Climate-Driven High Primary Production and Contrasting Export Production in the Eastern North Pacific Subtropical Gyre

The eastern North Pacific subtropical gyre (NPSG) contributes significantly to global primary production (PP) and export production (EP). In this study, we have investigated the impact of the North Pacific gyre oscillation (NPGO) mode on the temporal changes in the relationship between PP and EP in the eastern NPSG, using long-term time series of oceanographic observations at Station ALOHA. The positive NPGO phases (N2+: 1998–2004, N4+: 2007–2013), exhibiting a deeper mixed layer depth (MLD), coincided with high PP. Moreover, the N2+ phase showed high EP, associated with an increase in the nano-sized phytoplankton group, and inorganic and organic nitrogen-to-phosphorus ratios. However, multiple physical and biogeochemical factors, such as thermocline depression, increase in pico-sized phytoplankton groups, smallest-sized mesozooplankton, and heterotrophic bacteria, have induced low EP during the N4+ phase, despite deep MLD and increased PP conditions. Enhanced stratification under prolonged warming indicates that the surface eastern NPSG may experience a permanent shift toward small cells.


INTRODUCTION
The eastern region of the North Pacific subtropical gyre (NPSG), a well-known crucial open-ocean ecosystem, is characterized by basin-scale ocean circulation. Therefore, the physical characteristics of the water column and the biological productivity in the eastern NPSG can be directly and indirectly attributed to the variations in the two leading North Pacific climate modes: the Pacific decadal oscillation (PDO; which is primarily associated with the changes in the strength of the Aleutian low-pressure system) and the North Pacific gyre oscillation (NPGO; which is directly connected to the variations in the north-south dipole pattern of sea-level pressure) (Mantua et al., 1997;Di Lorenzo et al., 2008;Lukas and Santiago-Mandujano, 2008;Chhak et al., 2009;Messié and Chavez, 2011;Letelier et al., 2019) (detailed summary of the physical-biogeochemicalecological conditions associated with the PDO and NPGO can be found in Supplementary Table 1). Previous studies have reported that the PDO is closely related to the interannual variations in sea surface temperature (T) anomalies in the eastern NPSG (Miller et al., 1994;Mantua et al., 1997;Johnstone and Mantua, 2014); however, biological responses may have a weak connection with PDO (Bond et al., 2003;Di Lorenzo et al., 2008;Dave and Lozier, 2010;Saba et al., 2010). Meanwhile, recent studies have revealed that the NPGO develops favorable physical conditions for phytoplankton growth by enhancing nutrient supply to the surface layer (i.e., euphotic zone), resulting from deep mixed layer depths (MLDs) through wind forcing and lateral advection by highly saline waters (Di Lorenzo et al., 2008;Chhak et al., 2009;Dave and Lozier, 2010;Saba et al., 2010;Valencia et al., 2016). However, a few intensive and integrated studies of NPGOdriven biogeochemical factors, such as biological productivity and nutrient dynamics, have been conducted.
The primary production (PP) is the total amount of organic matter produced by phytoplankton in the surface layer, and the fraction of PP exported from the surface layer is called "export production" (EP) or "biological pump" (Falkowski et al., 2003). Owing to the large proportion of the ocean surface, the eastern NPSG contributes significantly to global PP and EP. Hence, it is crucial to understand the temporal variations in the relationship between PP and EP in the eastern NPSG (Emerson et al., 1997;Karl, 1999). Previous studies have revealed that deep carbon sequestration occurs in the eastern NPSG, indicating high efficiency of EP via low remineralization in surface waters, thereby rapidly sinking the particles into the deep (Eppley and Peterson, 1979;Scharek et al., 1999;Karl et al., 2012;Mouw et al., 2016). However, detailed information regarding temporal changes in the relationship between PP and EP and its controlling factors in the eastern NPSG is lacking. Furthermore, both climate change and anthropogenic effects are aggravating in this region, thereby creating increasingly complex and unpredictable ocean ecosystems (Barnett et al., 2005;Pörtner et al., 2019;Litzow et al., 2020).
In this study, we investigated (1) how the relationship between PP and EP has evolved in relation to the NPGO mode, and (2) what factors have changed from high PP and EP (i.e., "coupling") to high PP and low EP (i.e., "uncoupling") conditions, using long-term time series of oceanographic observations conducted at Station ALOHA (22.75 • N and 158 • W), which was established as a site chosen to be representative of the expansive eastern NPSG (Karl and Lukas, 1996).

Station ALOHA Data
We analyzed the physical and biogeochemical datasets of the eastern NPSG based on water samples obtained by the Hawaii Ocean time-series (HOT) group at Station ALOHA at approximately monthly intervals starting from January 1992 (Karl, 1999). Station ALOHA data with high temporal resolution were downloaded from the HOT-DOGS website. 1 Water samples were collected using a Niskin rosette sampler. All details regarding the data, methods, and programs are publicly available on the HOT program website. 2 To understand the response of the physical properties of eastern NPSG to NPGO, the variations in T and salinity (S) in the surface layer (here, defined as 0-100 m) and MLD during the study period were investigated. The potential T and practical S were obtained using a Seabird CTD sensor system (Karl and Lukas, 1996). The monthly average T and S data were calculated from the surface layer after removing the outliers, that is, the data points (i.e., ∼1%) falling above three standard deviations (STDs) from the mean values of climatological data over the study period (Garcia et al., 2013). To quantify the warming trend in the eastern NPSG during the study period, the linear trends of the surface and subsurface T were estimated using a simple least-squares method (Tofallis, 2009). Surface T anomalies were standardized after linear detrending to remove the warming trend. Annual T and S data were calculated by averaging the monthly mean values of each year for the surface layers. The MLD represents the depth marking an offset of 0.125 kg m −3 in the potential density relative to its value at a depth of 10 m (Huang and Russell, 1994). The potential density was estimated as a function of T and absolute S (Millero and Poisson, 1981;Fofonoff and Millard, 1983), where the latter was calculated from practical S (McDougall et al., 2012). To investigate the stratification of the water column, the mean T averaged from the subsurface layer of 100-200 m (T 100−200 ) and the 20 • C isothermal depth (Z 20 ), corresponding to the upper thermocline (Sakamoto et al., 2004;White et al., 2007), were estimated using linearly interpolated T profiles.
The biogeochemical and ecological responses of the eastern NPSG to the NPGO mode were analyzed by comparing the variability in biological (PP), geological [contents of sedimenttrap-based particulate organic carbon flux (POC 150 ), particulate organic nitrogen flux (PON 150 ), and particulate organic phosphorus flux (POP 150 )], chemical [dissolved inorganic nitrogen (DIN), dissolved inorganic phosphorus (DIP), low-level nitrate + nitrite (LLN), and low-level phosphate (LLP)], and ecological (surface phytoplankton size fractions, heterotrophic bacteria abundance, and smallest-sized mesozooplankton biomass) factors. To consider the major sources of sinking POC caught in sediment traps deployed at 150 m, we investigated biogeochemical and ecological parameters for the upper 100 m of the water column, which reflect the organic carbon produced in the upper winter MLD (December, January, and February; 81.5 ± 25.8 m) (Lampitt et al., 2008;Yoon et al., 2018).
Primary production samples were measured via the 14Cradiotracer method (Nielsen, 1952) and vertically integrated into the upper 100 m following linear interpolation of the values from sampling depths (≥4), using the values between the first sampling point from 10 m and the last sampling point within 120-180 m.
Mesozooplankton was collected using oblique tows of a plankton net of 1 m 2 (202-µm mesh netting) from the surface to 200 m (Landry et al., 2001;Landry, 2002). The dry weight biomass of the mesozooplankton was determined after thawing and oven drying (at 60 • C for at least 24 h). The smallest-sized mesozooplankton (i.e., 0.2-0.5 mm) was sorted by gently wet sieving through nested Nitex screens of 0.5 and 0.2 mm.
We estimated POC 150 , PON 150 , and POP 150 contents collected at 150 m in a free-floating sediment trap array designed to hold 12 replicate cylindrical traps . To investigate the efficiency of the "biological pump, " the EP ratio (%) was estimated as the ratio of monthly mean POC 150 to monthly mean PP (Eppley and Peterson, 1979;Dunne et al., 2007). PON 150 :POP 150 was calculated as the ratio of monthly PON 150 to monthly POP 150 . Annual ratios were calculated as the mean of the monthly ratios.
Surface DIN and DIP concentrations were determined colorimetrically using an autoanalyzer Karl and Lukas, 1996). We calculated the monthly surface ratio of DIN to DIP (i.e., DIN:DIP) from monthly mean nutrient concentrations averaged over the upper 100 m. Annual averaged DINs:DIPs were calculated by averaging the monthly mean DIN:DIP. As the surface eastern NPSG is extremely nutrient-limited, the nutrients were analyzed at nanomolar levels, that is, LLN and LLP, which are typically detected using high-sensitivity analytical techniques (Cox, 1980;Karl and Tien, 1992). The nutrient data of the surface layer were averaged after removing the outliers, that is, points lying at more than three STDs from the climatological averages (Garcia et al., 2013).

Climate Modes
The PDO index, the first leading mode of T in the North Pacific Ocean, was obtained from the University of Washington 3 (Mantua et al., 1997). The NPGO index, the second leading mode of the Northeast Pacific sea surface height (SSH), was obtained from the Georgia Institute of Technology 4 (Di Lorenzo et al., 2008).

Statistical Analysis
To remove any potential seasonal weighting of observational datasets at the ALOHA station driven by non-uniform sampling intervals, uniform physical and biogeochemical time series of monthly intervals was generated by filling the missing data points with linear interpolation using MATLAB (fillmissing function, version R2019a, MathWorks). Furthermore, monthly physical and biogeochemical parameters were analyzed with removed seasonality using MATLAB's seasonal adjustment function of an additive decomposition model (Findley et al., 1998).
To compare the physical and biogeochemical parameters between the phases of the NPGO mode, an average threshold for the study period  was applied to define the high and low values of each parameter. For statistical comparisons, a statistical concept Z-score (i.e., Z i = (x i − µ)/σ, where x i = measured value, µ = mean, and σ = STD) was calculated (Kreyszig, 2009). The (+) or (−) signs of the Z-scores indicate whether the measured values are above the mean (i.e., positive) or below the mean (i.e., negative), respectively. The sum of Z-scores was calculated by summing the Z-score values during each phase of the NPGO mode. Furthermore, a two-sample Student's t-test was applied with the null hypothesis that there was no difference in averages between the two groups, using MATLAB's ttest2 function (MathWorks, 2021).

High Primary Production and Contrasting Export Production During the Periods of Positive North Pacific Gyre Oscillation Phases
The role of the NPGO mode in driving the physical features that stimulate phytoplankton growth in the surface waters of the eastern NPSG was first investigated (Figure 1). Previous studies have reported that the positive phase of NPGO, which is associated with an intensified high-pressure system over the eastern NPSG (i.e., decrease in SSH), is characterized by cool T, whereas the negative phase exhibits the opposite (see Supplementary Table 1; Di Lorenzo et al., 2008;Chhak et al., 2009). During the study period , the surface eastern NPSG showed five phases in NPGO mode: N 1 − (1992-1997), N 2 [2007][2008][2009][2010][2011][2012][2013], and N 5 − (2014-2018) ( Figure 1A) (detailed description of other climate modes is provided in Supplementary Text 1 and Supplementary Figure 1). Moreover, the surface water at Station ALOHA shows anomalous shifts from warm to cool T and from low-to-high S conditions ( Figure 1B) while transitioning from the negative to positive NPGO (i.e., N 1 − to N 2 + and N 3 − to N 4 + ; Figure 1A), which is accompanied by increasing intensity of the trade winds (detailed descriptions of NPGO-related physical variations can be found in Supplementary Text 2) (Di Lorenzo et al., 2008;Chhak et al., 2009). Accordingly, during N 2 + and N 4 + periods, relatively cold and hypersaline conditions would contribute to an apparently deep MLD (>59.5 m) (Figure 1B), providing more favorable conditions for phytoplankton growth, compared to other NPGO phases (Saba et al., 2010). To further investigate the response of the eastern NPSG physical properties to the NPGO mode during the study period, the sums of the Z-scores of T, S, and MLD during each phase of the NPGO mode were analyzed (Supplementary Figures 2a-c). The signs of the Z-scores showed reverse patterns in physical properties between different phases of the NPGO mode, particularly during positive NPGO phases, indicating (+) signs in S and MLD (i.e., hypersaline and deep MLD), which are considered to be important drivers of increasing PP in the eastern NPSG (Di Lorenzo et al., 2008;Chhak et al., 2009;Dave and Lozier, 2010;Saba et al., 2010;Valencia et al., 2016). Therefore, to examine whether both the positive NPGO phases (N 2 + and N 4 + ), under deep MLD and hypersaline conditions, were indeed relevant to the increase in PP and EP in the eastern NPSG, compared to the average over the study period, a comparative analysis on changing relationship between PP and EP was performed by focusing on the positive phases of the NPGO.
The periods of N 2 + and N 4 + phases, exhibiting deep MLDs, coincided with high annual PP relative to the mean annual PP of 497 mg C m −2 d −1 (Figure 2A). Although they showed differences in magnitude between the N 2 + and N 4 + phases, there was no statistically significant difference between the two phases (Supplementary Table 2). The sum of Z-scores of PP during each phase of the NPGO mode also indicated distinct (+) signs (i.e., high PP) during both positive NPGO phases (Supplementary Figure 2d). However, the negative NPGO phases (N 1 − and N 3 − ), exhibiting shallow MLDs, coincided with (−) signs (i.e., low PP), except for the N 5 − phase, showing exceptionally high PP values. Error bars indicate the STDs, the dotted black line indicates the average during the study period, and bars and circles indicate annual averages exceeding (red) and lower (blue) than the study period average, respectively. POC 150 , particulate organic carbon flux; PP, primary production; EP, export production.
Meanwhile, EP estimates based on POC 150 were high (>27 mg C m −2 d −1 ) during the N 2 + phase, indicating an enhanced biological pump ( Figure 2B). This high POC 150 flux period is highly consistent with the period of the deepening MLDs and high PP, suggesting that a large proportion of the POC 150 flux was originated in the increased PP (i.e., coupling between PP and EP). However, the N 4 + phase exhibited frequent low POC 150 flux (<27 mg C m −2 d −1 ) despite the presence of deep MLDs and high PP conditions. There were statistically significant differences in the POC 150 flux estimates between the periods of N 2 + and N 4 + (p < 0.05; Supplementary Table 2 and Supplementary Figure 2e). The EP ratio (represented as POC 150 divided by PP) was also used to determine the efficiency of the biological pump in the eastern NPSG (Eppley and Peterson, 1979;Dunne et al., 2007). The average EP ratio for the 1992-2018 period was 5.7%, which is similar to the previous estimates for oligotrophic NPSG (Henson et al., 2011;Herndl and Reinthaler, 2013). As shown in Figure 2C, the frequent high EP ratios (>5.7%) during the N 2 + phase indicate an increase in the efficiency of the biological pump. However, the N 4 + phase resulted in frequent occurrences of lower values (<5.7%), indicating a significant difference with the N 2 + phase (Supplementary Table 2). This result suggests an uncoupling relationship between PP and EP during the N 4 + phase.

Factors Causing the Uncoupling Between Primary Production and Export Production During the N 4 + Phase
The uncoupling between PP and EP (i.e., high PP and low EP) during the N 4 + phase is likely due to rapid remineralization as well as slow sinking of POC in the surface waters (Marsay et al., 2015). Variations in the size of phytoplankton structure have been attributed as a key factor controlling the EP. Largesized phytoplankton cells (>2 µm) typically have high export efficiency associated with fast sinking particles, whereas smallsized phytoplankton cells (<2 µm) have low export efficiency associated with slow sinking particles (Sathyendranath et al., 2014;Basu and Mackey, 2018;Tréguer et al., 2018). Additionally, EP is regulated via food web processes, such as heterotrophic bacterial activity and zooplankton grazing (Ducklow, 2001;Steinberg et al., 2008;Herndl and Reinthaler, 2013).
The surface waters of the eastern NPSG are constantly characterized by strong stratification and nutrient depletion with an MLD shallower than the nitracline depth (Eppley et al., 1973;Dave and Lozier, 2010). Therefore, the variability of the MLD alone in the eastern NPSG has been considered to have limited influence on the change in the phytoplankton size structure (Karl and Lukas, 1996;Dave and Lozier, 2010). Fluctuations in SSH are also known to be related to the changes in biogeochemical parameters (Uz et al., 2001). However, a recent study reported that PP and EP at Station ALOHA did not correlate with the SSH anomaly, which is associated with the vertical displacement of nutrients (Barone et al., 2019). This highlights that PP and EP variabilities are not related to changes in the position of the pycnocline, but to the injection of nutrients into the euphotic zone associated with the isotherm (or isopycnal) uplift (Sakamoto et al., 2004;Barone et al., 2019). Variations in the thermocline in the North Pacific, which is located near the nitracline, can influence the nutrient flux changes from deep waters to surface waters, resulting in decreased nutrient supply following the deepening of the thermocline and increased nutrient supply following the shallowing of the thermocline (Karl, 1999;Miller et al., 2004;Sakamoto et al., 2004).
Therefore, to investigate the physical and biogeochemical factors that change from high PP and EP (i.e., "coupling") during the N 2 + phase to high PP and low EP (i.e., "uncoupling") conditions during the N 4 + phase, the N 2 + vs. N 4 + environmental conditions were compared using the variability of the Z 20 representing the upper thermocline layer, pigment-based phytoplankton-sized groups, surface DIN:DIP ratios, sedimenttrap-based PON 150 :POP 150 ratios, surface heterotrophic bacteria abundance, surface low-level nutrients, and the biomass of the smallest-sized mesozooplankton (i.e., 0.2-0.5 mm). First, Z 20 is significantly deeper (>168.5 m) in the eastern NPSG during the N 4 + phase than during the N 2 + phase ( Figure 3A and Supplementary Table 2). This suggests that a relatively shallow thermocline and deep MLD may have contributed to the high EP ratio during the N 2 + phase, likely facilitating the vertical transport of nutrients from within the thermocline to the nutrient-depleted surface waters. However, during the N 4 + phase, a low EP ratio may have been caused by the thermocline depression even under a deep MLD of the eastern NPSG, resulting in a decreased upward nutrient flux with the thermocline being relatively far from the MLD (Karl, 1999;Miller et al., 2004;Sakamoto et al., 2004).
Second, the pigment-based phytoplankton-sized groups (i.e., three different phytoplankton-sized groups; pico-sized group: 0.2-2 µm, nano-sized group: 2-20 µm, and micro-sized group: 20-200 µm) also vary in magnitude between the N 2 + and N 4 + phases ( Figure 3B). The mean fraction of the picosized phytoplankton group was increased by ∼10% during the N 4 + phase, whereas that of the nano-sized phytoplankton group was decreased by ∼10% relative to the N 2 + phase. In addition, the cell abundance of the pico-sized phytoplankton Prochlorococcus (as determined by dual-beam flow cytometry) was significantly higher (>1.8 × 10 5 cells ml −1 ) during the N 4 + phase than during the N 2 + phase (Supplementary Figure 3 and Supplementary Table 2). Low levels of Prochlorococcus during the N 2 + phase were comparable with the findings from a previous study that showed a decreased abundance of Prochlorococcus was associated with an increase in vertical gradients of inorganic nutrients during the low SSH anomaly, linked to positive NPGO (Barone et al., 2019). Although pico-sized phytoplankton contributes significantly to POC export through aggregationdisaggregation processes (Close et al., 2013), these results suggest that the enhanced EP ratio during the N 2 + phase may have been primarily derived from the increase in the relative biomass of the nano-sized phytoplankton group, resulting in the relatively rapid and deep sinking of POC (Figures 2C, 3B). This relationship between the EP ratio and nano-sized phytoplankton supports the emerging important role of nano-sized phytoplankton groups in biological pump functioning in the Northeast Pacific (Juranek et al., 2020). One particular reason for the low EP ratio during the N 4 + phase may be the significantly increased contribution of the pico-sized phytoplankton group to the POC 150 flux with a simultaneous decrease in the nano-sized phytoplankton group.
Third, the surface DIN:DIP and PON 150 :POP 150 ratios increased continuously during the N 2 + phase (DIN:DIP, r 2 = 0.42, p < 0.05; PON 150 :POP 150 , r 2 = 0.18, p < 0.05; Figure 3C), which coincided with an increasing trend in the nano-sized phytoplankton biomass (r 2 = 0.26, p < 0.05; Figure 3B). There was no evident trend in the pico-sized phytoplankton biomass (Supplementary Figure 4). In contrast, decreasing trends in the surface DIN:DIP (r 2 = 0.27, p < 0.05) and PON 150 :POP 150 (r 2 = 0.36, p < 0.05) ratios were observed during the N 4 + phase (Figure 3C), accompanied by an increasing trend in pico-sized phytoplankton biomass (r 2 = 0.11, p < 0.05; Figure 3B); however, there was no significant trend in nano-sized phytoplankton biomass (Supplementary Figure 4). These results are consistent with previous findings that the biochemical compositions of phytoplankton cells (i.e., carbohydrates, proteins, and lipids contents) are dependent on the water-column nitrogen-to-phosphorus (N:P) ratio, thereby affecting the changes in the size of phytoplankton (Morris et al., 1974;Takagi et al., 2000;Kim et al., 2020). A largesized phytoplankton group typically has a high protein content under high N:P conditions, whereas a small-sized phytoplankton group has high carbohydrate and lipid contents under low N:P conditions. Phytoplankton pigment composition may also vary not only with changes in the phytoplankton community structure, but also with light quality-related changes (i.e., photoadaptation) (Karl and Church, 2014;Letelier et al., 2017). Previous studies reported that higher phytoplankton pigments in the winter season and around the subsurface occurred based on light adaption (Winn et al., 1995;Karl and Church, 2014;Letelier et al., 2017;Karl et al., 2021). However, the DIN:DIP ratio and pigment data for both the upper part (0-45 m) and lower part (70-100 m) of the surface layer showed temporal coherence (Supplementary Figure 5), indicating that changes in pigment composition and nutrients during N 2 + and N 4 + phases were not influenced by photoadaptation. Thus, decreasing trends in N:P ratios during the N 4 + phase may have caused an increase in the pico-sized phytoplankton group, leading to low EP. Fourth, during the N 2 + phase, the high EP ratio ( Figure 2C) was associated with the overall low abundance of surface heterotrophic bacteria (<4.6 × 10 5 cells ml −1 , as determined by dual-beam flow cytometry) (Figure 3D and Supplementary Table 2), implying the low recycling of POC in the upper waters, thereby accompanied by low concentrations of remineralized nutrients (LLN < 4.5 nM and LLP < 52 nM) ( Figure 3E and Supplementary Table 2). In contrast, the low EP ratio in the N 4 + phase ( Figure 2C) coincides with the relatively high abundance of surface heterotrophic bacteria (>4.6 × 10 5 cells ml −1 ) and high levels of LLN (>4.5 nM) and LLP (>52 nM) (Supplementary Table 2 and Figures 3D,E), thereby indicating a high dependence on bacterial remineralization rather than POC export.
Fifth, the biomass of smallest-sized mesozooplankton ( Figure 3F) varied considerably with that of pico-sized phytoplankton ( Figure 3B and Supplementary Figure 3), indicating the direct and indirect dependence of energy transfer to zooplankton on phytoplankton size (Pomeroy, 1974;Azam et al., 1983;Laws et al., 2000). The relatively high contribution of nano-sized phytoplankton during the N 2 + phase was associated with a decrease in the biomass of the smallest-sized mesozooplankton group (<0.20 g dry weight m −2 ); however, the high contribution of pico-sized phytoplankton during the N 4 + phase caused an increase in the biomass of the smallest-sized mesozooplankton group (>0.20 g dry weight m −2 ). This is consistent with previous studies suggesting that small-sized zooplankton feed efficiently on small-sized phytoplankton and that organic material is immediately recycled in surface waters (Sommer et al., 2002;Finkel et al., 2009).
Overall, the results suggest that multiple physical and biogeochemical factors, such as the depression of the thermocline, shift toward the pico-sized phytoplankton group, and increase in smallest-sized mesozooplankton and heterotrophic bacteria, would have induced rapid recycling of organic matter, the grazing pressure increase in the surface water, and subsequent decrease in EP during the N 4 + phases, despite the conditions of deep MLD and increased PP.
A recent study reported that euphotic depth-integrated PP exhibited a long-term increasing trend over the 30-year Station ALOHA observation period, which was most likely caused by increasing atmospheric deposition of pollutant nutrients (Karl et al., 2021). Our results also showed high PP values during N 5 − , even under negative NPGO phases (Supplementary Figure 2d). However, the sum of Z-scores of EP ratio during each phase of the negative NPGO mode indicated a weak (−) sign (i.e., low EP ratio) during N 5 − phase (Supplementary Figure 6), even under high PP and POC flux conditions (Supplementary Figures 2d,e), but weak (+) signs (i.e., high EP ratio) during N 1 − and N 3 − phases (Supplementary Figure 6). A slight decrease in EP ratio during N 5 − was related to the relatively deep Z 20 , a slight decrease in the fraction of nano-sized phytoplankton, and an increase in heterotrophic bacteria, smallest mesozooplankton, LLN, and LLP concentrations, compared to that during N 1 − phase under a relatively high EP ratio condition (Supplementary Figure 6). To understand whether this pattern was loosely modulated by natural climate variability due to accelerating anthropogenic disturbances (Barnett et al., 2005;Pörtner et al., 2019;Litzow et al., 2020), further observation of biogeochemical parameters at the ALOHA station is needed.
On the other hand, the T 100−200 exhibited an increase (r 2 = 0.3, p < 0.05) throughout the study period (Supplementary Figure 7), implying a gradual decline in nutrient supply to the surface due to enhanced stratification. Accordingly, the EP ratio displayed a significant decreasing trend over the 30-year observation period, probably leading to higher regenerated PP and lower EP due to intensive remineralization processes, despite the conditions of increasing PP and atmospheric deposition (Karl et al., 2021). Therefore, under prolonged warming, the strengthening trend of the subsurface thermal stratification suggests that, even with the positive NPGO phase, the future surface eastern NPSG may likely experience a permanent shift toward small cells, accompanied by strong remineralization via the microbial loop. These shifts may induce a low EP ratio, which represents the low efficiency of the biological pump in the eastern NPSG, eventually contributing less to the global carbon cycle (Figure 4). Therefore, analyzing the changes in the future environmental and ecological conditions of the surface eastern NPSG to climate modes under accelerating anthropogenic effects is significant.

CONCLUSION AND SUMMARY
In this study, the impact of positive NPGO phases (N 2 + : 1998-2004 and N 4 + : 2007-2013), with deep MLD and increased PP conditions, on the temporal changes in the relationship between PP and EP was investigated using long-term timeseries observations of the surface physical and biogeochemical factors in the eastern NPSG. During the N 2 + phase, deepened MLD induced high PP and EP, associated with an increase in the nano-sized phytoplankton group and N:P ratios. However, even under deepening MLD and high PP conditions, the N 4 + phase was characterized by a low EP. Complex and multiple factors, such as deepening thermocline, increase in the picosized phytoplankton group, smallest-sized mesozooplankton, and heterotrophic bacteria abundance, and a decrease in N:P ratios, have induced an uncoupling feature (i.e., high PP and low EP). Thus, under high stratification induced by continuous warming, the results suggest a shift of the surface ecosystem in the eastern NPSG toward small cells, leading to a reduction in the biological pump. Additional observational data are required to improve our understanding and analyses of future changes in the environmental and ecological responses of the eastern NPSG to climate modes under accelerating climate change.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
J-EY and I-NK developed the idea, interpreted the data, and wrote the manuscript. J-EY conducted the literature survey and analysis. All authors discussed and approved the final manuscript.

ACKNOWLEDGMENTS
We are sincerely grateful to the editor and the two reviewers. We thank all those who have contributed to the HOT program since 1988.