The SCS is located between the tropical western Pacific Ocean and the Asian continent, and it is under the control of the East Asian monsoon (EAM) system. In the boreal winter, the interaction between the low-pressure system over the western Pacific warm pool and the Siberian high-pressure system triggers the formation of a strong East Asian winter monsoon (EAWM) (Figure 1). The cyclonic circulation is driven by the EAWM in the entire basin. While in boreal summer, the prevailing southwesterly monsoon (i.e., East Asian summer monsoon, EASM) drives anticyclonic circulation formation in the whole basin (Xiu et al., 2010) (Figure 1). Therefore, different surface circulation patterns in the winter and summer seasons are driven by the EAM system in the SCS. Modern physical oceanographic observations have shown that anticyclonic eddies dominate the northern part of the SCS in summer (Xiu et al., 2010; Guo et al., 2015), which drives the thermocline and halocline to deepen and become thicker (Chen et al., 2011).

The effect of the EAM system on the water hydrodynamics in the SCS in the geological periods has been investigated. Dong et al. (2015) reconstructed the thermocline depth of the southern SCS since 180 ka using the temperature difference (ΔT) between the reconstructed surface seawater temperature by utilizing U 37 K’   and the reconstructed subsurface seawater temperature by using  TEX 86 H . This indicated that the thermocline depth was shallower during the glacial period and deeper during the interglacial period due to the effect of the EAWM. The strength of the EAWM varied within a glacial-interglacial timescale, which could be indicated by loess grain size (>32 μM) from the Xifeng of China and the vertical sea surface water temperature difference in the northern and southern SCS (Guo et al., 2009). Dong et al. (2015) also found that the EAWM had significantly strengthened during the glacial period and weakened during the interglacial period since 175 ka. They also believed that the positive wind stress curl provided by the enhanced EAWM during the glacial period enhanced cyclonic circulation and strengthened upwelling, thus lifting the depth of the thermocline in the southern SCS (Dong et al., 2015). The same process has also been confirmed in modern observations (Qu et al., 2007).

Mesoscale eddies influence the marine carbon and nitrogen cycle by regulating the distribution of nutrients, in turn affecting the biota (Zehr & Kudela, 2011). As a key player in marine nitrogen cycle, the abundance and activity of AOA are affected by temperature, light, availability of ammonia, and its competition (McElroy, 1983) with phytoplankton, especially in shallow coastal waters (Herfort et al., 2007; Q. Liu et al., 2018; Urakawa et al., 2014). Some of these factors can be altered by mesoscale processes, e.g., upwelling (Molina et al., 2010), water mass mixing (Müller et al., 2018) (Haas et al., 2021), and in turn, modify the AOA community. In geological time, the physical and hydrological properties of the surface waters in the SCS have been proposed to influence the growth of shallow AOA (Dong et al., 2019), thus impacting the marine nitrogen cycle.

The [2]/[3] ratio was investigated in both the northern and southern SCS and showed a distinct variation pattern since 30 ka. In the southern SCS (Figure 2A), [2]/[3] ranged from 7.0–9.4 with a mean value of 7.9 and progressively increased in MIS 2. In MIS 1, [2]/[3] ranged from 5.9–10.6 with an average of 7.3 and showed a decreasing trend. However, it had a different pattern in the northern SCS (Figure 2B). In core MD12-3428, the [2]/[3] ranged from 5.3–6.3 with an average of 5.8 in MIS 2, while in MIS 1, [2]/[3] ranged from 6.1–7.2 with a mean value of 6.6. Similarly, in core ODP 1147 retrieved from the northern SCS, [2]/[3] ranged from 4.69 to 6.42 with a mean value of 5.75 in MIS 2, and ranged from 5.67 to 6.42 with 6.16 on average in MIS 1. Both cores in the northern SCS showed increasing trends of [2]/[3] since 30 ka, indicating that they have unique and identical regional implications under which the controlling mechanisms were different from those of the southern SCS.

The mechanisms that control the [2]/[3] and its implications had been interpreted in the southern SCS by Dong et al. (2019). The biomarker [2]/[3] reflects the abundance of Shallow-water clade Thaumarchaeota which determines the ammonia oxidation. Specifically, shallow-water clade Thaumarchaeota, with lower [2]/[3], have better ammonia concentration adaptation due to their phylogenetic distinction compared to their deep-water clade relatives (Dong et al., 2019) and take part in relatively active ammonia oxidation (Xu et al., 2018). Therefore, a lower [2]/[3] indicates higher shallow-water clade AOA abundance and a higher ammonia oxidation rate (Dong et al., 2019). In core MD05-2897, a similar variation trend between [2]/[3] and δ¹⁵N since 30 ka indicates that the coupling of ammonia oxidation and nitrogen fixation in the southern SCS (Figures 2A,C). δ¹⁵N decreased from MIS 2 to MIS 1. This indicates an enhancement of nitrogen fixation which, to some extent, is probably contributed by the consumption of oxygen due to the enhancement of ammonia oxidation indicated by the decreasing trend of [2]/[3] since the last glacial period (Figure 2C).

Considering the depth segregation of ammonia oxidation and nitrogen fixation, i.e., ammonia oxidation in the nutricline and nitrogen fixation mainly in the nutrient-depleted layer above the nutricline (Michael Beman et al., 2012; Du et al., 2017), Dong et al. (2019) concluded that the two pathways were not directly but indirectly coupled. Global marine nitrogen data-based investigation showed that sea surface temperature and subsurface minimum O₂ are two main factors that influence nitrogen fixation variations on a global scale (Luo et al., 2014). Nitrogen fixation was enhanced during anoxic events in deep geological time (Kuypers et al., 2004). The exuberant growth of ammonia-oxidizing bacteria consumes oxygen in the water column, which would facilitate the N₂-fixing enzyme activity (Zehr & Kudela, 2011; Luo et al., 2014; Dong et al., 2019). Similarly, in the northern SCS, the nitrogen fixation has increased since 30 ka indicated by decreasing δ¹⁵N value of foraminifera from core MD 12–3433 (Figure 2D). A consistent nitrogen fixation variation trend has occurred throughout the SCS since 30 ka, which might be controlled by similar physical and chemical factors, e.g., stratification. However, the increasing [2]/[3] value from the glacial to the interglacial period in cores from the northern SCS indicates the weakening of ammonia oxidation during the interglacial period (Figure 2B). The decoupling of nitrogen fixation and ammonia oxidation implies that the heterogeneity of regional surroundings between the northern and southern SCS has acted on the ammonia metabolic changes of AOA communities, and resulted in this regional differentiation within the glacial-interglacial timescale (Table 1).

The hydrodynamic processes exhibit discrepancies between the southern and northern SCS under the effect of the EAWM system. In the southern SCS, the surface circulation exhibits a cyclonic structure in winter, driven by the winter monsoon, and Coriolis force and combined with surrounding topographic constraints, causing the upwelling in the southern SCS, which brings nitrogen-limited and phosphorus-rich deep water to the upper layers (Tan & Shi, 2006; Wong et al., 2007; Ren et al., 2017). During glacial periods, the enhanced EAWM provided a positive wind stress curl that built up the cyclonic circulation and strengthened the upwelling, thus lifting the depth of the thermocline and halocline, and leading to a thinner thermocline and halocline (Chen et al., 2011; Dong et al., 2015). During the interglacial periods, the weakening of the EAWM weakened the upwelling, therefore enhancing surface stratification and shifting thermocline and DCM layer downward.

In the northern SCS, however, no significant seasonal upwelling region has been observed. Similar to that in the southern SCS, thermocline depth was deeper during the interglacial period than that in the last glacial period, but this was not because of the weakening of upwelling driven by a weaker EAWM (Figure 3A). Instead, the anticyclonic current loop, which was driven by the EASM dominating this region, drove the thermocline and halocline deeper (Xiu et al., 2010; Guo et al., 2015). Therefore, monsoon-triggered water mixing controls the surface circulation in the northern SCS: in the interglacial period, the strengthening of the EASM enhanced vertical mixing and deepened the thermocline, which, conversely, was thinner and shallower during the last glacial period in the northern SCS. Coincidentally, Li et al. proposed that vertical mixing driven by the EAWM dominantly determined the hydrological structure in the northern SCS over the last 356 ka (Li et al., 2013). Intensification of the EAWM drives stronger vertical mixing, resulting in deeper mixed layer depth (MLD) and a thicker thermocline and halocline, while the weakening of wind turbulence resulted in a shallower thermocline and halocline depth. The evidence indicates that the northern and southern SCS are controlled by different surface circulation systems and thus lead to a different hydrological structure.

Why do different hydrological structures during the glacial and interglacial periods have diverse characteristics in regard of the nitrogen cycle? Factors that affect the growth of shallow-water AOA are considered to be complex and influenced by light and the depth of the upper water column, as well as ammonia concentration and ammonia effectiveness. The AOA prefer to live in the trophic halocline where have adequate ammonia mainly produced by remineralization of organic matter and weak light intensity (Xue et al., 2004; Herfort et al., 2007; Michael Beman et al., 2012; Liu et al., 2018). In the southern SCS, during the glacial periods, the strengthening of the EAWM drove the cyclonic eddy formation and lifted the thermocline and DCM due to the upwelling. Therefore, shallow-water clade AOA communities shrank because the light was more intense in the shallower seawater. In contrast, during the interglacial period, the weakening of the EAWM led to enhanced stratification and weakened upwelling, which shifted the DCM downward. Shallow-water clade AOA, therefore, obtained a steadier water environment and space to survive with weaker light. Additionally, the ammonia concentration was also higher in the deeper layer, which in general resulted in their vigorous growth.

However, in the northern SCS, anticyclonic eddies pushed the nutrient-poor and oxygen-rich surface water down to the deeper water column during the interglacial period. An oxygenated and turbulent water environment is unfavorable for AOA survival. This low-nutrient surface-water mixing failed to stimulate primary productivity during the intergalcial period (Fig. 3B, He et al., 2008), which thus reduced the remineralization-derived carbon dioxide and ammonia for AOA growth, (Ning et al., 2008; Xiu & Chai, 2011; Hu et al., 2014; Guo et al., 2015). Moreover, enhanced export POC flux triggered by anticyclonic eddies has been observed in modern settings (Zhou et al., 2013), which means weaker remineralization in the eddies and hence, lower ammonia and carbon dioxide available to AOA (Figure 4A). In the glacial period, reduced anticyclonic mixing combined with high productivity reinforced the remineralization of fixed organic carbon (Figure 3B), therefore increasing the NH₃ and CO₂ concentration that supported the shallow-water clade AOA to growth (Figure 4B). In short, reduced anticyclonic activity promoted the stratification, which stimulated the shallow-water clade AOA growth. To some extent, the [2]/[3] ratio can indirectly reflect the relative strength of remineralization in the upper water column because remineralization of organic carbon derived from primary productivity provides both energy and carbon source for AOA, e.g., NH₃ and CO₂. Higher remineralization, therefore, leads to a higher ammonia oxidation rate, which is embodied in a lower [2]/[3] ratio (Figure 4B).

The decoupling of ammonia oxidation and nitrogen fixation in the northern SCS suggests that rather than being a determinant, ammonia oxidation can only affect the efficiency of nitrogen fixation to a certain extent. Factors that shape the nitrogen cycle can be intricate and vary between regions and timescales. The same variation trend of the glacial-interglacial nitrogen cycle in the southern and the northern of the SCS suggests that the nitrogen fixation efficiency may be controlled by the same physicochemical processes, such as thermal stratification, i.e., enhanced nitrogen fixation when the nitricline deepens, which is also consistent with modern observations (Church et al., 2009; Zehr & Kudela, 2011; Luo et al., 2014). Mesoscale physical processes (e.g., mesoscale eddies, temperature, and stratification) have an important effect on marine primary productivity and the microbial community dynamics. In addition, anticyclonic eddies have been proposed to motivate the N cycle in the ocean and enhance the N₂ fixation in the subtropical Pacific (Fong et al., 2008; Church et al., 2009) and subtropical Atlantic (McGillicuddy et al., 2007), probably because of the enhancement of diazotroph growth. In the northern SCS, anticyclonic eddies inhibit the growth of shallow-water clade AOA while enhancing the N₂-fixing bacteria during the interglacial period, indicating a decoupling between ammonia oxidation and nitrogen fixation. The processes within the N cycle, as well as their influencing factors, are subtle and intricate, and physical factors that shape the biotic community composition are not ignorable (Fong et al., 2008; Church et al., 2009). We proposed that mesoscale eddies might affect the inner cycle of the marine N cycle not even on a temporal and small spatial scale but on the longer geological timescale. Even though the mechanisms behind physical/biological interactions are not well defined (Zehr & Kudela, 2011), determining these physical processes is essential for understanding the primary production and biological responses to climate change and N cycling on a geological timescale.