Estimating Denitrification Rates in the East/Japan Sea Using Extended Optimum Multi-Parameter Analysis

Denitrification rates in the East/Japan Sea (EJS) were examined with extended Optimum Multi-Parameter (eOMP) analysis. The potential denitrification locations expected from the eOMP analysis occurred only in the Ulleung Basin (UB) and near the Tatar Strait (TtS) of the Eastern Japan Basin (EJB). Estimated denitrification rates were ~0.3 3 and ~4 11 μmol N m-2 d-1 in the UB and in the EJB, respectively. These rates agree with previous published results. The EJS’s rates were lower than reported for other marginal seas. However, considering the rapid EJS response to climate change, we predict that denitrification may be enhanced in the near future.


INTRODUCTION
The East/Japan Sea (EJS) is a semi-enclosed marginal sea in the northwestern North Pacific Ocean that exhibits many dynamic characteristics, e.g., deep-water formation, subpolar front, eddy, and gyre circulation (Talley et al. 2006).The EJS is often referred to as a "Miniature Ocean" (Kim et al. 2001).The EJS is also well known as an optimal location to study modern climate change due to the short hundred year time scale residence time (Kim and Kim 1996;Kim et al. 2001).
Denitrification (NO 3 -→ NO 2 -→ N 2 O/N 2 ), which occurs in low oxygen environments, is the most significant mechanism responsible for nitrogen loss in the ocean environment (Brandes et al. 2007).Denitrification reduces the bioavailable nutrient N/P ratio because nitrate serves as an oxidant instead of O 2 (Anderson and Sarmiento 1994).Denitrification has not been considered a significant process in the EJS nitrogen cycle because of high oxygen concentrations, which would not be conducive to conventional denitrification in this region.The EJS uniquely shows that dissolved oxygen concentrations are greater than 220 μmol kg -1 at 1000 m depth in the North Pacific Ocean (see Fig. 1 of Talley et al. 2006).
A number of recent studies have indicated that radical changes in the EJS are induced by climate change.Examples of these trends include: increases in air and sea temperatures (Gamo 1999;Min and Kim 2006), deepening of the oxygen minimum layer (Kim and Kim 1996;Chen et al. 1999;Kim et al. 2001;Kang et al. 2004, Kim et al. 2004), a change in the deep water formation mode (Gamo et al. 2001;Kang et al. 2003;Chae et al. 2005), and a prediction that anoxia may occur during the next century (Chen et al. 1999;Kang et al. 2004).Both indirect and direct evidence suggest the possibility for denitrification in the EJS (Talley et al. 2001;Yanagi 2002;Lee et al. 2007;Tishchenko et al. 2007;Jeong et al. 2009).
Several lines of evidence indicating denitrification signals on the continental slopes of the Ulleung Basin (UB) and the Eastern Japan Basin (EJB) were recently suggested by intensive nutrient profile analysis.Denitrification rates were estimated to be ~3 -33 μmol N m -2 d -1 (Kim et al. 2012).Two hypotheses were presented to explain denitrification in the EJS: (1) the formation of mirco-reducing environments in the bottom waters, and (2) aerobic denitrification as a newly discovered process.
At present, verification of the respective hypotheses is difficult for the EJS, but examining denitrification rates with alternate approaches may provide mechanistic insights.
The extended Optimum Multi-Parameter (eOMP) analysis estimated denitrification rates from coastal to marginal to open ocean scales (Hupe and Karstensen 2000;Kim and Min 2013;Kim et al. 2013).The main goals of this study are: (1) to investigate potential denitrification locations, and (2) to estimate denitrification rates using the eOMP analysis.

DATA
Two basin-wide hydrographic cruises were conducted in the summer of 1999 in the EJS via the Circulation Research of East Asian Marginal Seas II (CREAMS II) program (Fig. 1).More specific information on the cruise is available in Talley et al. (2004).The data used for this study were obtained from http://sam.ucsd.edu/.The parameters used for the eOMP analysis are latitude, longitude, pressure, potential temperature (PT), salinity (S), dissolved oxygen (DO), nitrate (N), phosphate (P), silicate (Si), total alkalinity (TALK), and dissolved inorganic carbon (DIC).The DIC values were calculated from total alkalinity and total pH through the CO2SYS program (Van Heuven et al. 2011).Only data below 300 m were analyzed to minimize seasonal variation influences.The EJS basin is divided into the Japan Basin (JB), the UB, and the Yamato Basin (YB).The JB is divided into the Western Japan Basin (WJB) and the EJB with a boundary along 135°E, for more convenient analysis presentation (Fig. 1).

THE EOMP ANALYSIS
In general, the oceanic measurements are determined by physical mixing and biogeochemical processes together (Anderson and Sarmiento 1994).The eOMP analysis provides one approach to consider physical mixing and biogeochemical processes together.Since then basic OMP analysis was developed (Tomczak and Large 1989), and improved and named the eOMP analysis (Poole and Tomczak 1999;Hupe and Karstensen 2000).The Si and TALK equations in the eOMP analysis were recently modified (Kim et al. 2013).The eOMP analysis matrix from is given as: (1) where the matrix A is defined as the physicochemical characteristics of the source water types, column X is composed of mixing ratios (x i ) among the source water masses, the ( ) ( )  amount of remineralized phosphate (ΔP remi ), the amount of denitrification (ΔN deni ), the amount of inorganic carbonate dissolution (ΔC inorg ), and the amount of inorganic silicate dissolution (ΔSi inorg ).The column b is the observations, and column R is the residuals of parameters.The last row that consists of '1' is to constrain the mass conservation (∑x i = 1).The ratios of r : : : : indicate the Redfield ratios, and r D is the amount of phosphate produced by denitrification and is given as 1/104 (Gruber and Sarmiento 1997).The eOMP analysis is based on an over-determined system, meaning that there are more equations than unknown variables, so the solutions (x i , ΔP remi , ΔN deni , ΔC inorg , and ΔSi inorg ) are found by minimizing residuals with non-negativity (Tomczak and Large 1989).Here, the amount of denitrification (ΔN deni ) among biogeochemical changes is the primary focus in this study.

Physicochemical Characteristics of Source Water Types
The physicochemical characteristics of eight different water masses in the EJS were defined using geographical locations and water properties including temperature, salinity, and dissolved oxygen: (1) North Korea Surface Water (NKSW; surface water), ( 2 (Kim and Lee 2004;Kim et al. 2010a).The source water types (1) -( 4) are surface water, (5) -( 7) are intermediate water, and ( 7) -( 8) are deep water.Because only data collected deeper than 300 m is considered, only TMW, LCW, ESIW, and ESPW were used in the eOMP analysis (Table 1).
Since each parameter shows different accuracy and precision, the eOMP analysis assigns a weight value for each parameter.The weight equation is given as (Kim and Lee 2004;Kim and Min 2013;Kim et al. 2013): where j v is the standard deviation of parameter j calculated from the physicochemical characteristics of the source water types, and accuracy j is the measurement error of parameter j.The weight values used for the eOMP analysis are summarized in Table 1.Note that the weight of DIC was assigned with the same value as that of TALK, because the DIC was estimated from TALK and the total pH.

Residuals of Mass Conservation
The last column consisting of '1' in Eq. (1) constrains the mass conservation ( ).This constraint serves as a tracer to examine the eOMP analysis validation results (Tomczak and Large 1989;Hupe and Karstensen 2000).The residuals of mass conservation (R MC ) are calculated as follows: (%) R x 1 100 Overall the residuals (cases 1 -4) were defined within ~3%.

Potential Denitrification Locations Expected
The eOMP analysis suggested that the eight locations (ΔN deni > 0), intersected collectively from cases 1 -4, were locations where bottom water denitrification may occur in EJS (Fig. 2).Six locations (1 -6) were in the UB, and two stations (7 -8) were in the EJB.Overall the denitrification spatial distribution estimated from the eOMP analysis was similar to that from the nitrate profile analysis (Kim et al. 2012), and the estimated denitrification is higher in the EJB than in the UB.The mean ΔN deni values of the eight locations, averaged from the cases 1 -4, ranged from about 51 ± 1 to 530 ± 330 μmol m -3 at the bottom layer (Table 2).
The rates of bottom water denitrification are estimated for the eight locations following the section 4.2.

Basin
Table 2. Mean magnitude and rates of denitrification estimated at eight potential denitrification locations expected from the eOMP analysis.

Comparisons with Other Marginal Seas
Based on previous results (Kim et al. 2012) and this study, the EJS's denitrification rates range from 0.2 -33 (0.3 ± 0.1 to 26 ± 7) μmol N m -2 d -1 .The EJS's rates are comparable, but generally low, relative to those in other marginal seas.The denitrification rates estimated in the Adriatic Sea, the Arabian Sea, the Baltic Sea, the Black Sea, the North Sea, and the Okhotsk Sea ranged from 19 -1151 μmol N m -2 d -1 (Degobbis et al. 1986)

Implication for a Climate Feedback Loop in the EJS
Although the mechanism driving denitrification in the EJS is not yet understood, we speculate that EJS denitrification are enhanced by the rapid responses of EJS to climate change -for examples, increase in sea surface temperature, slowdown of deep-water formation system, and in turn a decrease in oxygen content in the deep/bottom waters (Kim and Kim 1996;Gamo 1999;Chen et al. 1999;Gamo et al. 2001;Kim et al. 2001Kim et al. , 2004;;Kang et al. 2003Kang et al. , 2004;;Chae et al. 2005).If these changes occur continuously in the EJS, the oxygenated deep/bottom waters will be changed progressively to low oxygen waters, a condition favoring denitrification (NO 3 -→ NO 2 -→ N 2 O/N 2 ).Nitrous oxide (N 2 O), a strong greenhouse gas (Solomon et al. 2007), is produced during the denitrification process.As a result, we expect that a positive feedback loop may be formed in the EJS in the future (Fig. 5), if climate change patterns continue or increase in the EJS environment.

SUMMARY AND FUTURE STUDY
The main purpose of this study was to estimate denitrification rates in the EJS using the eOMP analysis and compare the results with those estimated previously using nitrate profile analysis (Kim et al. 2012).The results indicate that the potential denitrification locations expected from the eOMP analysis were located in UB and EJB.Mean denitrification rates were estimated as ~0.3 -3 and ~4 -11 μmol N m -2 d -1 in the UB and the EJB, respectively.The rates are consistent with those estimated by the nitrate profile analysis.However, big questions remain about denitrification in the EJS, such as (1) "Where exactly do the denitrification signals originate (bottom water vs. sediment)?"and (2) "What mechanisms drive the denitrification process?".At present, direct evidence is not yet available to elucidate these questions.We suggest that a synthesized survey, including O 2 , nutrients (NO 2 -, NO 3 -, and PO 4 3-), isotopes (δ 15 N and δ 18 O), and molecular information, is needed to identify denitrifying bacteria at the denitrification locations predicted in the EJS.
Considering the rapid responses of EJS to climate change, denitrification may be enhanced in the near future.Therefore, a denitrification study will be an important key to understanding the EJS's nitrogen cycle, and indicating changes expected in EJS deep/bottom environments.

Fig. 1 .
Fig. 1.Map showing the hydrographic stations of 1999 Summer Circulation Research of East Asian Marginal Seas II (CREAMS II) program and topography of the East/Japan Sea (EJS).The EJS is consisted of three basins; JB (Japan basin), YB (Yamato Basin), and UB (Ulleung Basin), one rise; YR (Yamato Rise), and four straits; KR (Korea Strait), TS (Tsugaru Strait), ST (Soya Strait), and TtS (Tatar Strait).The Japan Basin at 135°E was divided into the Western Japan Basin (WJB) and the Eastern Japan Basin (EJB) to facilitate analysis.

Fig. 3 .
Fig.3.An illustration to estimate denitrification rate.It is assumed that denitrification occurs within the triangular area (= 1/2 × ΔN deni × ΔH).The time information (ΔT) between the upper and the lower boundary is estimated from the relative age concept.

Fig. 5 .
Fig. 5.An expected loop of climate feedback associated with denitrification in the EJS.

Table 1 .
The physicochemical characteristics of source water types and weights used for the eOMP analysis in the EJS.