In the northwestern Pacific marginal seas, there has been a rapid temporal increase and spatial variability in the relative abundance of dissolved inorganic nitrogen over dissolved inorganic phosphate. The cause and mechanisms of this temporal and spatial variation is under debate. Recently, atmospheric deposition of nitrogen has been shown to be the major cause of the spatio-temporal variation in the concentration ratio of dissolved nitrogen and phosphate. We show that the transport by ocean currents is a more crucial factor causing the spatio-temporal variation in the ratio of dissolved nitrogen and phosphate in the study area.
 A shift in the ratio of dissolved nitrogen and phosphate (N:P ratio), resulting in N-limiting or P-limiting, can change ecological dynamics in oceans. A long-term trend of an increasing N:P ratio in the northwestern Pacific marginal seas, which comprise the East China Sea (ECS), the Yellow Sea (YS), and the East Sea (ES)/Sea of Japan, has been documented in many previous studies [Wang, 2006; Siswanto et al., 2008]. Thus, the important issue raised is the question of what is the primary driver of the change in the N:P ratio, or excess N (=N*), which is defined as N* = N – RN:P × P (where N and P are concentration of dissolved inorganic nitrogen (DIN) and dissolved inorganic phosphate (DIP), and RN:P is a constant [Wong et al., 1998; Kim et al., 2011a]).
 Most recently, Kim et al. [2011a] explained that one of the main factors in the spatio-temporal variation is the atmospheric deposition of N (hereafter, air-N), which is directly associated with human activities such as NOx emission and fertilizer use. Their conclusion relies on the temporal correlation between the air-N flux and the excess N in the seawater. The Taiwan Current (TC), the Kuroshio Current (KC), and the Tsushima Warm Current (TWC) have been considered major determinants of the water properties, including heat, salt, and nutrients, and the resulting ecosystem in the ECS and the ES [Morimotoet al., 2009; Onitsuka et al., 2007; Yanagi, 2002; Zhang et al., 2007]. However, Kim et al. [2011a] did not consider N transported by the ocean currents (hereafter, current-N) as a factor in the increase in N. If current-transport from the adjacent seas is considered a main contributor, identifying the cause of the N:P ratio shift in a receptor sea becomes more complicated. The N:P ratio shift would reflect changes in both biological/chemical (N-fixation, denitrification, and regeneration) and physical (current flow change, mixing, sedimentation, and wind-strength associated Ekman transport) processes within the adjacent provider seas. Furthermore, the effects from climate change (increase in precipitation and river discharge) and human activities (NOx emission, fertilizer consumption, land-use change, and N-saturation in watershed soil) within the watersheds providing nutrients to the marginal seas also need to be considered.
 We hereby suggest that the current-N accounts for not only a large portion of the total input of N in the northwestern Pacific marginal seas but also the spatio-temporal variability in the observed excess N in the marginal seas.
 We collected historical data for the input fluxes of N, including the current-driven input flux, the riverine discharge, and the atmospheric deposition, from previous studies (Table 1). DIN was assigned by a sum of nitrate and nitrite for seawater and river water, and by a sum of nitrate, nitrite, and ammonium for air deposition. For DIN and DIP concentrations in the northwestern Pacific marginal seas, we used the same database as that used by Kim et al. [2011a], which has been archived for the past 30 years at the Japan Meteorological Agency (http://www.data.kishou.go.jp/kaiyou/db/vessel_obs/data-report/html/ship/efile_NoS2_e.html) and at the Korea National Fisheries Research and Development Institute (NFRDI, at http://kodc.nfrdi.re.kr/page?id=obs_04_01). For comparison purposes, we also followed the same data processing procedure followed by Kim et al. [2011a]; only concentrations greater than 0.1 μM for DIN and 0.01 μM for DIP, which are the approximate analytical detection limits of N and P, were used for data analysis and RN:P of 13.06 ± 0.03 was used for N* [Kim et al., 2011a]. The areas assigned for calculation of N fluxes are 4.83 × 105 km2 for the ECS, 4.04 × 105 km2 for the YS, and 6.03 × 105 km2 for the ES.
Table 1. Comparison of the DIN Input Fluxes for Individual Paths to the ECS and the ES (unit: Tmol yr–1)
aBased on all values measured in the Korean Strait during 1999–2003 (see section 3) which were selected out of the last 30 years (1980s to 2010s in Figures 2C and 2D) for comparison with other data. To produce total flux across the strait, average annual flux during 1994–1998 was considered for the eastern channel because of no measurement in that channel since 1999.
bBased on seasonal measurements in the Korean Strait (see Figure 1) in 1999–2000 (see section 4).
cNitrate-based flux value.
dAir-N was recalculated from ~70 mmol m–2 yr–1 for the ECS, and ~60 mmol m–2 yr–1 (for DIN) for the ES.
eThe accumulated flux values over the last 30 years [Kim et al., 2011a, Table S2] were converted to average annual flux values. The area ratio of ECS to total area of ECS and YS was considered in calculation for the ECS.
 The volume transports of the TC and the intrusion of the KC to the ECS across the shelf break are 2.59 Sv (1 Sv = 106 m3 s–1) and 0.89 Sv in summer, and 1.22 Sv and 1.81 Sv in winter [Cho et al., 2009]. The intrusion of the KC to the ECS becomes weaker in summer when an eastward Ekman transport develops as a result of the southerly wind, while it becomes stronger in winter when northerly wind prevails [Zhang et al., 2007; Siswanto et al., 2008].
 The DIN influx to the ECS by the TC is estimated to be 0.27 Tmol yr–1 (1 Tmol yr–1 = 1012 mol yr–1) in summer and 0.20 Tmol yr–1 in winter, whereas by the KC, it is 0.19 Tmol yr–1 in summer and 0.43 Tmol yr–1 in winter [Zhang et al., 2007]. Several studies also estimated a similar range of DIN input fluxes to the ES by the TWC, a branch of the KC, through the Korea Strait (mean depth < 120 m) (Table 1). The TWC-transported nitrate flux into the ES was reported to be 0.2−0.4 Tmol yr–1 in 1999−2003 [Chung et al., 2000], which is similar in magnitude to the flux of nitrate transported-out from the ECS to the ES (0.37−0.54 Tmol yr–1) in the same period [Zhang et al., 2007].
 The current-N flux to the ES through the Korea Strait was also calculated by multiplying the annual mean DIN concentration and the annual mean volume transport in the western and the eastern channels of the strait, respectively. The DIN concentrations were only taken along the cross line of the Korea Strait (see the later section), and the annual mean volume transport of 2.64 Sv (with 60:40 ratio in west and east channel) was used [Cho et al., 2009]. The estimated mean DIN-flux was 0.32 Tmol yr–1 during 1999−2003 through the western channel, and 0.13 Tmol yr–1 during 1994−1998 through the eastern channel, totaling to 0.45 Tmol yr–1 throughout the entire strait.
 Riverine DIN discharge by the Changjiang River (river-N) was based on measurements at the Datong hydrological station. The mean annual river-N flux in the 2000s was 0.108 Tmol yr–1 [Dai et al., 2011], which is comparable to 0.047 Tmol yr–1 estimated by Kim et al. [2011a] based on the total load of N over the last 30 years.
 With regard to the atmospheric deposition of DIN, Kim et al. [2011a] estimated 1.7 Tmol of air-N to the ECS/YS and 0.79 Tmol to the ES, supposedly the result of 30 years of deposition. These estimates correspond to 0.031 Tmol yr–1 of flux for the ECS and 0.026 Tmol yr–1 for the ES. Air-N flux of 0.037 Tmol yr–1 in summer and 0.062 Tmol yr–1 in winter were also estimated for the ECS [Zhang et al., 2007]. Zhang et al.  analyzed the spatial distribution of the dry and wet atmospheric deposition flux of nitrate, nitrite, and ammonium species in the northwestern Pacific Ocean, including the ECS, the YS, and the ES. Their estimated flux was ~70 mmol m–2 yr–1 in the ECS and ~60 mmol m–2 yr–1 in the ES (corresponding to ~0.05 Tmol yr–1 and ~0.036 Tmol yr–1, respectively). The reported air-N fluxes for the ES were in good agreement with those reported by Kim et al. [2011a].
 The N input fluxes cited in Table 1 are based on the estimates for similar periods (the late 1990s to the mid-2000s). The river-N and the air-N fluxes to the ECS are on the same order of magnitude, consistent with previous studies [Uno et al., 2007; Nakamura et al., 2005]. According to the N-budget box-model for the ECS [Zhang et al., 2007], a major DIN input originates from the current-driven transport with the air-N and the river-N fluxes, accounting for only 6% and 10% of the total DIN input flux, respectively. Yanagi  showed that over 97% of the total N input to the ES originated from current advection through the Korea Strait.
4 Current-Transported N Flux Through the Korea Strait Based on Direct Measurements
 We estimated the dissolved inorganic nutrient fluxes passing through the Korea Strait by using the data obtained from an array of six bottom-moored acoustic Doppler current profilers across a certain section between May 1999 and March 2000 [Teague et al., 2002] and from four hydrographic surveys across the same section during the mooring period in May, June [Talley et al., 2004], and October in 1999, and March in 2000 (Figure 1A). The largest flux occurs near the coasts of Korea (station N1) with flux minima halfway between the coasts of Korea and Japan. The nitrate flux through the western channel of the Korea Strait is larger (55–71% of the total fluxes) than that through the eastern channel (on average, the ratio of the western versus the eastern channel is 60:40) (Figure 1B), which is comparable to other studies [Onitsuka et al., 2007; Morimoto et al., 2009].
 The net northward flux of nitrate through the Korea Strait shows a large seasonal variation ranging from 0.12 to 0.72 Tmol yr–1 (Figures 1B and 1C). The nitrate flux shows a maximum in October and a minimum in March, and it increased by approximately 120% from June to October 1999. The percentage of increases in the volume transport and the section-averaged nitrate concentration between June and October was 22% and 65%, respectively, indicating that the increase in the nitrate concentration contributes more to the increase in the nitrate flux.
 The seasonal variation in the nitrate flux shows an in-phase relationship with those of the volume transport and the nitrate concentration, but an out-of-phase relation with the surface salinity (Figure 1C). The maximum nitrate flux occurred in October 1999 when the surface salinity is at a minimum owing to Changjiang diluted water, while the minimum flux of nitrate occurred in March 2000 when the surface salinity was the highest. This seasonal pattern in the current-N flux was also observed in other studies [Onitsuka et al., 2007; Morimoto et al., 2009]. The increase rate of N* is also strongly correlated with the distribution of salinity in the whole area of northwestern Pacific marginal seas, as shown in next section.
 While the air-N flux ranges from 0.026 to 0.05 Tmol yr–1 for both the ECS and the ES [Kim et al., 2011a; Zhang et al., 2007, 2011], the current N influx into the ECS amounts to ~0.5 Tmol yr–1 (Table 1). Similarly, N influx of ~0.4 Tmol yr–1 enters the ES through the Korea Strait by the TWC (Table 1). These current-N fluxes for both seas are an order of magnitude greater than the air-N fluxes, consistent with previous results [Morimoto et al., 2009; Onitsuka et al., 2007; Uno et al., 2007; Yanagi, 2002; Zhang et al., 2007]. The total input of DIN to the marginal seas could be larger if other input paths are included, such as submarine groundwater discharge [Kim et al., 2011b] and the N2 fixation [Bashkin et al., 2002], although quantifying these input fluxes is difficult at this stage.
 According to recent works regarding the temporal change in N* in the ECS, increments of the annual river-N and the air-N over the last 30 years can be estimated as ~0.1 Tmol yr–1 [Dai et al., 2011] and ~0.03 Tmol yr–1 (assuming a threefold rate of increase in air-N during this period [Uno et al., 2007]), respectively (Figure 2). Thus, river-N can be an important factor in the observed increase in N* in the ECS.
 Our analysis shows that the current-N flux by the TWC has approximately doubled since the mid-1980s, which has resulted in the increase in N* (Figure 2). The increment in the annual current-N flux to the ES by the TWC was ~0.15 Tmol yr–1 over the last 30 years, which is comparable to the total increment (~0.13 Tmol yr–1) of the annual river-N and the air-N fluxes to the ECS during the same period. Therefore, the N* increase over time in the whole area of the ES is considered to have been more influenced by the N influx from the ECS by the ocean current than by the local air-N flux to the ES.
 The rate of N* increase showed large spatial variation with a relatively high rate in the ECS and along the Korean coast in the ES [Kim et al., 2011a]. Relatively high N* increase rates (Figure 3A) are observed along the path of the Changjiang River diluted water with a statistically significant correlation (R = 0.75) between the rate of N* increase and surface salinity (Figure 3B). Low-salinity in the northwestern Pacific marginal seas results mainly from the mixture of the Changjiang River discharge and surrounding seawater [Chang and Isobe, 2003; Chen et al., 1994; Wang, 2006]. A large current-N flux occurs in the Korea Strait during autumn, corresponding to the maximum Changjiang River discharge in summer (Figure 1) [Morimoto et al., 2009; Onitsuka, 2007]; this implies a considerable introduction of the river-N to the ES. Both the current-N flux and its increase rate (as well as N* increase rate) are much greater in the western channel than in the eastern channel (Figures 2C, 2D, and 4). Considering that the TWC through the western channel mainly flows along the east coast of Korea, these observations imply that the spatial pattern of the N* increase rate is likely associated with current-driven transport of N.
 N:P ratio may exceed the Redfield ratio (16:1) in the ECS and the ES in several decades if current observed trend of N enrichment continues. Some area in the ECS and its adjacent seas, showing relative N enrichment trend, have already experienced changes in phytoplankton standing stock and community composition, benthic hypoxia, and dramatic increase of harmful algae since the late 20th century [Lin et al., 2005; Li et al., 2009; Zhou et al., 2008]. Further studies to monitor the nutrient budget and their stoichiometry, and the potential impacts on ecosystem by N* change are needed.
 Transport by the major current system in the northwestern Pacific marginal seas has been shown to be the primary contributor to the temporal and spatial changes in N*. Especially, variations in N* in the ES is strongly connected with the N budget in the neighboring ECS via the ocean currents, implying that the nutrient status in the ES can be influenced by climate change as well as biological, chemical, physical processes in neighboring ECS and its bearing watershed. Multidisciplinary studies covering the ECS, the YS, and the ES are inevitable to elucidate N:P ratio change associated with human activities.
 This research was a part of the project titled “Ocean Climate Change: Analyses, Projections, Adaptation (OCCAPA)” funded by the Ministry of Land, Transport and Maritime Affairs, Korea.