIE with high pCO2, high dissolved inorganic carbon in seawater, relatively high sea level, and limited distribution of the Antarctic ice sheet (Holbourn et al., 2014; Miller et al., 2020; Sosdian et al., 2020, Figure 3). The original positive feedback loop cannot easily explain these paleoclimatic and paleoceanographic phenomena. Recently, Sosdian et al. (2020) proposed a revised negative feedback loop and explained the OC burial during the broad MCIE as a critical step in mitigating the abnormally high pCO2 triggered by the massive eruption of the large igneous province. This model is convincing for the MCO, considering the contemporaneous appearance of the Columbia River Basalt Group eruption and the carbon perturbation (Hooper et al., 2002; Kasbohm & Schoene, 2018, Figure 4). It should be noted, however, that this negative feedback may only make sense during the MCO, as massive eruptions only occurred in the early to middle stages of the MCO, and the MCIE of the MMCT period involved high-latitude cooling and an apparent pCO2 drop (Holbourn et al., 2014; Rae et al., 2021). Moreover, our quantitative estimates of the Sunda Shelf favored accelerated OC burial in response to global cooling (Figure 3). Similarly, enhanced opal accumulation and more OC-rich deposition also occurred in the East Pacific during the MMCT (Flower & Kennett, 1993; Föllmi et al., 2005; Holbourn et al., 2014). These studies all endorse the original Monterey Hypothesis during the climate transition.

Therefore, we argue that different climate feedbacks governed the carbon cycle perturbation of the middle Miocene against distinct climate backgrounds. As revealed by high-resolution dating and modeling, the rapid eruption of the Columbia River flood basalt and associated “cryptic” degassing were synchronous with and may cause the abrupt pCO2 raising of the MCO (Armstrong Mckay et al., 2014; Hooper et al., 2002; Kasbohm & Schoene, 2018; Longman et al., 2022, Figure 4). Global sea level raised during this greenhouse stage. Then, massive OC burial occurred on the drowned continental shelves mitigating the high pCO2 and constituting a negative feedback (Figure 3c). The reconstruction of pCO2 confirmed its gradual decrease trend after the eruption of the Columbia River flood basalt (Rae et al., 2021, Figure 3). However, the pCO2 of the late MCO remained relatively high around the average value (762.92 ppm). Following this, the MMCT ensued, responding to the obliquity-paced high-latitude cooling (Holbourn et al., 2007). The dominance of the negative feedback in the global carbon cycle gave way to another positive feedback, as the original Monterey Hypothesis proposed (Figure 3d). Subsequently, a rapid reduction in pCO2 of about 220 ppm occurred within about 1 million years, as facilitated by the increased shelf OC burial. Regarding the nine CM events, positive feedback was performed at eccentricity cycle scales (Holbourn et al., 2007; Laskar et al., 2004; Sosdian et al., 2020), but the fluctuation amplitude of δ13C in single events increased significantly when the overall positive feedback of the MMCT overlapped (Figure 4).

The temporal resolution of our study could not reveal the Sunda Shelf's role in individual CM events at eccentricity scales. Nevertheless, the Sunda Shelf was clearly pivotal in the long-standing MCIE and pCO2 evolution trend during the middle Miocene. According to our calculations, even assuming that the intensity of potential volcanic degassing did not significantly weaken during the MMCT, the accelerated OC burial on this wide tropical shelf could contribute about 16.29%–32.40% of the global pCO2 drop.

6 Conclusions

In this study, we quantified the middle Miocene OC burial of the Sunda Shelf to decipher its role in this period's profound carbon cycle perturbation and climate transition. Approximately 205.36–281.72 × 103 and 148.23–197.74 × 103 Gt sediments were accumulated during the MCO and MMCT with estimated OC of 0.62–3.26 × 103 and 0.33–1.50 × 103 GtC, respectively. The buried OC there accounted for 2.19%–11.48% and 2.90%–13.17% of the net global OC burial during the two respective stages. Comparing the OC burial rates between the MCO and the MMCT demonstrated that OC was buried faster during the climate transition by an increment of 0.07–0.14 × 103 GtC/Myr. Even assuming the volcanic degassing rate of the MMCT remained as large as that of the MCO, the accelerated OC burial on the Sunda Shelf may result in an additional 34.77–69.16 ppm pCO2 sequestration, accounting for at least one-sixth of the global pCO2 drop. We found that the long-standing MCIE could be maintained by OC burial on continental shelves via different feedbacks in distinct climate backgrounds. During the MCO, OC burial commencing on widely drawn shelves formed a negative climate feedback and mitigated the abnormally high pCO2 caused by the massive volcanic eruption. In contrast, after the high-latitude cooling and sea level drop, a positive climate feedback occurred with accelerated OC burial, facilitating further global cooling and pCO2 reduction. Expanded terrestrial carbon reservoirs and the subsequent rapid burial that were previously neglected promoted the high OC burial efficiency of the MMCT.

Acknowledgments

This work was inspired by discussions with our colleagues from the Deep-Time Digital Earth (DDE) group and was supported by the National Natural Science Foundation of China (42050102), the National Key Research and Development Program of China (2018YFE0202402), the Shanghai Science and Technology Innovation Action Plan (20590780200), and the China Geological Survey Program (ZD20220606). We thank Prof. Chao Ma (Chengdu University of Technology) for his suggestions on data analysis. We are also grateful to Prof. Sarah Feakins and the anonymous reviewers for their constructive comments.

    Conflict of Interest

    The authors declare no conflicts of interest relevant to this study.

    Data Availability Statement

    The data used in this paper can be obtained from the figshare website (available from https://doi.org/10.6084/m9.figshare.21343662.v2).