CC BY-NC-ND
0Contents lists available at ScienceDirect
Ecological Modelling
journal homepage www.elsevier.com/locate/ecolmodel
Interannual variability of the phytoplankton community by the changes in vertical mixing and atmospheric deposition in the Ulleung Basin, East Sea: A modelling study
Soonmi Leea b, Sinjae Yoo
a Jeju International Marine Science Center for Research and Education, Korea Institute of Ocean Science & Technology, Gujwa Iljudongro 2670, Jeju, South Korea
b Ocean Science and Technology School, Korea Maritime and Ocean University/Korea Institute of Ocean Science & Technology Joint Program, Busan 606-791, South Korea
ARTICLE INFO
Article history: Received 2 July 2015 Received in revised form 21 November 2015 Accepted 25 November 2015
Keywords:
Phytoplankton functional types ERSEM
Winter vertical mixing Atmospheric deposition East Sea
ABSTRACT
The East Sea (Japan Sea) ecosystem has experienced a significant warming and ever-increasing anthropogenic atmospheric deposition of nitrogen during recent decades. To understand the impacts of such environmental changes on the planktonic community, we set up a zero-dimensional European Regional Seas Ecosystem Model (ERSEM) in the Ulleung Basin, East Sea for the years 2001-2012. The model results show that as the winter maximum mixed layer depth (MMLD) changes, the growth and grazing loss of phytoplankton functional types (PFTs) are affected differently, resulting in differential success of PFTs in the upper mixed layer. Diatoms pre-empted the early spring growth by better utilization of light and nitrate. Diatoms' advantages lessened as the MMLD decreased. Flagellates and picophytoplankton showed mixed responses to decreased MMLD. Their net primary productivity (NPP) and peak biomass decreased but their annual biomass increased due to decreased grazing. Dinoflagellates always did better when MMLD decreased. The model results also indicate that with an increase in atmospheric deposition, the picophytoplankton and the flagellates increased in summer, whereas the dinoflagellates and the diatoms decreased. For the study period, the atmospheric deposition in the Ulleung Basin increased the annual net primary production by 4.58% (mean; range 3.77-10.58%). Biological variables showed the largest responses in summer with high year-to-year variability. Picophytoplankton increased the most (summer increase mean: 23.23%; summer increase range: 9.12-42.6%) while dinoflagellates decreased the most (summer decrease mean: -2.33%; summer decrease range: -9.09 to 10.13%). The changes in flagellates and diatoms were much less. Taking the results together, it is likely that as the warming and atmospheric deposition continue to intensify into the future; the phytoplankton community in the region will shift to smaller phytoplankton with consequent changes of food web structure to follow.
© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
The composition of the phytoplankton community plays an important role in controlling marine biogeochemical cycling and primary production (Doney et al., 2002; Falkowski et al., 2003). The phytoplankton groups have different nutrient requirements and utilization methods (Falkowski et al., 2004; Litchman et al., 2007; Jennings et al., 2008; Finkel et al., 2009) and are grazed by
* Corresponding author at: Jeju International Marine Science Center for Research and Education, Korea Institute of Ocean Science & Technology, Gujwa Iljudongro 2670, Jeju, South Korea. Tel.: +82 64 798 6070; fax: +82 64 798 6085. E-mail address: sjyoo@kiost.ac.kr (S. Yoo).
different zooplankton groups (Sterner and Elser, 2002). For example, diatoms transport carbon to the deep ocean through rapid sinking and have a great effect on carbon export production. Their frustules are an important determinant of the silicon cycle (Smetacek, 1999). Diatoms are also an important food web element supporting large fish populations. Therefore, it is important to predict present and future changes in phytoplankton community composition to understand how the marine ecosystem is influenced by environmental changes such as warming and anthropogenic forcing.
In recent decades, climate warming and the increase in anthropogenic nutrient inputs to surface waters have been major issues in the oceans. Recent models have begun to include many different phytoplankton functional types (PFTs) and zooplankton
http://dx.doi.org/10.1016/j.ecolmodel.2015.11.012
0304-3800/© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Fig. 1. Map of the study area with bathymetric contours. The black circle indicates 104-09 station (37° N, 130.6°E) of the KODC. The rectangular box shows the model domain. The shaded arrows denote major currents of the area—TWC: Tsushima Warm Current, EKWC: East Korean Warm Current.
functional types (ZFTs) in an attempt to understand the effect of these marine environmental changes on the marine ecosystem and biogeochemical cycling. Several studies have suggested that microphytoplankton were reduced when the mixed layer depth (MLD) became shallower in the north-western Mediterranean basin (Auger et al., 2014) and large phytoplankton such as diatoms might decrease as the vertical mixing is attenuated by climate warming (Lewandowska et al., 2014). The atmospheric nitrogen inputs entering the open ocean could increase the primary production and the effects of increasing atmospheric deposition are expected to continue to grow in the future (Duce et al., 2008).
The East Sea (Sea of Japan), surrounded by the Korea Peninsula, Japan and Russia, is a semi-enclosed marginal sea of the northwestern Pacific (Fig. 1). It has some features characteristic of the ocean, such as thermohaline circulation, western boundary current, meso-scale eddies and coastal upwelling, so it is sometimes referred to as a "miniature ocean" (Ichiye, 1984). The East Sea has also undergone two major marine environmental changes over the last several decades and can serve as a model case to study the impacts of the two major decadal changes mentioned above.
Firstly, the East Sea has been in a warming trend in the upper 1000 metres since the late 1980s (Kim et al., 2001). Like other mid-latitude regions, vertical mixing is a dominant source of nutrient supply in the East Sea (Yentsch, 1990; Dugdale et al., 1992; Onitsuka and Yanagi, 2005). If the vertical mixing in the middle latitudes decreased through climate warming, it would limit the nutrient supply and decrease the total phytoplankton near the surface (Doney, 2006). Secondly, the N* (the relative abundance of N over P on the basis of the N:P ratio, 13:1) has steadily increased in the surface waters since the 1980s (Kim et al., 2011b). Kim et al. (2011b) suggested that the atmospheric deposition by Asian NOx emission related to rapid industrialization might have caused the increase in N*. The increase in anthropogenic nitrogen inputs to surface waters might bring about changes in the N:P ratio and eventually influence the marine ecosystem.
Despite the evident decadal environmental changes, the responses of the East Sea ecosystem have not been adequately studied. Most studies have focused on the spatial and temporal variability in the biomass of total phytoplankton or some groups mainly due to the lack of long-term observation of phytoplank-ton community composition. These studies put forward hypotheses about the key physical processes that underlie the variability (Kim et al., 2000, 2007; Yamada et al., 2004, 2005; Chiba et al., 2008). The lack of detailed observation has prompted the use of ecosystem models to test hypotheses about the marine environmental changes affecting the phytoplankton community composition. Nutrient-phytoplankton-zooplankton-detritus (NPZD)-type models have been used in some studies to understand the seasonal variations of the phytoplankton blooms and nutrients dynamics (Onitsuka and Yanagi, 2005; Onitsuka et al., 2007, 2009). These models have only one nutrient component, nitrogen, and one or two phytoplankton groups, so they cannot take account of the complex dynamics of PFTs and ZFTs.
In this study, we used a zero-dimensional European Regional Seas Ecosystem Model (ERSEM) whose structure includes four nutrient elements, four PFTs and three ZFTs to understand the dynamics of the phytoplankton community clearly and to look more deeply into the biological processes. The key questions that we address are: (1) how do the changes in vertical mixing and atmospheric deposition alter the phytoplankton community composition and the primary production? (2) what is the relative contribution of the major sources to the annual nutrient supply?
2. Data and methods
2.1. Data sources
In order to investigate the interannual variability of PFTs, the station 104-09 (37.0°N, 130.6°E) of the Korea Ocean Data Center (KODC), located in the middle of the Ulleung Basin, was selected as a representative site of the model domain (Fig. 1). At this station, there are various available data including long-term oceanographic
Table 1
Data sources for model input or validation.
Data provider
Description
Chlorophyll a concentration
HPLC pigment Temperature Salinity Cloud cover
Nutrients for vertical mixing
Ocean Biology Processing Group, National
Aeronautics and Space Administration (NASA)
Korea Institute of Ocean Sciences & Technology
Korea Ocean Data Center
Korea Ocean Data Center
Korea Meteorological Administration
Korea Institute of Ocean Sciences & Technology
Nutrients for atmospheric deposition Literature reviews
Point (37.0°N, 130.6°E), 2001-2012, monthly reconstructed data using Ocv6 algorithm
Points (36.5-37.3°N, 130.1-131.5°E), 2000-2010, survey data Point (37.0°N, 130.6°E), 2001-2012, 0-400 m bimonthly survey data Point (37.0°N, 130.6°e), 2001-2012, 0-400 m bimonthly survey data Point (37.5°N, 130.9°e), 2001-2012, daily survey data Points (36.5-37.3°N, 130.1-131.5°E), 2000-2010, 0-1500m survey data
Uno et al. (2007); Zhang et al. (2011); KMA (2013)
data (since 1965) observed bimonthly by the KODC. Table 1 is an overview of the data sources used. The nutrient concentrations and the high-performance liquid chromatography (HPLC) pigment data were obtained from the Korea Institute of Ocean Sciences & Technology (KIOST) surveys conducted by Eardo in Nov. 2000, Apr. 2001, Oct. 2001, Sep. 2002, Jul. 2005, Apr. 2006, Aug. 2007, Feb. 2008, Oct. 2008 and Jul. 2010 in the Ulleung Basin. To verify the model outputs, the surface chlorophyll a concentration (37.0°N, 130.6°E) from merged satellite data of Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and Moderate-resolution Imaging Spectroradiometer (MODIS)-Aqua was used. The data merge was done using the Ocean Color (OC) v6 algorithm (http://oceancolor. gsfc.nasa.gov/REPROCESSlNG/R2009/ocv6).
2.1.1. Light
To calculate the surface photosynthetically available radiation (PAR) (Fig. 2a), the daily cloud cover monitored at Ulleung island (37.5°N, 130.9°E) was obtained from the Korea Meteorological Administration (KMA). The cloud cover data ranged from 1 to 10 and was divided by 10 to calculate the fractional cloud cover. With the fractional cloud cover, the surface PAR was calculated using an astronomical equation (Rosati and Miyakoda, 1988) in the ERSEM.
2.1.2. Mixed layer depth
In the Ulleung Basin, the development of the mixed layer is affected by not only convection and wind forcing from the sea surface but also horizontal heat flux from the Tsushima Warm Current (TWC) (Onitsuka et al., 2007). Thus, observational data were utilized to derive the MLD instead of MLD simulated by a one-dimensional model. The MLD was reproduced from bimonthly KODC data of temperature and salinity from 2001 to 2012 by variable density threshold method (Sprintall and Tomczak, 1992). This method allows for temperature and salinity stratification and defines the MLD as the depth where the increase in density from the surface reference depth, 10 m, equals an increase in density equivalent to a decrease in temperature of 0.2°C (Lim et al., 2012). The monthly MLD data was also calculated from the bimonthly data according to the linear interpolation method (Fig. 2a). Finally, the maximum MLD from December to February for the period 2001-2012 (MMLD) was calculated for each year to assess the interannual variability of winter vertical mixing in the Ulleung Basin.
2.1.3. Nutrients
To estimate the nutrient concentrations below the MLD, the fitted curves from 0 m to 1500 m were obtained by non-linear method (Fig. 2b). The equations for nitrate, phosphate and silicate concentrations at the boundary are written as:
Nitrate = 26.1620 - 25.5451 • e(-0 00359 MLD) (1)
Phosphate = 2.1805 - 2.09409 • e(-a00331-MLD) (2)
Silicate = 90.6642 - 88.2544 • e(-0 00108 MLD) (3)
During this study period, the N:P ratio from the observational data including the surface data was 12.2:1 (Fig. 2c). However, the N:P ratio in the Ulleung Basin is generally known as 13:1 (Talley et al., 2004; Kim and Kim, 2013), so the nitrate concentrations at the boundary in the Ulleung Basin were calculated by N:P ratio, 13:1, instead of the observed nitrate concentration.
2.1.4. Atmospheric deposition
To investigate the effect of atmospheric deposition on PFTs in the Ulleung Basin, the slope of nitrate deposition, 0.007, was obtained from Uno et al. (2007) using the linear regression method (Fig. 2d). The nitrate deposition was estimated using the values compiled by Zhang et al. (2011)(their Table S2). The ammonium deposition was
Table 2
Assignment of CHEMTAX groups to PFTs.
Algal groups Typical cell size PFTs
Diatoms 20-200 |xm Diatoms
Prymnesiophytes 2-20 |xm Flagellates
Pelagophytes
Cryptophytes
Cyanobacteria 0.2-2 |xm Picophytoplankton
Chlorophytes
Dinoflagellates 20-200 |xm Dinoflagellates
set to the equivalent value to the nitrate deposition according to the ratio from KMA (2013).
2.1.5. Net primary production
The primary production model by Kameda and Ishizaka (2005), a type of vertically generalized production model (VGPM), was used for net primary production (NPP) in the Ulleung Basin during this study period. This algorithm is as follows:
PPeu = 0.66125 .pBpt • £ +04 1 • Chl • Dirr .Zeu (4)
where, PPeu is the daily integrated primary production in the euphotic depth (mgCm-2 d-1), ?Bpt is the optimal photosynthesis rate (mg C (mg Chl)-1 h-1), E0 is the PAR at the surface (E m-2 d-1), Chl is the chlorophyll a concentration, Dirr is the photoperiod (h) and Zeu is the euphotic depth (m). But in this study, MLD, M(t) was used instead ofZeu to describe the NPP in the mixed layer. ?Bpt in the above model was a function for both temperature and chlorophyll a concentration and was described as:
pB _ 0.071T - 3.2 • 10-3T2 +3.0 •10-5T3
opt" Chl
+ (1.0 + 0.17T - 2.5 -10-3T2 -8.0 •10-5T3 (5)
where, Tis sea surface temperature (SST, °C).
2.1.6. Analysis of phytoplankton pigments
The HPLC pigment data were processed by CHEMTAX, a matrix factorization program. The class-specific pigment ratios for each diagnosticbiomarker, modified by Lee et al. (2011), were used in the CHEMTAX calculation. Seven algal groups estimated on the basis of the pigment contents were assigned to one of the PFTs (Table 2).
2.2. Model description
2.2.1. Physical aspects
The seasonal pattern of water column structure in the Ulleung Basin was more or less regularly repeated, despite the temperature fluctuating widely between years (data not shown). The water column structure was typically stratified into two layers during warm seasons and became practically one layer above the permanent thermocline after late autumn. Thus, we used a simple two-layer model to represent the physical structure in the Ulleung Basin. We used Evans and Parslow's (1985) formulation to describe the changes in vertical mixing. We followed the biological activity in the upper box and assumed the lower box to be the nutrient pool, which remains constant all year round. With a deepening mixed layer after summer, nutrient is supplied into the upper layer by two processes: entrainment and diffusive mixing across the thermo-cline. The function h(t) = dM(t)/dt was used to calculate the time rate of the change in MLD, M(t) and the variable h+(t) = max(h(t), 0) was utilized to calculate the dilution of phytoplankton (Fasham, 1993). It is assumed that the phytoplankton are detrained with a shoaling mixed layer. On the other hand, the function h(t) was used for zooplankton, because the zooplankton become more concentrated
Fig. 2. (a) Monthly mean of MLD and surface PAR, (b) Vertical gradient of the nitrate, phosphate and silicate concentration interpolated from in situ data. White circles indicate the mean of nutrient concentrations with depth. Black circles indicate the seasonal means of nutrient concentrations (W: Winter, S: Spring, M: Summer, A: Autumn). Bars denote ±95% confidence intervals, (c) Correlation between nitrate and phosphate concentration in the Ulleung Basin (r = 0.98). The solid line indicates N:P ratio from the observational data (12.2:1). The dashed line indicates Redfield's ratio (16:1), (d) Estimated atmospheric deposition in the Ulleung Basin (white circles: nitrate, black circles: nitrate + ammonium). Red asterisks are nitrate atmospheric deposition from Uno et al. (2007) and Zhang et al. (2011). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)..
Fig. 3. A schematic diagram of the ERSEM indicating the carbon/nutrient pathways between functional groups (4 phytoplankton, 3 zooplankton and 1 bacteria). Numbers indicate relative prey availability for each consumer.
due to their swimming capability within the mixed layer. Diffusive mixing across the thermocline, m was parameterized by a constant factor. In the model, 1.2 m d-1 was chosen for m, because there are various physical processes such as eddies, storm events and TWC advection in the Ulleung Basin (Onitsuka and Yanagi, 2005). The whole diffusion term, K is finally given by
m + h+(t)
And the light intensity in MLD, /mid is expressed as:
Wd = My *> • e-keM(t)dz
where, /0 is the irradiance at the sea surface (Wm-2) and ke is the light extinction coefficient.
The atmospheric deposition term is given by:
M(t) • Nmass
Dn3rn M(t)Nmass
(8) (9)
where, An3 is the nitrate flux by atmospheric deposition (mmolNm-3d- 1), Dn3 is the atmospheric deposition of nitrate (mgNm-2d-1) and Nmass is the molecular weight (gmol-1) of nitrogen. The deposition ratio of ammonium to nitrate, rn was used to estimate the ammonium flux by atmospheric deposition, An4. It was set to 1.0, which was assumed for the composition of atmospheric deposition.
2.2.2. Biogeochemical model parameter setting
The biogeochemical model is based on ERSEM-2004 (Blackford et al., 2004), which is a state-of-the-art marine biogeochemical model connecting carbon and nutrient fluxes with functional-type dynamics. The ERSEM has been proven capable of simulating the
marine ecosystem response to climate change in waters of various trophic states. An important feature of the ERSEM is the decoupling between carbon and nutrient dynamics through variable cell quotas. This facilitates the simulation of variable stoichiometry and the evaluation of the nutrient limitation effect. These features are very suitable for the purpose of this study. Fig. 3 shows a schematic diagram of the pelagic ERSEM. A detailed description and equations of this model can be found in Blackford et al. (2004).
Primary producers and consumers consisted of four and three size-based functional types (diatoms 20-200 |m, flagellates 2-20 |im, picophytoplankton 0.2-2 |im, dinoflagellates 20-200 | m, heterotrophic nanoflagellates 2-20 | m, microzooplankton 20-200 |im, mesozooplankton >200 |im), while the decomposer is modelled through one functional type. The functional types in the ERSEM are described by the same physiological and population processes but different parameterizations. The growth of PFTs partly depends on their light affinity to acclimatize to variable ambient light intensity. The response of PFTs to nutrient conditions is also determined by their competitive ability, which is quantifiable by nutrient affinity. For example, diatoms have the maximum photosynthesis rate and a high capacity to tolerate low light but have a lower competitive ability about nutrients (Falkowski, 1980). Therefore, parameters for PFTs can be adjusted in accordance with their strategies.
For parameter values, we started from the standard set used in Blackford et al. (2004). We performed sensitivity analysis for all parameters using increments or decrements of ±10% and ±30%. Based on the sensitivity analysis, we fine-tuned parameters. The specific parameters for the pelagic ecosystem components are given in Tables 3 and 4. The selected values, except for several parameters, are within the range of previous studies (Broekhuizen et al., 1995; Baretta-Bekker et al., 1995, 1997; Blackford et al., 2004; Vichi and Masina, 2009). The initial slope of the P-I curve and the maximum chlorophyll-to-carbon cell ratio were taken
Table 3
Model parameters for PFTs. P1: Diatoms, P2: Flagellates, P3: Picophytoplankton, P4: Dinoflagellates. Parameter notation follows Blackford et al. (2004).
Description Notation Unit Range P1 P2 P3 P4
Maximum assimilation rate (10 °C) Tass d-1 0.35-6.98 3.7 2.5 2.7 1.5a
Initial slope of P-I curve a mgCmgChl-1 (Wm-2)-1 d-1 0.56-9.62 2.98a 2.11 2.40 0.98
Exudation under nut. stress Pex - 0.05-0.2 0.15 0.15 0.15 0.15
Activity respiration ractr - 0.1-0.25 0.15 0.15 0.15 0.15
Maximum chlorophyll-to-carbon cell ratio $max mgChlmgC-1 7.0E-3-7.8E-2 2.0E-2 2.0E-2 2.0E-2 2.0E-2
Minimum chlorophyll-to-carbon cell ratio ^min mgChlmgC-1 7.0E-3-5.0E-2 1.5E-2 1.5E-2 1.5E-2 1.5E-2
Minimal N/C ratio QPmin mmol N (mgC)-1 3.78E-3-6.87E-3 3.78E-3 3.78E-3 3.78E-3 3.78E-3
Minimal p/c ratio qpPmin mmol P (mgC)-1 1.965E-5-4.29E-4 3.54E-4 2.36E-4 1.97E-4 3.54E-4
Affinity for nitrate ano3 (mgC)-1 m-3d-1 Set-2.5 E-2 3.5E-3 2.5E-3a 2.0E-3 3.0E-3
Affinity for ammonium anh4 (mgc)-1 m-3 d-1 2.5 E-3-2.5 E-1 4.0E-3 5.0E-2 7.0E-2 2.5E-3
Affinity for phosphate ap (mgC)-1 m-3 d-1 2.5 E-3-2.5 E-1 4.0E-3 5.0E-3 7.0E-3 3.5E-3
Maximal silicate-to-carbon ratio qPSirdf mmol Si (mgC)-1 1.0E-2-3.0E-2 1.0E-2
a Same as in Blackford et al. (2004).
from literature reviews (Geider et al., 1997; MacIntyre et al., 2002). The affinities for nutrient were adjusted according to nutrient utilization strategies (Wafar et al., 2004; Litchman et al., 2007; Litchman and Klausmeier, 2008). Finally, the food web structure was modified from the standard configuration (Blackford et al., 2004) to reflect the ecosystem conditions in the Ulleung Basin (Fig. 3). The numbers beside the arrows in Fig. 3 indicate the percentage of food availability of the particular trophic level.
3. Results
3.1. Overall performance of the model
Fig. 2a shows the monthly mean time series of MLD and surface PAR at station 104-09 in the Ulleung Basin for the period 2001-2012. The MLD shows a large interannual variability. The MMLD ranged from 29.8 m in February, 2010 to 193.6 m in February, 2006. The range of minimum MLD during summer was from 10.3 m in July, 2005 to 14.2 m in August, 2006. In comparison with the MLD, the surface PAR showed a smaller interannual variability. The maximum surface PAR ranged from 127.4 Wm-2 in June, 2006 to 151.3 Wm-2 in June, 2002. The minimum surface PAR in winter ranged from 28.5 Wm-2 in December, 2011 to 33.7 Wm-2 in January, 2001.
We first examine the interannual variability of chlorophyll a concentration as a bulk indicator of phytoplankton dynamics. The simulated time series of chlorophyll a concentration is in good agreement with the merged satellite data from SeaWiFS and MODIS-Aqua (r = 0.69, p <0.001; Fig. 4a). In the model, the peak magnitude of chlorophyll a concentration in winter/spring showed a remarkable interannual variability, ranging from 0.84mgCm-3 to 1.67 mg Cm-3. The highest peak magnitude of chlorophyll a concentration appeared in April, 2004, whereas the lowest peak magnitude of chlorophyll a concentration appeared in February, 2010.
When the monthly mean time series of the model chlorophyll a concentration was compared with the satellite data, the best matched year was 2004 and the worst matched year was 2012 (Fig. 4b). On the whole, the model chlorophyll a concentration tends to overestimate the satellite chlorophyll a concentration due to an early initiation of spring bloom in the model (Fig. 4c). There are certain shortcomings of this model set-up that could produce errors. First of all, the MLD time series used was based on bimonthly observations. This could not adequately reproduce the MLD change in a short timescale. As a zero-dimensional model, it cannot deal with event-scale turbulence or non-constant vertical structure of chlorophyll a concentration. The study region is also under the influence of TWC and warm anticyclonic eddies that could affect the nutrient dynamics (Kim et al., 2011a). This model is not adequate to deal with such horizontal or vertical processes.
The PFT composition from model outputs was compared to the phytoplankton community composition from HPLC data, which are too few and uneven to derive confident seasonal means (Fig. 5). Nevertheless, the PFT composition in spring was reproduced well by the model, with a domination of diatoms. However, the model underestimated the picophytoplankton composition in summer and the flagellate composition in autumn, and overestimated the diatom and dinoflagellate composition in summer and autumn.
3.2. Seasonal cycle
The model recapitulates typical seasonal cycles of ecosystem variables in the study area (Fig. 6). The standard deviation of MLD increased steadily from August when the monthly mean of MLD was the shallowest. The standard deviation was the largest in February after which the MLD decreases again. The nitrate concentration was the highest in February and the lowest in July. The nitrate concentration in February showed considerable fluctuation in accordance with the change in winter MLD. Unlike the nitrate concentration, the ammonium concentration showed the highest
Table 4
Model parameters for ZFTs and bacteria. Z4: Mesozooplankton, Z5: Microzooplankton, Z6: Heterotrophic nanoflagellates, B1: Bacteria. Parameter notation follows Blackford et al. (2004).
Description Notation Unit Range Z4 Z5 Z6 B1
Maximum assimilation rate (10 °C) Tass d-1 0.5-10 0.8 2.0 3.0
Food concentration where relative uptake is 0.5 h mgCm-3 40-300 60 40 60
Assimilation efficiency ae - 0.2-0.6 0.4 0.5a 0.4a 0.4a
Lower threshold (mgCm-3) for feeding Zminfood mgCm-3 1-100 5 20 30
Excreted fraction of uptake eu - 0.30-0.55 0.55 0.40 0.30
Temperature independent. Mortality rate rmort d-1 0.02-0.05 0.02 0.05 0.05 0.05
Fraction of excretion going to DOM Pdom - 0.5-1.0 0.5a 0.7 1.0
Q10 value Q10 - 2.0-3.0 2.0a 2.0a 2.0a 2.95
a Same as in Blackford et al. (2004).
Fig. 4. (a) Simulated total chlorophyll a concentration (black circles) as compared with satellite chlorophyll a concentration (grey shade. r=0.69, p < 0.001), (b) Taylor diagram for evaluation of the model results. The red circle (total) indicates the statistic values for the whole period (r= 0.69), (c) Difference between the model and satellite chlorophyll a concentration. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
Fig. 5. Comparison of PFT composition by observation and model. Colours denote the PFTs (red: diatoms, blue: flagellates, green: picophytoplankton, violet: dinoflagellates). Number of data is shown in the lower right corner of each panel. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
Time (monthly) Time (monthly)
Fig. 6. Seasonal cycle of MLD observation, chlorophyll a concentration (grey: satellite, golden brown: model), simulated nitrogen (red: nitrate, blue: ammonium, green: nitrate + ammonium), NPP (grey: satellite, golden brown: model), simulated PFTs (red: diatoms, blue: flagellates, green: picophytoplankton, violet: dinoflagellates) and simulated ZFTs (red: mesozooplankton, blue: microzooplankton, green: heterotrophic nanoflagellates) in the Ulleung Basin. The shaded areas indicate standard deviation. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
concentration in March and the fluctuation range was narrower than that of the nitrate concentration. The model is not able to describe the characteristic of the Ulleung Basin where the nitrate drawdown is observed during warm seasons (Fig. 6), because the diffusive mixing across the thermocline was parameterized by a constant factor.
Like the satellite data, the magnitude and the standard deviation of chlorophyll a concentration and NPP in the model were the largest in April. The model result also showed that each group of PFTs and ZFTs exhibited a different seasonal cycle. Diatoms were the earliest growing group in the model and generally dominated in the cold season. They were also the most variable group among the PFTs in spring. The flagellates and the picophytoplankton bloomed in late spring, but their peak magnitudes were smaller and the fluctuation ranges were narrower than those of diatoms. The seasonal succession progressed in the order of diatoms, picophytoplankton, flagellates and dinoflagellates in the Ulleung Basin. ZFTs also increased following the blooms of PFTs according to prey availability (Fig. 3).
3.3. The responses of phytoplankton functional types to the changes in winter vertical mixing
To investigate the causes of the interannual variability of ecosystem variables, we examine the relationship between MMLD and the annual mean of various plankton variables (Fig. 7). The annual means of all the PFT NPP and peak magnitude except for dinoflag-ellates showed an increasing trend whether statistically significant or not. The peak magnitude of diatoms responded most sensitively to the changes in MMLD among the PFTs (r =0.81, p <0.01). The annual mean of total phytoplankton NPP and peak magnitude showed a significant relationship with the changes in MMLD
(NPP r = 0.65, p < 0.05; peak magnitude r = 0.87, p < 0.001). However, the annual mean of total phytoplankton biomass did not show a statistically significant relationship with MMLD (r =0.40, p >0.05). This is because only the diatoms showed an increasing trend while the other three groups showed a decreasing trend, which results in a less significant relationship. Only the annual means of diatom biomass and flagellate biomass were statistically significant (diatoms r = 0.66, p < 0.05; flagellates r = -0.58, p < 0.05).The annual means of flagellate NPP and picophytoplankton NPP increased but the annual mean biomass decreased because of increased grazing as MMLD increased. However, the annual mean of dinoflagellate NPP and the annual mean biomass decreased as MMLD increased. This difference in response among the PTFs can be explained by competitive abilities for light, nutrients and grazing (see Section 4). The annual mean of zooplankton biomass did not respond either to the changes in MMLD possibly due to their complicated food web and swimming capability.
To see the impacts of varying winter vertical mixing clearly, we compared the two extreme years, 2006 and 2010, which respectively had the deepest and shallowest winter vertical mixing during this study period (Fig. 8). In 2006, deep vertical mixing maintained low biomass of phytoplankton and zooplankton in winter, but the phytoplankton groups, except for dinoflagellates, increased fast in spring. In 2010, on the other hand, both phytoplankton and zooplankton maintained high biomass in winter.The temporal changes and the seasonal cycle of phytoplankton and zooplankton were less conspicuous. More importantly, the composition of PFTs in 2006 and 2010 was very different. The diatoms in the spring of 2006 had the largest proportion of PFTs (71%), and they made the biggest contribution (59%) to the annual mean of total phytoplankton biomass. On the other hand, in the spring of 2010, the proportion of flagellates and picophytoplankton was higher than in 2006 (67%), and
Fig. 7. Relationship between biological variables (annual mean of NPP, peak magnitude of phytoplankton biomass, annual mean of plankton biomass) and MMLD. The correlation coefficient with p-value is shown in the lower right corner of each figure.
they made the biggest contribution (52%) to the annual mean of total phytoplankton biomass.
3.4. The responses of phytoplankton functional types to atmospheric deposition
In the next experiments, we calculated the percentage increase of seasonal and annual means with and without atmospheric deposition to investigate the effect of the atmospheric deposition on biological variables (Fig. 9, data not shown). Two patterns are noteworthy: the percentage increase of ammonium concentration was higher than that of nitrate concentration (2.5-fold). Specifically, it was much higher in summer (24.0-fold). The fluctuation range was also wider than that of nitrate concentration. The percentage increase of biological variables with atmospheric deposition was the most variable in summer (median range -7.86 to 23.23%) among the seasons. The percentage increase of picophytoplankton biomass was the highest (annual median 6.47%) among the PFTs and it reached 42.60% in the summer of 2012. On the other hand, the percentage increase of dinoflagellate biomass was the lowest (annual median -3.20%) among the PFTs. In particular, the atmospheric deposition had a negative effect on the diatom biomass and the dinoflagellate biomass in summer. The diatom biomass decreased by 15.57% in the summer of 2012 and the dinoflagellate biomass decreased by 9.09% in the summer of 2010. The percentage increase of microzooplankton biomass was the highest (annual median 4.65%) among the ZFTs, but was smaller than that of pico-phytoplankton biomass. The percentage increase of NPP was more than twice that of chlorophyll a concentration and total phyto-plankton biomass in the model. The NPP increased up to 32.32% in the summer of 2010, and the percentage increases of annual mean of NPP reached 10.58% in 2010.
We performed two additional simulation experiments to supplement the scanty observational data. First, we investigated the effect of different deposition ratios of ammonium to nitrate. The response of PFTs was followed by changes in the ratio from 0.1 to 2 (Fig. 10). With the increase in ammonium deposition, the pico-phytoplankton biomass and the flagellate biomass increased nearly linearly, but the diatom biomass and the dinoflagellate biomass decreased. The percentage change in diatom biomass became negative from the ratio 0.30 and that of dinoflagellate biomass became negative from the ratio 0.75 and up. Although the percentage increase of dinoflagellate biomass was high when the deposition ratio of ammonium to nitrate was small, it became less than that of picophytoplankton and flagellate biomass when the ratio increased above 0.2 and 0.5, respectively. Among the PFTs, the percentage increase of picophytoplankton biomass was the highest (max 33.08%).
In the second experiment, we increased the atmospheric deposition in the range 2.15-42.47mgNm-2 d-1 to see the effect of the increase in anthropogenic atmospheric deposition (Fig. 11). The percentage increase of N* and that of the N:P ratio increased continuously through the range. The NPP showed the highest increase among biological variables, up to 31.11%. Except for dinoflagellates and flagellates, all biological variables increased nearly linearly at low atmospheric deposition and levelled off at higher atmospheric deposition. The percentage increase of picophytoplankton biomass was the highest (max 30.71%) among the PFTs, whereas the percentage increase of dinoflagellate biomass was the lowest (min -17.83%) among the PFTs. Interestingly, the percentage increases of flagellate biomass drastically decreased from the point where the atmospheric deposition increased 10-fold. In ZFTs, the percentage increase of microzooplankton biomass was the highest (max
Fig. 8. Composition of phytoplankton biomass (upper panel, red: diatoms, blue: flagellates, green: picophytoplankton, violet: dinoflagellates) and zooplankton biomass (lower panel, red: mesozooplankton, blue: microzooplankton, green: heterotrophic nanoflagellates) by seasons. The size of the circle for seasons indicates relative biomass and the size of the circle for annual is normalized to the larger of the two. The percentage in the upper right corner of each figure indicates season composition (%). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
20.94%). All ZFTs increased nearly linearly at low atmospheric deposition and levelled off at higher atmospheric deposition.
4. Discussion
4.1. Effects of the changes in vertical mixing
We used the zero-dimensional model in combination with long-term observation data to look into the biological processes responding to the changes in winter vertical mixing and atmospheric deposition in the Ulleung Basin. Overall, the model recapitulates the basic pattern of phytoplankton seasonal cycles and its interannual variability in the upper layer of the Ulleung Basin (Figs. 4a and 6). As the winter vertical mixing increases, so does the annual mean of NPP, peak magnitude and biomass (Fig. 7). The opposite trend results when the winter vertical mixing decreases. Generally, the phytoplankton biomass is limited by deep mixing in winter. The deep mixing in winter drags down the phytoplankton into the deeper layer where the light is limited. The phytoplankton are not able to grow in this circumstance, even though the deep mixing injects abundant nutrients up to the upper layer. The deep mixing in winter also causes greater dilution of zooplankton abundance and lessens the grazing pressure. Consequently, the grazing of zooplankton is not able to catch up with the growth of phytoplankton, so the phytoplankton are able to bloom much faster in spring. However, when MMLD becomes shallower, the light conditions for photosynthesis improve in winter.
The nutrients are depleted earlier and the grazing by zooplankton increases faster as a consequence of phytoplankton growth in winter.
In this light, we examine how the light and the nutrient are affected by the changes in winter vertical mixing. When winter vertical mixing increased, water column-averaged PAR also decreased (Fig. 12a). On the other hand, when winter vertical mixing increased, so did the nutrient concentration (Fig. 12b). In the nitrogen dynamics, the effect of the changes in winter vertical mixing on ammonium concentration was weaker than that on nitrate concentration. The changes in the annual means of nitrate concentration showed a close relationship with MMLD (r2 = 0.80, p <0.001). With ammonium concentration, the relationship was also significant but weaker (r2 = 0.50, p<0.01).
The nitrate concentration was closely related to the nitrate flux by the entrainment process, which increased as the winter vertical mixing deepened (Fig. 12c). On the other hand, the ammonium regeneration is closely linked to biological processes. Among the biological processes, the zooplankton excretion accounted for 72.69% of total ammonium regeneration (Table 5) and was the main factor determining the interannual variability of ammonium regeneration (Fig. 12d). Banse (1995) suggested that zooplanktonmediated processes such as excretion and sloppy feeding make a large contribution to ammonium regeneration, so they play an important role in controlling the phytoplankton dynamics. This effect of zooplankton excretion explains why the winter vertical mixing has a weaker influence on ammonium concentration than on nitrate concentration.
Winter
Nitrate
-1-1-1—
Spring Summer Autumn
Annual
-1-1-1—
—i-1-1—
Ammonium
0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 12 3
L ♦ o- H I h I •
à <£> <£> ^ <$> ^ <§> * o
Chi-NPP Phyto-Diatoms-Flag-Pico Dino-Zoo Hetero Micro-Meso
* -ill A L
[1.. 11 o- T JL L.
Z I L IT l IT L
L. -01- T i «
J 1 L 111 1 i.
J . i 1 m fr
1 • 1 {h J i
1 §--f- t IU ir 1.
L „ _ IT Jl . I I.
1 n. If ! i r JU
11 I i 4 11
20 40 0 20 40 0 20 40 0 20 40 Percentage increase (%)
Fig. 9. Percentage increases of seasonal and annual means of nitrate concentration, ammonium concentration, chlorophyll a concentration, NPP, total phytoplankton biomass, PFT (Diatoms, Flag, Pico, Dino) biomass, total zooplankton biomass and ZFT (Hetero, Micro, Meso) biomass attributed to atmospheric deposition.
Table 5
Composition (%) of ammonium regeneration sources for the whole study period.
Zooplankton
Bacterial excretion
Remineralizaron
Mesozooplanton excretion
Microzooplanton excretion
Heterotrophic nanoflagellate excretion
The model shows that not only the bloom magnitudes but also the composition of PFTs may change in response to the changes in winter vertical mixing. The diatoms responded most positively to the changes in winter vertical mixing among the PFTs. On
M . _ - ----- - •
^ as „
1-1-i-1-r
0.1 0.5 1.0 1.5 2.0
The ratio of ammonium:nitrate
Fig. 10. Percentage increases of PFT biomass (red: diatoms, blue: flagellates, green: picophytoplankton, violet: dinoflagellates) according to the changes in the deposition ratio of ammonium to nitrate. The vertical dashed line denotes the base ratio of ammonium to nitrate. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
the other hand, the other PFTs' behaviour seems to depend on diatoms which pre-empt the early spring stage. When winter vertical mixing increased, NPP of flagellates and picophytoplankton increased, while their biomass decreased (Fig. 7). However, both NPP and biomass of dinoflagellates decreased. We now examine the cause of the different responses by groups. The model parameters regarding the resource utilization and competitive ability of PFTs were set in accordance with observed data from previous studies (e.g. Geider et al., 1997; Maclntyre et al., 2002; Wafar et al., 2004; Litchman et al., 2007; Litchman and Klausmeier, 2008). The parameters reflect the fact that the diatoms have high competitive ability in lower light conditions and a faster nitrate uptake rate (Table 3 and Fig. 13). The flagellates and picophytoplankton have a preference for ammonium over nitrate. They are also limited by a higher grazing rate, so their peak magnitudes are smaller and the fluctuation ranges are narrower than those of diatoms (Fig. 6). Although we used the specified values, the competitive ability and grazing rate of PFTs showed significant variance around the fitted curves through the complex effects of external environmental changes and the internal condition of organisms (Fig. 13).
As the winter vertical mixing has opposite effects on light and nutrients, the balance between these two effects will determine the competitive advantage of PFTs. With a shallower MLD, the light availability in winter improved (Fig. 12a) while the ambient nutrient level decreased (Fig. 12b). Although the light utilization of diatoms was much higher than other PFTs (Fig. 13), diatoms
Fig. 11. Percentage increases of N:P ratio (triangles) and N* (circles), chlorophyll a concentration, NPP, total phytoplankton biomass, PFT biomass, total Zooplankton and ZFT biomass in response to the changes in atmospheric deposition.
Fig. 12. (a) Relationship between Feb-Apr mean of PAR in the upper mixed layer and MMLD, (b) Relationship between annual mean of nitrogen concentration and MMLD (red: nitrate, blue: ammonium), (c) Relationship between annual mean of nitrate flux and MMLD, (d) Composition of ammonium regeneration (red: mesozooplankton excretion, blue: microzooplankton excretion, green: heterotrophic nanoflagellate excretion, violet: bacteria excretion, orange: remineralization). r2 with p-value is shown in the lower right corner of each figure (upper value for nitrate and lower value for ammonium, '": p <0.001, ": p <0.01, *: p <0.05). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).
Fig. 13. Monthly means of photosynthesis rate and annual means of nutrient uptake rates (upper panels, red: diatoms, blue: flagellates, green: picophytoplankton, violet: dinoflagellates) and monthly grazing rate of PFTs (lower panels). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version ofthis article.).
lost their competitive advantage as their nutrient uptake decreased when MMLD decreased (Fig. 14a). The nutrient limitation of diatoms started at a deeper MMLD (higher nitrate concentration) than that of other PFTs (Fig. 14b). The maximum NPP of diatoms occurred at a lower light level while that of other PFTs occurred at a much higher light level (Fig. 14c). Therefore, a deeper MMLD conferred diatoms advantages both in light and nutrients. A deeper MMLD would also enhance the annual NPP of flagellates and picophytoplankton because it would increase the total nitrogen flux in the upper layer (Fig. 12c). As a result, the relative contribution of the PFTs in the spring production would depend on MMLD. The annual mean NPP of diatoms, flagellates and pico-phytoplankton increased with increasing MMLD as more nutrients were supplied. However, the annual mean NPP of dinoflagellates decreased with increasing MMLD (Fig. 7) because the growing duration of dinoflagellates decreased (data not shown). The graZing rate on diatoms showed a smaller range with the peak at the medium level of NPP (Fig. 14d). This is because when their NPP became high, the graZers could not catch up with their growth. On the other hand, the grazing rate on flagellates and picophytoplankton continuously increased in a wider range (Fig. 14d). Because of this high grazing rate, their biomass decreased as MMLD increased (Fig. 7). Therefore, the biomass of flagellates and pico-phytoplankton decreased with increasing MMLD because of the increasing grazing rate. On the other hand, the biomass of dinoflag-ellates decreased with increasing MMLD because of decreased NPP.
4.2. Effects of atmospheric deposition
Next, we examine the effects of the increasing atmospheric deposition on PFTs. The model showed that the atmospheric deposition increased the nitrogen concentration in the upper layer. This effect was pronounced in summer when the nitrogen supplied by winter vertical mixing was depleted. The percentage change of ammonium concentration was much higher than that of nitrate concentration because of its relatively lower concentration in the mixed layer. In addition, as mentioned above, the changes in atmospheric deposition ratio of ammonium to nitrate altered the
composition of PFTs (Fig. 10). Throughout the range of ratios, the PFTs that have a high competitive affinity for ammonium benefitted more from ammonium supply by atmospheric deposition. Thus, the picophytoplankton responded most sensitively to the increased ammonium deposition.
We compared the percentage increases of NPP by atmospheric deposition with other studies from various regions (Table 6). The model result shows that the percentage increase of NPP was the highest in summer (23.48%) when the nitrogen entrained by vertical mixing was limited. The phytoplankton have a high carbon-to-nitrogen ratio in summer when the nitrogen is limited, but they are able to reduce the carbon-to-nitrogen ratio by getting the nitrogen supply from atmospheric deposition. This provided significant benefits to phytoplankton growth in summer.
The atmospheric deposition in the Ulleung Basin increased the annual mean of NPP by about 4.58%. The estimate from this study is higher than that of Duce et al. (2008), which suggested that the atmospheric deposition accounts for 3% of the new primary production in the global open ocean. However, the value is lower than the estimate from the southern East Sea (37°N, 135°E) by Onitsuka et al. (2009), despite the atmospheric deposition in this model run being much higher. They concluded that the atmospheric deposition potentially accounts for 8.77% of the annual mean of primary production. They used a simple NPZD model with only one nutrient component, nitrogen. Their calculation might have overestimated the effect of atmospheric deposition on primary production, because their model did not take account of the limitation of nutrients other than nitrogen.
In this experimental model run, we applied a single rate for atmospheric deposition each year disregarding the seasonal variations. However, the atmospheric deposition in the Ulleung Basin shows seasonal variations, which depend on weather conditions. In winter, the atmospheric deposition is low even though anthropogenic fine dusts and aerosols from the continent blow into the Korean Peninsula via the north-west monsoon. The atmospheric deposition increases slowly in spring, and becomes high in summer due to the increase in the wet deposition. Thus, if we consider the seasonal variability of atmospheric deposition, the effect of atmospheric deposition on primary production in summer will increase more.
Fig. 14. (a) Relationship between annual mean proportion of nitrogen uptake (PFT/total) and MMLD. Red circles represent nitrate, blue ammonium. r2 with p-value is shown in the lower right corner of each figure (upper value for nitrate and lower value for ammonium, ***: p <0.001, **: p <0.01, *: p <0.05), (b) Relationship between annual means of nutrient limitation (Eq. (8) of Blackford et al., 2004) and MMLD, (c) Relationship between Feb-Apr means of NPP and Feb-Apr means of PAR within the upper mixed layer, (d) Relationship between Feb-Apr means of grazing rate and Feb-Apr means of NPP.
4.3. Contribution of major sources of nitrogen
Previous studies have postulated that there could be various sources of nutrients in the Ulleung Basin such as vertical mixing (Onitsuka and Yanagi, 2005), TWC transport (Onitsuka et al., 2007), coastal upwelling (Yoo and Park, 2009), atmospheric desposition (Onitsuka et al., 2009), Changjiang diluted water (Isobe et al., 2002), Nakdong River discharge (Suh, 2006) and mesoscale eddies (Hong
et al., 2013). We conducted numerical simulations to see the relative contribution of the major sources of the nutrient supply on a regional scale.
First, the relative contribution from vertical mixing, atmospheric deposition and TWC transport in the Ulleung Basin was calculated. Kim et al. (2013) showed that the nitrogen flux by TWC transport in the East Sea was about 13.45 times greater than that by atmospheric deposition. For the study period, the nitrogen flux
Table 6
Estimates of primary production enhancement due to atmospheric deposition from various studies.
Study area Period Season Percentage increase (%) References
Southern North Sea 1999 Summer -5.5 De Leeuw et al. (2003)
Southern North Sea 2000 Summer 0.6-2.5 Spokes and Jickells (2005)
Global open ocean 2000 Annual -3 Duce et al. (2008)
North Atlantic Gyre 2002 Autumn 0.1-4.7 (dry) Baker et al. (2007)
0.6-2.9 (wet)
Southern North Sea 1996-2002 Annual 13.8-15 Troost et al. (2013)
Strait of Dover 2003 Spring -3.1 Boulart et al. (2006)
Southern East Sea (37° N, 135° E) 1996-2003 Winter 5.22 Onitsuka et al. (2009)a
Spring 5.58
Summer 11.7
Autumn 12.56
Annual 8.77
Mediterranean Sea 1999-2004 Annual -5 Lazzari et al. (2012)
Eastern Mediterranean Sea 2003- 2004 (including P) Winter 1-26 Christodoulaki et al. (2013)
Summer 33-34
South China Sea 2005- 2010 Annual 20 Kim et al. (2014)
Ulleung basin (37°N, 130.6°E) 2001-2012 Winter -0.08 This study
Spring 1.35
Summer 23.48
Autumn 5.23
Annual 4.58
a Reproduced seasonal value based on the monthly model output in the southern East Sea.
Table 7
Relative contribution (%) of the three sources of nitrogen flux. The nitrogen flux by vertical mixing consists of only nitrate and the nitrogen flux by atmospheric deposition and TWC transport includes both nitrate and ammonium.
Vertical mixing Atmospheric deposition TWC transport
59.10 2.83 38.07
Table 8
Changes in annual mean N:P ratio and annual mean N* by biochemical process in vertical mixing simulation for the whole study period.
Processes N:P ratio N*(mmolNm-3)
Physical process 13.00:1 0.00
Biochemical process 14.06 (13.42-15.87):1 0.03 (-0.01-0.14)
ratio of vertical mixing to atmospheric deposition from the model result was 20.88:1. Taking this into consideration, the relative contribution from the three sources is given in Table 7—59.1% was by vertical mixing, 2.8% by atmospheric deposition and 38.1% by TWC transport.
The TWC supplies heat, salt, fresh water and nutrients horizontally to the East Sea through the Korea Strait (Morimoto et al., 2009). Onitsuka et al. (2007) suggested that the horizontal nutrient flux through the western channel of the Korea Strait is the main source of nutrient supply and its contribution accounts for more than 60% of the annual primary production. While their model did not include atmospheric deposition added to TWC transport, our estimates reflect atmospheric deposition added to TWC following Kim et al. (2013). We have already described how the changes in vertical mixing and atmospheric deposition can affect the bloom magnitudes as well as the composition of PFTs via changing the balance in nutrient supply. From this result, we suggest that the TWC transport can affect the peak timing/duration in phytoplankton blooms and the composition of PFTs.
The changes in the balance of nutrient supply from outside have impacts on the N:P ratio and N*. The model result shows that the N:P ratio increased by 4.57% and the N* increased by 5.43% with 2.15 mg N m-2 d-1 of the nitrogen flux by atmospheric deposition. They increased steadily with the increase in atmospheric deposition (Fig. 11). Kim et al. (2013) argued that the nitrogen flux by TWC transport was an order of magnitude greater than that by atmospheric deposition and the increase in N* seems to be associated with TWC transport.
Furthermore, the model result suggests that the N:P ratio or N* in surface waters would be affected significantly by biochemical processes such as the nutrient uptake of dominant phytoplankton type or the excretion of zooplankton and bacteria (Table 8). Previous studies have shown that the internal N:P stoichiometry of zooplankton and bacteria is fairly constant (Sterner and Elser, 2002)
and it is easy to predict the changes in N* caused by the changes in their biomass. But it is difficult to predict the changes in N* by phytoplankton uptake as a function of biomass, because the internal N:P stoichiometry of phytoplankton differs between PFTs (Quigg et al., 2003; Finkel et al., 2009) and it is more flexible (Kooijman et al., 2004).
In this study, we addressed how the changes in the balance of nutrient supply from outside alter the PFTs and N* in the upper mixed layer. In the future, more studies are needed to investigate the effect of the changes in balance of nutrient supply and the effect of the internal N:P stoichiometry of phytoplankton comprehensively to understand the cause of the increase in N* in the Ulleung Basin.
5. Conclusions
We investigated the impacts of varying winter vertical mixing and atmospheric deposition on the planktonic community using a zero-dimensional ERSEM. The model results show that the changes in winter vertical mixing had varying effects on the competitive advantages of PFTs and thereby altered the succession process of PFTs in the upper mixed layer. With shallow winter vertical mixing, the increased PAR and the decreased nutrient in the upper mixed layer undermined the competitive advantage of diatoms to the advantage of other PFTs. With the increase in atmospheric deposition, the flagellates and picophytoplankton increased whereas the diatoms and dinoflagellates decreased. For the study period 2001-2012, the atmospheric deposition in the Ulleung Basin increased the annual mean of NPP by about 4.58%. The relative contribution of vertical mixing, atmospheric deposition and TWC transport was estimated at 59.10:2.83:38.07% for the same period. Taken together, the model suggests that the shallower MLD in the Ulleung Basin may result in a shift toward a dominance of small phytoplankton such as flagellates and pico-phytoplankton and the increasing atmospheric deposition would
amplify such changes. Certainly there are many shortcomings in our model set-up. These include the limitation of zero-dimensional models in general, parametrization of MLD and uncertainties in bio-geochemical parameters that are difficult to measure in the ocean. Specifically our model cannot address the horizontal processes or the event-scale turbulences that might be important in the winterspring period in the region. Notwithstanding these shortcomings, our study shows how the two important environmental changes in the region, warming and atmospheric deposition, can influence the planktonic community.
Acknowledgements
We are greatly indebted to ERSEM Development Team whose efforts made our research possible. We thank Drs. ChanJoo Jang and M. Butenschon for their comments on earlier results of the study. We thank NASAfor SeaWiFS and MODIS-Aqua satellite data, KODC for temperature and salinity data, KMAfor cloud cover data and Plymouth Marine Laboratory (PML) for ERSEM. We also thank the members of the Marine Ecosystem Dynamic Laboratory (MEDL), KIOST, who were responsible for the nutrient and HPLC pigment observation data. This research was supported by a grant from the 2015 KIOST projects, #PN66150, #PE99305 and #PE9935A. This paper is a contribution to SCOR WG137 "Patterns of Phytoplankton Dynamics in Coastal Ecosystems".
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ecolmodel.2015. 11.012.
References
Auger, P.A., Ulses, C., Estournel, C., Stemmann, L., Somot, S., Diaz, F., 2014. Interannual control of plankton communities by deep winter mixing and prey/predator interactions in the NW Mediterranean: results from a 30-year 3D modeling study. Prog. Oceanogr. 124,12-27. Baker, A.R., Weston, K., Kelly, S.D., Voss, M., Streu, P., Cape, J.N., 2007. Dry and wet deposition of nutrients from the tropical Atlantic atmosphere: links to primary productivity and nitrogen fixation. Deep Sea Res. Part I Oceanogr. Res. Pap. 54, 1704-1720.
Banse, K., 1995. Zooplankton—Pivotal role in the control of ocean production. ICES
J. Mar. Sci. 52, 265-277. Baretta-Bekker, J.G., Baretta, J.W., Ebenhoh, W., 1997. Microbial dynamics in the marine ecosystem model ERSEM II with decoupled carbon assimilation and nutrient uptake. J. Sea Res. 38,195-212. Baretta-Bekker, J.G., Baretta, J.W., Rasmussen, E., 1995. The microbial food web in
the European Regional Seas Ecosystem Model. J. Sea Res. 33,363-379. Blackford, J.C., Allen, J.I., Gilbert, F.J., 2004. Ecosystem dynamics at six contrasting
sites: a generic model study. J. Mar. Syst. 52,191-215. Boulart, C., Flament, P., Gentilhomme, V., Deboudt, K., Migon, C., Lizon, F., Schapira, M., Lefebvre, A., 2006. Atmospherically promoted photosynthetic activity in a well-mixed ecosystem: significance of wet deposition events of nitrogen compounds. Estuar. Coast. Shelf. Sci. 69,449-458. Broekhuizen, N., Heath, M.R., Hay, S.J., Gurney, W.S.C., 1995. Modellingthe dynamics
of the North Sea's mesozooplankton. J. Sea Res. 33,381-406. Chiba, S., Aita, M.N., Tadokoro, K., Saino, T., Sugisaki, H., Nakata, K., 2008. From climate regime shifts to lower-trophic level phenology: synthesis of recent progress in retrospective studies of the western North Pacific. Prog. Oceanogr. 77,112-126.
Christodoulaki, S., Petihakis, G., Kanakidou, M., Mihalopoulos, N., Tsiaras, K., Tri-antafyllou, G., 2013. Atmospheric deposition in the Eastern Mediterranean. A driving force for ecosystem dynamics. J. Mar. Syst. 109-110, 78-93. De Leeuw, G., Spokes, L., Jickells, T., Skjoth, C.A., Hertel, O., Vignatic, E., Tamme, S., Schulze, M., Sorensen, L.L., Pedersen, B., Klein, L., Schlunzen, K.H., 2003. Atmospheric nitrogen inputs into the North Sea: effect on productivity. Cont. Shelf. Res. 23,1743-1755. Doney, S.C., 2006. Plankton in a warmer world. Nature 444, 695-696. Doney, S.C., Kleypas, J.A., Sarmiento, J.L., Falkowski, P.G., 2002. The US JGOFS Synthesis and Modeling Project—an introduction. Deep Sea Res. Part II Top. Stud. Oceanogr. 49,1-20.
Duce, R.A., LaRoche, J., Altieri, K., Arrigo, K.R., Baker, A.R., Capone, D.G., Cornell, S., Dentener, F., Galloway, J., Ganeshram, R.S., Geider, R.J., Jickells, T., Kuypers, M.M., Langlois, R., Liss, P.S., Liu, S.M., Middelburg,J.J., Moore, C.M., Nickovic, S., Oschlies,
A., Pedersen, T., Prospero, J., Schlitzer, R., Seitzinger, S., Sorensen, L.L., Uematsu, M., Ulloa, O., Voss, M., Ward, B., Zamora, L., 2008. Impacts of atmospheric anthropogenic nitrogen on the open Ocean. Science 320, 893-897.
Dugdale, R.C., Wilkerson, F.P., Barber, R.T., Chavez, F.P., 1992. Estimating new production in the equatorial Pacific Ocean at 150 W. J. Geophys. Res. 97, 681-686.
Evans, G.T., Parslow, J.S., 1985. A model of annual plankton cycles. Biol. Oceanogr. 3, 327-347.
Falkowski, P.G., 1980. Light-shade adaptation in marine phytoplankton. In: Falkowski, P.G. (Ed.), Primary Productivity in the Sea. Plenum Press, New York, pp. 99-119.
Falkowski, P.G., Katz, M.E., Knoll, A.H., Quigg, A., Raven, J.A., Schofield, O., Taylor, F.J.R., 2004. The evolution of modern eukaryotic phytoplankton. Science 305, 354-360.
Falkowski, P.G., Laws, E.A., Barber, R.T., Murray, J.W., 2003. Phytoplankton and their role in primary, new, and export production. In: Fasham, M.J.R. (Ed.), Ocean Biogeochemistry: the Role ofthe Ocean Carbon Cycle in Global Change. SpringerVerlag, Berlin Heidelberg, pp. 99-121.
Fasham, M.J.R., 1993. Modelling marine biota. In: Heimann, M. (Ed.), The Global Carbon Cycle. Springer-Verlag, Berlin Heidelberg, pp. 457-504.
Finkel, Z.V., Beardall, J., Flynn, K.J., Quigg, A., Rees, T.A.V., Raven, J.A., 2009. Phytoplankton in a changing world: cell size and elemental stoichiometry.J. Plankton Res. 32,119-137.
Geider, R.J., MacIntyre, H.L., Kana, T.M., 1997. Dynamic model of phytoplankton growth and acclimation: responses ofthe balanced growth rate and the chlorophyll a: carbon ratio to light, nutrient-limitation and temperature. Mar. Ecol. Prog. Ser. 148, 187-200.
Hong, G., Lee, D., Yang, D., Kim, Y., Park, J., Park, C., 2013. Eddy- and wind-sustained moderate primary productivity in the temperate East Sea (Sea of Japan). Biogeo-sciences 10,10429-10458.
Ichiye, T., 1984. Some problems of circulation and hydrography of the Japan Sea and the Tsushima current. In: Ichiye, T. (Ed.), Ocean Hydrodynamics of the Japan and East China Seas. Elsevier Science Publishers, Amsterdam, pp. 15-54.
Isobe, A., Ando, M., Watanabe, T., Senjyu, M., Sugihara, S., Manda, A., 2002. Fresh water and temperature transports through the Tsushima-Korea Straits. J. Geo-phys. Res. 107 (c7), 3065-3079.
Jennings, S., Melin, F., Blanchard, J.L., Forster, R.M., Dulvy, N.K., Wilson, R.W., 2008. Global-scale predictions of community and ecosystem properties from simple ecological theory. Proc. Biol. Sci. 275, 1375-1383.
Kameda, T., Ishizaka, J., 2005. Size-fractionated primary production estimated by a two-phytoplankton community model applicable to oceancolor remote sensing. J. Oceanogr. 61,663-672.
Kim, D., Yang, E.J., Kim, K.H., Shin, C.-W., Park, J., Yoo, S., Hyun, J.-H., 2011a. Impact of an anticyclonic eddy on the summer nutrient and chlorophyll a distributions in the Ulleung Basin, East Sea (Japan Sea). ICES J. Mar. Sci., http://dx.doi.org/10. 1093/icesjms/fsr178.
Kim,T.W., Lee, K., Najjar, R.G.,Jeong, H.D.,Jeong, H.J., 2011b. Increasing N abundance in the Northwestern Pacific Ocean due to atmospheric nitrogen deposition. Science 334, 505-509.
Kim, H., Yoo, S., Oh, I.S., 2007. Relationship between phytoplankton bloom and wind stress in the sub-polar frontal area of the Japan/East Sea. J. Mar. Syst. 16, 205-216.
Kim, K, Kim, KR., Min, D.H., Volkov, Y., Yoon, J.H., Takematsu, M., 2001. Warming and structural changes in the East (Japan) Sea: a clue to future changes in global oceans? Geophys. Res. Lett. 28,3293-3296.
Kim, S.K., Chang, K.I., Kim, B., Cho, Y.K., 2013. Contribution of ocean current to the increase in N abundance in the Northwestern Pacific marginal seas. Geophys. Res. Lett. 40, 143-148.
Kim, S.W., Saitoh, S., Ishizaka, J., Isoda, Y., Kishino, M., 2000. Temporal and spatial variability of phytoplankton pigment concentration in the Japan Sea derived from CZCS images. J. Oceanogr. 56, 527-538.
Kim, T.H., Kim, G., 2013. Factors controlling the C:N:P stoichiometry of dissolved organic matter in the N-limited,cyanobacteria-dominated East/Japan Sea. J. Mar. Syst. 115-116,1-9.
Kim, T.W., Lee, K., Duce, R., Liss, P., 2014. Impact of atmospheric nitrogen deposition on phytoplankton productivity in the South China Sea. Geophys. Res. Lett. 41, 3156-3162.
KMA (Korea Meteorological Administration), 2013. Report of global atmosphere watch. KMA, Korea. (https://www.climate.go.kr:8005/index.html).
Kooijman, S.A.L.M., Andersen, T., Kooi, B.W., 2004. Dynamic energy budget representations of stoichiometric constraints on population dynamics. Ecology. 85, 1230-1243.
Lazzari, P., Solidoro, C., Ibello, V., Salon, S., Teruzzi, A., Beranger, K., Colella, S., Crise, A., 2012. Seasonal and interannual variability of plankton chlorophyll and primary production in the Mediterranean Sea: a modelling approach. Biogeosciences 9, 217-233.
Lee, Y.W., Park, M.O., Kim, Y.S., Kim, S.S., Kang, C.K., 2011. Application of photo-synthetic pigment analysis using a HPLC and CHEMTAX program to studies of phytoplankton community composition. J. Korean. Soc. Oceanogr. 16,117-124 (in Korean).
Lewandowska, A.M., Boyce, D.G., Hofmann, M., Matthiessen, B., Sommer, U., Worm,
B., 2014. Effects of sea surface warming on marine plankton. Ecol. Lett. 17, 614-623.
Lim, S., Jang, C.J., Oh, I.S., Park, J., 2012. Climatology ofthe mixed layer depth in the East/Japan Sea. J. Mar. Syst. 96-97,1-14.
Litchman, E., Klausmeier, C.A., 2008. Trait-based community ecology of phytoplankton. Annu. Rev. Ecol. Evol. Syst. 39, 615-639.
Litchman, E., Klausmeier, C.A., Schofield, O.M., Falkowski, P.G., 2007. The role of functional traits and trade-offs in structuring phytoplankton communities: scaling from cellular to ecosystem level. Ecol. Lett. 10,1170-1181.
Maclntyre, H.L., Kana, T.M., Anning, T., Geider, R.J., 2002. Photoacclimation of photosynthesi irradiance response curves and photosynthetic pigments in microalgae and cyanobacteria. J. Phycol. 38,17-38.
Morimoto, A., Takikawa, T., Onitsuka, G., Watanabe, A., Moku, M., Yanagi, T., 2009. Seasonal variation of horizontal material transport through the eastern channel of the Tsushima Straits. J. Oceanogr. 65, 61-71.
Onitsuka, G., Uno, I., Yanagi, T., Yoon, J.H., 2009. Modeling the effects of atmospheric nitrogen input on biological production in the Japan Sea. J. Oceanogr. 65, 433-438.
Onitsuka, G., Yanagi, T., 2005. Differences in ecosystem dynamics between the northern and southern parts of the Japan Sea: analyses with two ecosystem models. J. Oceanogr. 61, 415-433.
Onitsuka, G., Yanagi, T., Yoon, J.H., 2007. A numerical study on nutrient sources in the surface layer of the Japan Sea using a coupled physical-ecosystem model. J. Geophys. Res. 112, C05042, http://dx.doi.org/10.1029/2006JC003981.
Quigg, A., Finkel, Z.V., Irwin, A.J., Rosenthal, Y., Ho, T.Y., Reinfelder, J.R., Schofield, O., Morel, F.M.M., Falkowski, P.G., 2003. The evolutionary inheritance of elemental stoichiometry in marine phytoplankton. Nature. 425, 291-294.
Rosati, A., Miyakoda, K., 1988. A general circulation model for upper ocean simulation. J. Phys. Oceanogr. 18,1601-1626.
Smetacek, V., 1999. Diatoms and the ocean carbon cycle. Protist. 150, 25-32.
Spokes, L.J., Jickells, T.D., 2005. Is the atmosphere really an important source of reactive nitrogen to coastal waters? Cont. Shelf Res. 25, 2022-2035.
Sprintall, J., Tomczak, M., 1992. Evidence of the barrier layer in the surface layer of the tropics. J. Geophys. Res. 97, 7305-7316.
Sterner, R.W., Elser, J.J., 2002. Ecological stoichiometry: the biology of elements from molecules to the biosphere. Princeton, NJ, Princeton University Press, New Jersey.
Suh, Y.S., 2006. Study on abnormal distribution of high chlorophyll a in the east sea of Korea in spring season using ocean color satellite remote sensing. J. Environ. Sci. 15, 59-66 (in Korean).
Talley, L.D., Tishchenko, P., Luchin, V., Nedashkovskiy, A., Sagalaev, S., Kang, D.J., Warner, M., Min, D.H., 2004. Atlas of Japan (East) Sea hydrographic properties in summer, 1999. Prog. Oceanogr. 61, 277-348.
Troost, T.A., Blaas, M., Los, F.J., 2013. The role of atmospheric deposition in the eutrophication of the North Sea: a model analysis. J. Mar. Syst. 125, 101-112.
Uno, I., Uematsu, M., Hara, Y., He, Y.J., Ohara, T., Mori, A., Kamaya, T., Murano, K., Sadanaga, Y., Bandow, H., 2007. Numerical study of the atmospheric input of anthropogenic total nitrate to the marginal seas in the western North Pacific region. Geophys. Res. Letts. 34, L17817.
Vichi, M., Masina, S., 2009. Skill assessment of the PELAGOS global ocean biogeochemistry model over the period 1980-2000. Biogeosciences 6, 3511-3562.
Wafar, M., L'hekguen, S., Raikar, V., Maguer, J., 2004. Nitrogen uptake by size-fractionated plankton in permanently well-mixed temperate coastal waters. J. Plankton Res. 26,1207-1218.
Yamada, K., lshizaka, J., Nagata, H., 2005. Spatial and temporal variability of satellite primary production in the Japan Sea from 1998 to 2002. J. Oceanogr. 61, 857-869.
Yamada, K., lshizaka, J., Yoo, S., Kim, H., Chiba, S., 2004. Seasonal and interannual variability of sea surface chlorophyll a concentration in the Japan/East Sea (JES). Prog. Oceanogr. 61, 193-211.
Yentsch, C.S., 1990. Estimates of'new production' in the Mid-North Atlantic. J. Plankton Res. 12,717-734.
Yoo, S., Park, J., 2009. Why is the southwest the most productive region of the East Sea/Sea of Japan? J. Mar. Syst. 78,301-315.
Zhang, J., Zhang, G.S., Bi, Y.F., Liu, S.M., 2011. Nitrogen species in rainwater and aerosols of the Yellow and East China Sea: effects of the East Asian monsoon and anthropogenic emissions and relevance for the NW Pacific Ocean. Global Biogeochem. Cycles. 25, GB3020, http://dx.doi.org/10.1029/2010GB003896.
