Influences of Physical Processes and Anthropogenic Influx on Biogeochemical Cycle in the Java Sea: Numerical Model Experiment

Available online at www.sciencedirect.com

ScienceDirect Procedia Environmental Sciences 33 (2016) 532 – 552

The 2nd International Symposium on LAPAN-IPB Satellite for Food Security and Environmental Monitoring 2015, LISAT-FSEM 2015

Influences of physical processes and anthropogenic influx on biogeochemical cycle in the Java Sea: numerical model experiment Alan F. Koropitana,*, Motoyoshi Ikedab a

Department of Marine Science and Technology, Faculty of Fisheries and Marine Sciences, Bogor Agricultural University, Dramaga, Bogor 16680, Indonesia b Graduate School of Environmental Sciences, Hokkaido University, Kita 10 Nishi 5, Kita-ku, Sapporo 060-0810, Japan

Abstract Three-dimensional coupled physical-biogeochemical model was utilized in order to investigate the influence of physical processes to biogeochemical cycle in the Java Sea. The biogeochemical model was consisted of nitrate, ammonium, phosphate, phytoplankton, zooplankton, pelagic detritus and benthic detritus. The coupled model could reproduce the basic condition of seasonal variability of surface Chl a distribution consistently with satellite data. Model results and satellite data clearly showed seasonal variability of Chl a distributions, influenced by monsoon, through water exchange with adjacent seas and nutrient supply from rivers discharges. Phytoplankton blooming during southeast monsoon is higher in general than northwest monsoon, due to upwelling event in the eastern Java Sea. On the other side, the role of nutrient riverine input during northwest monsoon (rainy season) is only limited in the region near river mouths or coastal regions. The calculated annual new productions (Rnew) suggested that the regenerated production is predominant in the Java Sea, except for some regions (e.g: Jakarta Bay and southcoast of Kalimantan) that is influenced by human activities in the land (anthropogenic perturbation). The anthropogenic impact through riverine input triggers high primary production in the regions, while it subsequently uptakes atmospheric CO2 in particular monsoon season. However, as a consequence of high sea surface temperature in the tropical region, annually the entire Java Sea acts as a source for CO2 even though the Java Sea is a net autotrophic. © 2016 2016The TheAuthors. Authors. Published by Elsevier B.V.is an open access article under the CC BY-NC-ND license Published by Elsevier B.V. This Peer-review under responsibility of the organizing committee of LISAT-FSEM2015. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of LISAT-FSEM2015 Keywords: coupled physical-biogeochemical model; Java Sea; riverine input; anthropogenic impact; primary production; CO2.

* Corresponding author. Tel.: +62-251-8623-644; fax: +62-251-8623-644. E-mail address: [email protected].

1878-0296 © 2016 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/). Peer-review under responsibility of the organizing committee of LISAT-FSEM2015 doi:10.1016/j.proenv.2016.03.106

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

533

1. Introduction The Java Sea has a unique role in human society because of its location in the middle of three main islands in the Indonesian archipelago: Kalimantan, Java and Sumatra (Fig. 1). Koropitan and Ikeda [1] have summarized that the dimensions of the Java Sea are roughly rectangular with the mean depth of 50 m. The water depths increase from 20 m off the coast of southeast Sumatra to more than 60 m in its eastern part. Through the northern open boundary, the Java Sea is linked with the three straits: Karimata, Gaspar and Bangka. The eastern and western open boundaries are connected with the Flores Sea and Sunda Strait, respectively. The roles of physical processes and anthropogenic impact on the biogeochemical cycle of the Java Sea have not been well understood. What role do the monsoonal currents and tidal mixing play in the material transport? How significant are nutrient fluxes from river discharges and adjacent seas in the effects on the marine ecosystem? In addition to the carbon fluxes, it is not clear yet whether the Java Sea is source or sink for the atmospheric CO2, while the atmospheric CO2 is rising during the last 50 years to 400 ppmV (at Mauna Loa Station). Related to the environmental problems in the Java Sea, on the other hand, the Java Sea is influenced by human activities on land. Mainly, the human populations are concentrated in Java Island with 57% of the total Indonesian population: ~206 millions in year of 2000. Talaue-McManus [2] reported that majority of the rivers in Java Island have influenced the Java Sea with moderate to heavy pollution. She concluded that land-based sources play a major role in coastal pollution, including the Java Sea where the wastes flow mostly from domestic, agricultural, and industrial sources. In addition, World Bank [3] reported that the pollution through river discharge of Java Island is divided into 44% of industry and 56% of human settlement. Little information is available about the human impact and role of rivers from Kalimantan and Sumatra Islands to the Java Sea. However, Susanto et al. [4] recently reported that the satellite Chl a images derived from SeaWiFS showed high concentrations along the south-coast of Kalimantan and southeast-coast of Sumatra, as well as the north-coast of Java during the whole years. Therefore, there is an indication of eutrophication among the coastal regions in the Java Sea. Especially in Jakarta Bay, Damar [5] pointed out the occurrence of hyper-eutrophic condition caused by high riverine input of nutrients. In order to investigate the interaction between physical processes and biogeochemical cycle in the Java Sea, here we perform a three-dimensional coupled hydrodynamic-biogeochemical model as an approach of a comprehensive study. In addition, the present research aims to resolve whether the Java Sea is a source or sink of atmospheric CO2. The recent paper by Cai and Dai [6] posed a question in the global estimation results suggesting the entire coastal seas or continental shelf as a sink. Cai and Dai [6] pointed out that the North Sea study by Thomas et al. [7] was not appropriate in extrapolating to the global scale. Furthermore, they proposed that some coastal regions such as U.S. South Atlantic Bight and northern South China Sea act as an annual CO2 source to the atmosphere. They also suggested that the tropical shelves are most likely sources of CO2 to the atmosphere due to the high annual surface temperature (as pointed out also by Feely et al. [8]), the lack of a strong spring bloom, inputs from marshes and mangroves, and/or reef formation. In this paper, we begin Section 2 with an introduction of the hydrodynamic model and discuss the physical phenomena as found in the previous study. The explanation of biogeochemical development is also summarized in this section. Section 3 examines the seasonal variability of observed riverine nutrient input in the Java Sea. In addition, comparisons of seasonal Chl a patterns in surface waters of the Java Sea were made between model result and satellite SeaWiFS images. Then annual budget of ecosystem compartments are discussed to analyze the role of adjacent seas and riverine nutrient input related to the biogeochemical processes and anthropogenic impact. In this section, the seasonal variability of primary production and CO2 flux are also examined in some specific regions within the Java Sea. Furthermore, sensitivity studies for phytoplankton biomass and CO2 flux are also made to analyze the biogeochemical response. A summary and discussion are presented in Section 4.

534

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

Fig. 1. Model region and rivers system surrounding the Java Sea.

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

535

2. Coupled Hydrodynamic-Biogeochemical Model 2.1. Hydrodynamic phenomena and model It has been well documented that the Indonesian seas are mainly under monsoon regime with two distinct monsoons [9]: northwest monsoon – NWM (December – February) and southeast monsoon – SEM (June – August). The other periods (months) are transition months between both distinct periods. However, as pointed out by Susanto et al. [4], recent observational data showed a change in the monsoon transition periods in which their durations are shorter for around one month than the previous knowledge. The transition month of NWM to SEM is April, while that of SEM to NWM is October. Therefore, the flow pattern of the Java Sea is more influenced by the monsoonal winds [9, 10]. The hydrodynamic model is the Princeton Ocean Model (POM), a primitive equation model with Boussinesq and hydrostatic approximations [11]. The specific information about the hydrodynamic model and its implementation and validation in the Java Sea has been summarized by Koropitan and Ikeda [1]. The model domain covers from 2q42’ to 8q14’ S and from 105q42’ to 114q42’ E. The model uses a Cartesian coordinate system in the horizontal (2’ by 2’ grid size) and a sigma coordinate system in the vertical. Along the vertical coordinate, the domain is divided into 21 unequal sigma levels (V = 0 at the sea surface and -1 at the sea bottom), with higher resolution near the surface and bottom layers. Therefore, the model has a total of 270 x 166 x 21 grid points. The tidal forcing of the model is imposed by the output from the global tidal model ORI.96 with assimilated TOPEX/Poseidon altimeter data [12]. The vertical eddy viscosity is computed using Mellor and Yamada’s method [13] level-2.5 turbulent closure scheme. Note that the present simulation uses the shear-dependent Smagorinsky formulation [14] for horizontal eddy viscosity with a coefficient of 3. Following the idea from Griffies and Hallberg [15], the case of the Java Sea indicates that tidal residual flow is comparable with monsoon-driven circulation for horizontal advection, while the coefficients of Smagorinsky [14, 16] were originally developed for the atmospheric model. Then, Griffies and Hallberg [16] suggested that the coefficient might vary from 2.2 to 4 for oceanic problems. In order to compare with satellite images, we have tried to tune the model by using coefficients of 2 and 3. In addition, the previous model in the Java Sea [1] reported important phenomena in that region, regarding physical processes. They found that the K1 tide clearly shows the lowest mode resonance in the Java Sea with intensification around the nodal point in the central region. The K1 tide produces a major component of tidal energy, which flows westward and dissipates through the node region near the Karimata Strait resulting in high tidal mixing. The enhanced tidal mixing in the central region (around nodal point) of the Java Sea is expected to play a crucial role in vertical exchange of nutrients and control of biological productivity. Combination of wind effects and tidal mixing seems to be important for the material transport in the Java Sea. The monsoonal flow pattern affects the horizontal transport, but tidal mixing plays a key role in vertical exchange. These environmental characters are concerned directly with the physical processes in the Java Sea. 2.2. Biogeochemical model The biogeochemical model is consisted of 8 compartments (Fig. 2), which are nitrate (NO3), ammonium (NH4), phosphate (PO4), phytoplankton (F), zooplankton (Z), pelagic detritus (PD), benthic detritus (BD) and dissolved inorganic carbon (or TCO2). The model was developed by combining the 5-compartment ecosystem model (NO3, NH4, F, Z and PD) from Wroblewski [17] and the 4-compartment ecosystem model (dissolved inorganic phosphate, F, Z and PD) which was originally developed by Denman and Peña [18] but modified for the nutrient type from N to P. We used sediment cohesive process for calculating BD, as demonstrated by Koropitan et al. [19] in Jakarta Bay. Moreover, we adopted the carbonate system model from OCMIP (Ocean Carbon-cycle Model Inter-comparison [40]) to calculate TCO2. The biogeochemical model is mainly calculated in N and uses the Redfield ratio for converting to P and C (C:N:P = 106:16:1), while the ratio of C/Chl a is set to be 50 based on its range of 27 – 67 [20]. A detailed description about this model is given in Appendix A. Biogeochemical parameters used in the present study are listed in Table 1. In general, the parameters are similar with Newberger et al. [21]. However, some specific parameters, such as half-saturation for phytoplankton uptake of

536

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

nutrients (NO3, NH4 and PO4) and optimal light intensity for phytoplankton growth were adopted from Damar [5], who used water samples from Jakarta Bay for a laboratory experimental study. The other potential source for nutrient cycling is from benthic processes. By using a model, Wei et al. [22] concluded that the benthic remineralization rate in the Bohai Sea is two times higher than that in the pelagic environment. Taking account of the strong tidal mixing in the Java Sea, the nutrient budget may be influenced by the benthic processes. Therefore we used the cohesive sediment model for the benthic processes where their parameter values are similar with Ningsih et al. [10] who applied them to the Java Sea. However, we omitted the effect of wind-wave on the bottom shear stress in calculating the erosion and deposition fluxes. Atmosphere

Air-Sea Flux (FCO ) 2

Red-field ratio PO

NH

4

TCO

4

2

Oxidation NO

Nutrient-Uptake (A F) 1

3

CO -Uptake 2

Min(NO +NH ,PO ) 3

Water column

F Respiration (A F)

4

4

Phytoplankton (F)

7

Grazing {(1-J)A Z} Z Excretion/ mortality (A Z)

3

4

Z Zooplankton

F Mortality (A F) 2

Egestion (JA Z) 3

Pelagic Decom-position (A D) 5

PD Pel. Detritus

Erosion (E) or Sedimentation (S) Bottom

Bottom Decom-position (A D) 6

BD Bot. Detritus

Fig. 2. A schematic diagram of the biogeochemical model which represents C (dashed line), N (full line) and P (dash-dotted line) cycles. The main calculation is based on N unit, while the Redfield ratio (C:N:P=106:16:1) is applied in converting material in the dotted box.

The coupled hydrodynamic-biogeochemical model was forced by the dominant constituent of K1 tide to consider tidal mixing as suggested by Koropitan and Ikeda [1], monthly climatological mean wind [23] and river discharges. In particular for the biogeochemical compartments, the model was forced by monthly riverine input of nutrients and solar radiation [24]. We used river discharge data from 21 rivers, which empty to the Java Sea from Sumatra, Java

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

537

and Kalimantan Islands (Fig. 1). The monthly river discharges were obtained by averaging the available daily data that vary from 1 to 10 years, except for Sebangau River in Kalimantan which has limited data. Sebangau River has smaller discharge compared to the whole data that we have, so that the data may not significantly affect the nutrient distribution. Furthermore, the river nutrient inputs were obtained by multiplying the river discharges and concentrations of nutrients in the rivers. However, the concentrations of nutrients are only available for several days of incidental observation in 1 year. Hence the monthly concentrations were interpolated based on availability of data. The detailed information of river discharges and concentrations of nutrients is shown in Table 2. Note that the hydrodynamic model has still homogeneous density, and that is the same as the tidal model from Koropitan and Ikeda [1]. Table 1. Parameters of the biogeochemical model. Parameter

Description

Value

RN2

NH4 oxidation rate

0.04 d-1

Vm

phytoplankton maximum uptake rate

1.5 d-1

Iopt

optimal light intensity for phytoplankton growth

143.9 watt/m2

D1

light dissipation coefficient of sea water

0.04 (mmol N/m3)-1 m-1

D2

phytoplankton self-shading coefficient

0.046 m-1

KN1

half-saturation for phytoplankton uptake NO3

2.5 mmol N/m3

KN2

half-saturation for phytoplankton uptake NH4

3.47 mmol N/m3

KN3

half-saturation for phytoplankton uptake PO4

0.47 mmol P/m3

<

NH4 inhabitation parameter

1.46 (mmol N/m3)-1

RPD

decomposition rate for pelagic detritus

0.5 d-1

RBD

remineralization rate for benthic detritus

0.2 d-1

mZ

phytoplankton specific mortality rate

0.1 d-1

mP

zooplankton specific excretion/mortality rate

0.145 d-1

Rm

zooplankton maximum grazing rate

0.52 d-1

/

Ivlev constant

0.5 (mmol N/m3)-1

J

fraction of zooplankton grazing egested

0.3

Wd

sinking rate for detritus

4 m/d

Wp

sinking rate for phytoplankton

0 m/d

rNP

ratio of N/P

16/1

rNC

Ratio of N/C

16/106

The initial concentrations of NO3 and PO4 were set to be homogeneously 0.3 and 0.2, respectively. These data were obtained from minimum values of annual means of NO3 and PO4 [25]. F initial condition was set based on surface Chl a from the European Service for Ocean Colour data set (case-1 water). We implemented annual mean from monthly data of the data set for period of 1997-2009 and 0.25° by 0.25° space resolutions. For the vertical layer of F initial condition, the annual mean data set was then extrapolated vertically using algorithm of Asanuma et al. [26]. In particular of TCO2, we adopted the annual gridded data of the global model results from Goyet et al. [27]. However, the initial concentrations of NH4, Z, and PD were assumed to be 0.1 mmol/m3 homogeneously, while the initial concentration of BD was assumed to be 15 mmol/m2 for the entire surface bottom layer. Related to the carbonate system model, there are several fixed parameters included in the calculation, as follows: annual silicate in the Java Sea with range of 9-11 mmol/m3 [25], total alkalinity of 2200 Pmol/kg from GLODAP data set (minimum value near the Indonesian seas region), sea surface temperature of 28.3°C, sea surface salinity of 34.6 psu, atmospheric mole fraction CO2 of 385 Patm, atmospheric pressure of 1 atm and pH range of 6.90 – 8.72. We do

538

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

not consider the silicate cycle in the model, since its value is higher in general for the Java Sea case, and hence, is insignificant as a limiting factor for phytoplankton growth. Table 2a. River discharges data. Station

River Name

Period of daily data*

1

Tulang Bawang

1992 and 1996

2

Seputih

1992

3

Sekampung

1992

4

Cibanten

1994-1996

5

Ciujung

1992-1995

6

Cisadane

1992-1997, 1999-2000

7

Angke [5]

December 2000 – November 2001

8

Priok [5]

December 2000 – November 2001

9

Marunda [5]

December 2000 – November 2001

10

Citarum

1992-2001

11

Garang

1992 -1996, 1999, 2000

12

Brantas

1992-1996, 2000

Barito

Assumed similar with 14

14

Kapuas

1993-1996, 2000

15

Kahayan

1993-1996, 1999

16

Sebangau [31]

Several days during 2002 - 2003

17

Katingan

Assumed similar with 15

18

Mentaya

Assumed similar with the low monthly river discharge of 15

19

Seruyan

Assumed similar with the low monthly river discharge of 15

20

Kumai

Assumed similar with the low monthly river discharge of 15

21

Kotawaringin

Assumed similar with the low monthly river discharge of 15

Sumatra

Java

Kalimantan 13

* Daily rivers discharge provided by: Directorate General for Water Resources, Ministry of Public Works, Indonesia. Several data can be downloaded from http://sda.pu.go.id/index.asp

In order to consider the tidal currents along open boundary, we applied particular values for each compartments during inflow direction and reversely the radiation conditions were applied during outflow direction, for their open lateral boundary conditions. For the particular value, the NO3, PO4, F and TCO2 were set similar with their initial conditions during inflow direction. Since there is no information for Z, PD and NH4 along open boundary, we then assumed that the Z and PD are similar with one-tenth and equal to F, respectively. Particularly for NH4, we applied the equal value to NO3. Time steps of the coupled model specified to the external and internal modes of the physical model are 10 s and 300 s, respectively, and the biogeochemical compartments are updated every 30 min. The coupled model was running totally for three years where the first two-year period was determined as a spin-up period, and then, the third year output was used for analyzing the annual biogeochemical cycle. The three-year simulation showed a quasi-steady state solution in the third year.

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

539

Table 2b. Nutrient concentrations data in the rivers. Station

River Name

Parameters

Period of data collection

Ref.

1

Tulang Bawang

BOD*

Oct, Nov, Dec 1995

2

Seputih

BOD*

Oct, Nov, Dec 1995

3

Sekampung

BOD*

Oct, Nov, Dec 1995

4

Cibanten

NO3, NO2, NH4, PO4

Nov 1995

Booij et al. [33]

5

Ciujung

NO3, NO2, NH4, PO4

Nov 1995

6

Cisadane

NO3, NO2, NH4, PO4

1999, Oct 2000, May 2001

Booij et al. [33] Machbub et al. [34], West Java Local Government [35]

7

Angke

NO3, NO2, NH4, PO4

8

Priok

NO3, NO2, NH4, PO4

9

Marunda

NO3, NO2, NH4, PO4

10

Citarum

NO3, NO2, NH4, PO4

11

Garang

NO3, NO2, NH4, PO4, BOD*

1985-1989 (12 times)

12

Brantas

NO3, NO2, NH4, PO4

May 2001

Sumatra Biro Bina Lingkungan Hidup Setwilda Tk. I Lampung Biro Bina Lingkungan Hidup Setwilda Tk. I Lampung Biro Bina Lingkungan Hidup Setwilda Tk. I Lampung

Java

Dec00, Feb01, Apr01, Jul01, Sep01, Nov01 Dec00, Feb01, Apr01, Jul01, Sep01, Nov01 Dec00, Feb01, Apr01, Jul01, Sep01, Nov01 May, Jun, Aug, Oct, Nov 2002

Damar [5] Damar [5] Damar [5] West Java Environmental Agency Ginting and Rahayu [36] Jennerjahn et al. [37]

Kalimantan 13

Barito

Assumed similar with the nutrient concentration of 14

14

Kapuas

NO3, PO4

Oct 2002

Hadi et al. [38]

15

Kahayan

NO3, NO2, NH4, PO4

Sep 2004, Mar 2005

Haraguchi et al. [39]

16

Sebangau

NO3, NO2, NH4, PO4

Sep 2004, Mar 2005

Haraguchi et al. [39]

17

Katingan

Assumed similar with the nutrient concentrations of 15

18

Mentaya

Assumed similar with the nutrient concentrations of 19

19

Seruyan

NO3, PO4

Hadi et al. [38]

20

Kumai

NO3, PO4

Hadi et al. [38]

21 Kotawaringin Assumed similar with the nutrient concentrations of 20 *) DIN and DIP were estimated by method of San Diego-McGlone et al. [33]

3. Model Results 3.1. Seasonal variability of nutrients riverine input and Chl a distrubutions Variability of river input of DIN (dissolved inorganic nitrogen) and DIP (dissolved inorganic phosphate) is mainly associated with the monsoon seasons as shown in the estimated monthly riverine input of nutrients (DIN and DIP) (Fig. 3). The riverine input is higher in general from November until May, representing rainy season (NWM), than dry season (SEM). DIN loads from Sumatra, Java and Kalimantan Islands are 10 – 20, 16 – 18 and 20 times higher than DIP, respectively. Thus, DIN from river input is generally 20 times higher than DIP. Among the three islands, Kalimantan has the highest contribution of riverine nutrient input due to several bigger rivers located there. On the other hand, concerning the composition of ammonium in DIN form, the sources from Java are dominant among the three islands.

540

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

Fig. 3. Estimated monthly nutrients riverine input in the Java Sea based on several observations: (a) DIN, and (b) DIP. Full line denotes river fluxes from south-coast of Kalimantan; dashed line denotes river fluxes from north-coast of Java; and dotted line denotes river fluxes from southeast-coast of Sumatra.

Seasonal variability of phytoplankton growth (Chl a) and its spatial distribution in the Java Sea are controlled by riverine input and material exchanges with adjacent seas. The riverine input mainly produces phytoplankton bloom along the coastal regions of Sumatra, Java and Kalimantan Islands, as shown by the model results (Fig. 4) and satellite data of surface Chl a (Fig. 5). In this case, the contribution of riverine nutrient input for the phytoplankton growth does not affect the entire Java Sea. The material exchanges, including nutrient, with adjacent seas play important roles in regulating the Chl a distributions and vary following monsoon seasons. The calculated monthly mean tide-and wind-driven circulation (Fig. 4) shows that the water column moves eastward and westward directions, respectively, following the monsoon wind during NWM and SEM. In addition, the current in the whole depth has the same direction. Therefore, the seasonal (monsoonal) flow pattern controls the material distributions in the Java Sea. During SEM, there is evidence of local upwelling at the southern end of the Makassar Strait and

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

541

western Flores Sea [4, 9, 28, 29] Therefore, the surface water containing rich nutrient is transported into the Java Sea during SEM, triggering higher production than NWM.

Fig. 4. The calculated seasonal mean plots of surface Chl a and currents for NWM (DJF) and SEM (JJA) for the cases of coefficient-3 (upper panel) and coefficient-2 (lower panel) in the horizontal eddy viscosity equation.

Model results of Chl a during NWM and SEM (Fig. 4) are qualitatively comparable with the satellite data (Fig. 5), particularly for the use of coefficient 2 for the horizontal eddy viscosity. In this case, the higher coeficient the wider dispersion pattern of material. So that, the tunning of coefficient in the model might be usefull for adjustment with the available satellite data (related with case-1 waters). Generally, model results could reproduce the less concentrations of Chl a during NWM than SEM, as shown in the satellite data. This is consistent with the previous explanation, regarding the upwelling in the eastern adjacent sea. 3.2. Annual mean of material fluxes and distributions In general, the calculated annual mean of material distributions shows high concentrations along the north-coast of Java and south-coast of Kalimantan. The calculated depth-averaged annual mean flow forced by wind and tide, Chl a, primary production, DIN and DIP distributions are shown in Fig. 6. The annual mean flow is in the westward direction with the maximum magnitude of around 0.09 m s-1. The strong currents are found along the north-coast of Java and the south-coast of Kalimantan. The present result is similar with the eight years mean tidal current of Schiller [30], who reported “the residual western boundary current” all along the coastline of the Asian landmass, including the Java Sea. In general, the material (Chl a and nutrients) distributions show influxes from the eastern

542

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

open boundary, following the annual mean flow pattern. The calculated annual mean of Chl a (Fig. 6a) shows high concentrations in the south-coast of Kalimantan, northcoast of Java and southeast-coast of Sumatra. However, the high primary productions (Fig. 6b) are only found at south-coast of Kalimantan and north-coast of Java. In comparison with DIN (Fig. 6c) and DIP (Fig. 6d) distributions, the primary production is controlled by Monod-type calculation in the nutrient uptake (eq. A.9), and does not reflect high primary production along southeast-coast of Sumatra. Therefore, the Chl a in the region is more influenced by horizontal tidal exchange at the western open boundary. As we summarized in the section 2.2, it is noted that the boundary values of F are set similar with their initial conditions during inflow direction or reversely determined by a radiation condition during outflow direction of tidal currents. Further, we examined the calculated annual new productions or Rnew (see Appendix A), as shown in Fig. 7. The value starts from 0 until 1, where the values of 0 and 1 mean dominance in NO3 and NH4, respectively. NO3 domination is affected by external sources, for example, riverine input or adjacent seas, where the riverine NO3 input is more characterized by agricultural activities using fertilizers. On the other hand, the NH4 domination is affected by regeneration processes (organic decomposition) and riverine input (especially in the coastal region/near river mouths) which is more characterized by domestic-urban pollution. Rnew distribution (Fig. 7) clearly shows that Jakarta Bay region is mostly affected by domestic-urban pollution as well as industry (as reported by World Bank [3] and Talaue-McManus [2]), while south-coast Kalimantan tends to be affected by combined domestic wastes and agricultural activities. This is also consistent with the composition of NH4 in the annual riverine DIN input data, as summarized previously.

Fig. 5. The seasonal mean fields of Chl a derived by satellite images (data sources: European Service for Ocean Colour, monthly Chlorophyll-a during 1997-2009, case-1 waters) for NWM (DJF) and SEM (JJA).

The other regions along the north-coast of Java are mostly influenced by riverine NO3 input that may be influenced by domestic and agricultural sources as pointed out by Talaue-McManus [2]. However, the anthropogenic impact seems strong in Jakarta Bay and south-coast of Kalimantan. The other coastal regions are also possibly affected by anthropogenic impacts, but no specific information is available due to lack of data in some river input. In main part of the Java Sea, the Rnew value has a range value below 0.5, so that, the regenerated production is predominant (as summarized before), except for the north-coast of Java Island and western area of the south-coast of Kalimantan.

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

543

Fig. 6. The calculated (a) depth-averaged of annual mean flow and Chl a distribution, (b) primary production, (c) DIN and (d) DIP distributions.

3.3. Seasonal variability of primary production and CO2 fluxes in Jakarta Bay (JB), south-coast of Kalimantan (SK) and central region of the Java Sea In the previous subsection, we examined the anthropogenic impact in JB and SK related to the nutrients riverine input. Now, we focus on the biogeochemical response in JB and SK through model results of primary production and CO2 flux variability (Fig. 8). In this case, the central region of the Java Sea becomes a control point. The high primary production in JB and SK shows a different seasonal cycle. SK is mostly influenced by the high riverine input during the rainy season of NWM (Dec-Mar), while JB is mostly controlled by the high riverine input and Chl a distribution from neighbour regions which is transported by westward currents during SEM. In addition, current velocity during SEM is higher than NWM, thus it affects the advection processes of nutrients and Chl a distribution from neighbour regions of JB. On the other hand, the central region shows higher primary production almost for the whole year, except for the transition period during post NWM. This is reasonable since the Java Sea receives many nutrients during SEM until NWM, as discussed in the section 3.1. There is no information about the observed primary productions in these three regions temporally and spatially, so that the seasonal variability of primary production in JB, SK and central part are more qualitatively described and interpreted. Our concern is more focused on the anthropogenic impact in JB and SK, caused by domestic-urban pollution, agricultural activities and domestic wastes. The highest primary production during SEM in JB is caused by upstream conditions of river influxes in the eastern part of JB. On the other hand, the variability in SK is affected by several river inputs. The primary production in the central part varies seasonally based on monsoonal flow patterns.

544

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

-2

De

Ja

Fe

Ma

Ap

Ma

Ju

Ju

Au

Se

Oc

mg C m hr

-2

mg C m hr

-1

-1

Fig. 7. The calculated annual mean field of new productions (Rnew). JB (Jakarta Bay) and SK (south-coast of Kalimantan) regions indicate the anthropogenic impact.

No

Fig. 8. The calculated annual cycle of primary production in selected positions within the Java Sea: JB (full line) and SK (bold full line) are scaled in left y-axis, and the central region (dotted line) is scaled in right y-axis.

The calculated CO2 fluxes and TCO2 variability in JB, SK and central part are shown in Fig. 9. Positive and negative CO2 fluxes denote release to atmosphere and oceanic uptake of CO2, respectively. Daily fluctuations of CO2 fluxes show a strong influence of tidal exchange. In general, the monthly moving average of CO2 fluxes is strongly associated with seasonal variability of TCO2 and primary production (Fig. 9). The higher primary production requires high nutrient and TCO2 assimilations (indirectly via Redfield ratio in the model) and leads to an under saturation condition of CO2 in the surface ocean (i.e. low pCO2). Therefore, SK and JB regions act as a sink for the atmospheric CO2 in NWM and SEM, respectively. JB region uptakes CO2 in longer periods of April – Dec associated with variability of primary production. For the annual mean of CO2 fluxes, the entire Java Sea acts as a source for the atmospheric CO2 in level amount of 4804 x 109 mol/year. This flux is more likely consequence of warming of subsurface water due to high SST (sea surface temperature), as a typical situation in the tropical region.

545

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

-2

De

Ja

Fe

Ma

Ap

Ma

Ju

Ju

Au

Se

Oc

No

Dec

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

De

Ja

Fe

Ma

Ap

Ma

Ju

Ju

Au

Se

Oc

No

mmol C m hr

mol C m

-3

-1

(a)

-2

mmol C m hr

mol C m

-3

-1

(b)

-2

mmol C m hr

mol C m

-3

-1

(c

Fig. 9. Model results of the annual cycles of CO2 fluxes (lower graph, right y-axis) and TCO2 concentration (upper graph, left y-axis) in (a) SK, (b) JB and (c) central region. Bold line in the middle is a representative of moving average (1-month). Positive (negative) net fluxes of CO2 denote a source (sink).

3.4. Sensitivity studies of phytoplankton biomass and CO2 fluxes To better understand the sensitivity of the models to the choice of parameter values, it is common to perform a sensitivity analysis using the equation, as follows:

S

'C / C 'Parameter / Parameter

(1)

546

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

S is a sensitivity index and C is concentration of biomass system/compartment. Negative S means an opposite tendency between a parameter and the system biomass. The sensitivity analysis shows the relative importance of the processes that play crucial roles in the biogeochemical cycles. Therefore, the changes in the parameters could be associated with the changes in natural or anthropogenic forcing mechanisms. We now focus on the sensitivity of phytoplankton biomass and CO2 fluxes (FCO2) by changing some parameters, as shown in Table 3. We additionally examine also the role of benthic flux by ignoring BD from the model system (biogeochemical compartments). For this purpose the model was run for two months, and the subsequent solution averaged over 5-K1 tidal periods is used for the sensitivity analysis. February is selected as a reference month for the physical-biogeochemical forcing mechanisms. For the phytoplankton biomass (Table 3a), the maximum nutrient uptake rate of phytoplankton is the most crucial parameter, and the second one is optimum light for phytoplankton growth. Their contributions account for more than 5% of the changes in phytoplankton biomass when the parameters are changed by 50%. In addition, the optimum light for phytoplankton growth has an opposite effect. On the other hand, the omission of BD compartment could reduce phytoplankton biomass by 7.8%. Table 3a. Sensitivity analysis of phytoplankton biomass. Parameter Zooplankton excretion/mortality Decomposition of pelagic detritus Decomposition of benthic detritus Optimum light for phytoplankton growth Maximum phytoplankton uptake rate Non-BD

Change rate of parameter (%)

Change rate of phytoplankton biomass (%)

Sensitivity index

50

0.51

0.01

50

0.03

0.00

50

2.44

0.05

50

-5.09

-0.1

50

13.23

0.26

-

-7.83

-

The same procedure was also applied to the FCO2 (Table 3b). We found that the FCO2 is more sensitive to the total alkalinity (as a fixed parameter or constant in the carbonate system model) and the atmospheric CO2. A small change of total alkalinity (5.6%) could reduce FCO2 for 81%, while the change in atmospheric CO2 (50 %) could reduce FCO2 to 48 %. Non-BD compartment only contributes to 2.3 % of the FCO2. This is contrast with the general mechanism according to carbonate model of OCMIP [40] in the enclosed system, where TCO2 and sea surface temperature are the most sensitivity parameters. Table 3b. Sensitivity analysis of CO2 fluxes (FCO2). Parameter

Change rate of parameter (%)

Change rate of FCO2 (%)

Sensitivity index

Zooplankton excretion/mortality

50

0.02

0.00

Decomposition of pelagic detritus

50

0.01

0.00

Decomposition of benthic detritus

50

0.34

0.01

Optimum light for phytoplankton growth Maximum phytoplankton uptake rate Atmospheric CO2

50

0.54

0.01

50

-1.46

-0.03

50

-47.86

-0.96

Total alkalinity

5.6

-80.88

-14.44

Non-BD

-

-

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

Based on the selected parameters for the sensitivity studies, the implication can be applied to other compartments, such as nutrients. The increased phytoplankton biomass due to an increase in the nutrient uptake rate will be followed by a decrease in nutrient, an increase in grazing activity, so that the mortality of phytoplankton and mortality/excretion of zooplankton will also increase. However, the change in algal biomass would not affect significantly to the decomposition processes because of its low contribution as shown in Table 3a. As for the effect of the nutrient uptake rate for phytoplankton on the CO2 flux (Table 3b), even though it has small contribution in the present result, the external forcing of nutrient loading from rivers plays a key role in carbon uptake by phytoplankton as discussed previously in the section 3.1. 4. Summary and Discussion The coupled physical-biogeochemical model has been demonstrated in investigating the physical influences on the biogeochemical cycle in the Java Sea. The coupled model could reproduce the basic condition of seasonal variability of surface Chl a distribution which is controlled by the monsoon events. The model results were also comparable with the satellite image data. Model results and satellite data clearly showed that the concentrations of Chl a in the Java Sea are mainly influenced by nutrients supply from the eastern boundary due to upwelling during SEM. On the other hand, the rainy season during NWM triggers much nutrients load through river discharges, resulting in phytoplankton bloom along the regions near coastline/river mouths. However, the effect of nutrient riverine input is only limited in the regions. As a consequence of the Monod-type, N and P are capable of limiting factors for the phytoplankton growth. Some regions showed a strong limitation of P, such as along the north-coast of Java and western area in the southcoast of Kalimantan. In this case, the P limitation is a consequence of high N riverine input. Compared to the offshore region, the entire Java Sea showed a strong N limitation. In particular of Jakarta Bay, the P limitation in the present result is consistent with the result of Damar [5] based on laboratory experiments. Jakarta bay is now facing a serious environmental impact due to human activities (anthropogenic impact). The riverine N input in the bay is more dominated by NH4 as a product of domestic-urban and industry pollutions. Damar [5] pointed out that Jakarta Bay region is now under hyper-eutrophic condition. The anthropogenic impact also seems serious in south-coast of Kalimantan. Besides the big rivers which have more capacity to transport nutrient, the specific environmental problem in that region is due to human activities in opening wetland. Galloway et al. [41] summarized that wetlands are so efficient in removing N reactive (inorganic and organic forms) through denitrification. However, the wetland removal can decrease water residence time and thus decrease denitrification within a river. Therefore, the coastal system will receive more N through the wetlandriver-coastal system and causing the eutrophication problems. The anthropogenic impact in Jakarta Bay and the south-coast of Kalimantan also could induce CO2 uptake (through biological uptake). However, the river and regeneration fluxes as well as the relatively warm water in the tropical region produce oversaturation condition of CO2 in the Java Sea. When we try to project the future tendency of this sensitive system, the present model needs an improvement in some compartments. Related to the carbonate system, TCO2, total alkalinity, pH and fugacity of CO2 are important parameters [42], since the individual species of the CO2 system in seawater cannot be measured directly. In the present simulation, we performed the total alkalinity as a diagnostic parameter with a fixed value. In addition, the river fluxes of TCO2 and organic carbon were not directly calculated in the model. At this stage, we showed the results of the sensitivity analysis, suggesting the importance of total alkalinity. By increasing a small amount of total alkalinity, it leads to large reduction in FCO2. Therefore, in spite of a lack of explicit inclusion by many biogeochemical coastal models, the information of total alkalinity from riverine input is more crucial in order to improve the present carbonate system model in future. The other necessary information is the specific organic carbon transport from the rivers into the Java Sea. On the other hand, the use of cohesive sediment processes is not enough to resolve the nutrient exchange of the water-sediment interaction in the bottom layer. This is important in maintaining the high concentration of Chl a in the bottom layer, as shown in the field observation but not reproduced in this model. Since the cohesive sediment model only accommodates the physical processes, the use of other model, such as parameterization in biological and chemical processes within sediment is crucial for model improvement in future. In order to describe the water-

547

548

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

sediment interaction, the model has to include temperature, salinity and consequent density distributions and perform more precise calculations of vertical mixing intensity near the bottom boundary layer. Acknowledgement We wish to thank Dr. Yasuhiro Yamanaka and Dr. Koji Suzuki for their valuable comments. This study is also partially supported by the JSPS Core University Program, collaboration between Hokkaido University and Research Center for Biology-the Indonesian Institute of Science. The first author is awarded a scholarship by Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.

Appendix A. Governing equations and formulations of the biogeochemical model The governing equations of this model are given as:

wNO3  ADV  DIF  Oxd  Rnew A1 F wt wNH 4  ADV  DIF  (1  Rnew ) A1 F  Oxd  A4 Z  A5 PD wt wPO4  ADV  DIF   A1 F  A4 Z  A5 PD rNP wt wF  ADV  DIF  A1 F  A2 F  A3 Z wt wZ  ADV  DIF  1  J A3 Z  A4 Z wt wPD  ADV  DIF  JA3 Z  A4 Z  A5 PD wt wBD ( S  E )  R BD BD wt wNO3 wNH 4 wTCO 2 ( )rCN  wt wt wt

(2) (3) (4) (5) (6) (7) (8) (9)

ADV and DIF are advection and diffusion processes in three-dimensional model, respectively. As a particular role of PD, the vertical advection includes also the vertical sinking velocity of PD (Wd). BD is applied as an ordinary differential equation which functions as a pool for sinking of PD.

A.1. Nutrient uptake (A1F) Nutrient uptake is dependent on nutrient concentrations, solar radiation and phytoplankton stock sizes [17, 18]. Individual of nutrient species adopted Michaelis-Menten formula (hyperbolic saturation function), while the nutrient limitation is a Monod-type (minimum concentration). The formula of nutrient uptake is as follows:

549

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

§ NO3 NH 4 PO 4 G Min¨¨ exp
A1 F

G

· ¸¸ F ¹

(10)

(11)

Steele [43] formula is adopted to the solar radiation or light intensity in the surface (I) which is variable to water depth: z I 0 exp§¨  ³ Ndz ·¸ ¹ © 0

I

(12)

Light extinction coefficient (ȡ) is a function of light attenuation by seawater and phytoplankton (shelf-shading):

N D1  D 2 F (13) A.2. Natural mortality of phytoplankton (A2F)

A2 F

mF F

(14)

A.3. Grazing (A3Z) Grazing is a function of phytoplankton and zooplankton, as follows:

A3 Z

R m 1  exp /F Z

(15)

Egestion is assumed to be proportional (J) to grazing activities. A.4. Natural mortality of zooplankton (A4Z)

A4 Z

mZ Z

(16)

A.5. Decomposition rate of pelagic detritus (A5PD)

A5 PD

RPD PD

(17)

A.6. Decomposition rate of benthic detritus (A6BD)

A6 BD

RBD BD

(18)

The rate of benthic fluxes are assumed to be proportional for C, N, and P as determined by the Redfield ratio. Thus, we applied the benthic fluxes at bottom boundary condition for NH4, PO4 and TCO2. A.7. Ammonium oxidation rate (Oxd)

Oxd

R N 2 NH 4

(19)

550

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

The new production (Rnew) is ratio of NO3 uptake from the total N uptake, as follows:

R new

NO3 exp
(20)

The bottom boundary condition for PD is determined by cohesive sediment process, where its summary and parameter values can be found in Koropitan et al. [19]. The parameter definitions and values of the above formulae are given in Table 1 and the text. A.8. Air-sea flux of CO2 (FCO2) The carbonate system model was fully adopted from OCMIP-3, as follows: FCO2=Kw (CO2air* – CO2surf*)

(21)

Kw = (1.09W – 0.333W2 + 0.078W3) (Sc/660)-0.5

(22)

Kw is a cubic function of the wind speed (W) using climatological winds [44]. Sc is the Schmidt number. CO2air* = D pCO2atm P/Po

(23)

where D is the CO2 solubility for water-vapor saturated air [mol/(m3 Patm)]; pCO2atm is the partial pressure of CO2 in dry air at one atmosphere total pressure (in Patm); P is the total air pressure at sea level (atm), locally (assumed to be 1 atm); P0 is 1 atm. In this simulation, pCO2atm are obtained by converting the value of atmospheric mole fraction CO2 in dry air (given as a fixed value). CO2surf* is calculated from the model’s surface TCO2, sea surface temperature (SST), sea surface salinity (SSS), and surface total alkalinity. In this simulation, SST, SSS and surface total alkalinity are given (fixed). Surface boundary condition for TCO2 is treated as a combination of FCO2 and virtual flux (Fv), given by: Fv = TCO2g (E - P)

(24)

where TCO2g is globally averaged surface concentrations of TCO2, E and P are evaporation and precipitation, respectively. Evaporation data are collected from monthly global data [25] which is averaged for only the Java Sea region. Precipitation data are plotted from the single graph of the annual cycle of precipitation in the Java Sea [45] in order to get the monthly data. They used several precipitation data from the several ground stations in Indonesia and made a classification into three main regions using EOF. Precipitation data of the Java Sea are included in one of the main regions which have strong monsoon impact. References 1. 2. 3. 4. 5.

Koropitan AF, Ikeda M. Three-dimensional modelling of tidal circulation and mixing over the Java Sea. Journal of Oceanography 2008;64(1):61-80. Talaue-McManus L. Transboundary diagnostic analysis for the South China Sea. EAS/RCU Technical Report Series No. 14. Bangkok: UNEP; 2000: 105 pp. World Bank. Indonesia forest, land and water; Issues in Sustainable Development UNDP. World Bank Report No. 7822-IND; 1989. Susanto RD, Moore II TS, Marra J. Ocean color variability in the Indonesian Seas during the SeaWiFS era. Geochemistry Geophysics Geosystems 2006;7:Q05021, doi:10.1029/2005GC001009. Damar A. Effects of enrichment on nutrient dynamics, phytoplankton dynamics and productivity in Indonesian tropical waters: a comparison between Jakarta Bay, Lampung Bay and Semangka Bay. PhD dissertation. Kiel: Christian-Albrechts-Universität; 2003.

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

28. 29. 30. 31.

32. 33. 34. 35. 36. 37.

551

Cai WJ, Dai M. Comment on “Enhanced open ocean storage of CO2 from shelfsea pumping”. Science 2004;306:1477c. Thomas H, Bozec Y, Elkalay K, de Baar HJW. Enhanced open ocean storage of CO2 from Shelf Sea pumping. Science 2004;304:1005-8. Feely RA, Sabine CL, Takahashi T, Wannonkhof R. Uptake and storage of carbon dioxide in the ocean: the global CO2 survey, special issue – JGOFS. Oceanography 2001;14(4):6-17. Wyrtki K. Physical oceanography of the Southeast Asian Waters, NAGA Report Volume 2, Scripps Institution of Oceanography. California: University of California; 1961:195 pp. Ningsih NS, Yamashita T, Aouf L. Three-dimensional simulation of water circulation in the Java Sea: Influence of wind waves on surface and bottom stresses. Natural Hazards 2000;21:145-71. Blumberg AF, Mellor GL. A description of a three dimensional coastal ocean circulation model. In Three-Dimensional Coastal Ocean Models, pp. 1-16. ed. by N. Heaps. Washington DC: American Geophysical Union; 1987. Matsumoto K, Ooe M, Sato T, Segawa J. Harmonic analysis of data from crossover points. A Collection of Global Ocean Tide Models CD ROM, distributed by JPL, PO. DAAC and CSR; 1996. Mellor GL, Yamada T. Development of a turbulence closure model for geophysical fluid problems. Review of Geophysics 1982;20:851–75. Smagorinsky J. Some historical remarks on the use of nonlinear viscosities. In Large Eddy Simulation of Complex Engineering and Geophysical Flows, pp. 3-36, ed. by. B. Galperin, S. A. Orszag. New York: Cambridge University Press; 1993. Griffies SM, Hallberg RW. Biharmonic friction with a Smagorinsky-like viscosity for use in large-scale eddy-permitting ocean models. Monthly Weather Review 2000;128:2935–46. Smagorinsky J. General circulation experiments with primitive equations. I. The basic experiment. Monthly Weather Review 1963;91:99– 164. Wroblewski JS. A model of phytoplankton plume formation during variable Oregon upwelling. Journal of Marine Research 1977;35:357-94. Denman KL, Peña MA. A coupled 1-D biological/physical model of the northeast subarctic Ocean Pacific with iron limitation. Deep Sea Research Part II 1999;46:2877-908. Koropitan AF, Ikeda M, Damar A, Yamanaka Y. Influences of physical processes on the ecosystem of Jakarta Bay: a coupled physicalecosystem model experiment. ICES Journal of Marine Science 2009;66:336-48. Rieman B, Simonsen P, Stensgaard L. The carbon and chlorophyll content of phytoplankton from various nutrient regimes. Journal of Plankton Research 1989;11(5):1037-45. Newberger PA, Allen JS, Spitz YH. Analysis and comparison of three ecosystem models. Journal of Geophysical Research 2003;108(C3):3061. doi:10.1029/2001JC001182. Wei H, Sun J, Moll A, Zhao L. Phytoplankton dynamics in the Bohai Sea-observations and modeling. Journal of Marine Systems 2004;44:233-51. Hellerman S, Rosenstein M. Normal monthly wind stress over the world ocean with error estimates. Journal of Physical Oceanography 1983;13:1093-104. Oberhuber JM. An Atlas Based on `COADS' Data Set, number 15; 1988. da Silva A, Young AC, Levitus S. Atlas of surface marine data 1994, volume 1: algorithms and procedures. NOAA Atlas NESDIS 6. Washington DC: U.S. Department of Commerce; 1994. Asanuma I, Tanaka T, Matsumoto K, Kawano T. Primary productivity model based on photosynthetically available radiation. In: R. J. Frouin, H. Kawamura, and M. Kishino (Eds.). Hyperspectral Remote Sensing of the Ocean Proc. SPIE 2001;4154:153-8. Goyet C, Healy RJ, Ryan JP. Global distribution of total inorganic carbon and total alkalinity below the deepest winter mixed layer depths. ORNL/CDIAC-127, NDP-076. Tennessee: Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, U.S. Department of Energy, Oak Ridge; 2000. Ilahude AG. On the occurrence of upwelling in the southern Makassar Strait. Marine Research in Indonesia 1970;10:3-53. Birowo S. Possible occurrence of upwelling in the Flores Sea and Bone Bay. Oseanologi di Indonesia 1979;12:1 – 12 (in Bahasa with English abstract). Schiller A. Effects of explicit tidal forcing in an OGCM on the water-mass structure and circulation in the Indonesian throughflow region. Ocean Modelling 2004;6:31-49. Tachibana H, Kano A, Iqbal R, Mori A, Hasegawa K, Saeki H, Takahashi H, Utosawa K, Sumawijaya N, Darung U, Dohong S, Limin S. Water quality and runoff characteristics of chemical components in the Sebangau River: Influence of the development of tropical peatland. In Anual Report for April 2003 - March 2004. Environmental Conservation and Land Use Management of Wetland Ecosystem in Southeast Asia. Core University Program between Hokkaido University, Japan and Research Center for Bilogy, LIPI, Indonesia Sponsored by JSPS; 2004. San Diego-McGlone ML, Smith SV, Nicolas VF. Stoichiometric interpretations of C:N:P ratios in organic waste materials. Marine Pollution Bulletin 2000;40(4):325-30. Booij K, Theo M, Hillebrand J, Nolting RF, van Ooijen J. Nutrients, trace metals, and organic contaminants in Banten Bay, Indonesia. Marine Pollution Bulletin 2000;42(11):1187-90. Machbub B, Moelyo M, Yusuf A, Irianto EW, Sudarna A, Suganda L, Ilyas MT. Pengelolaan kualitas air Cisadane. Research Report. Bandung: PUSLITBANG Sumber Daya Air, Departemen Permukiman dan Prasana Wilayah; 2001 (in Bahasa). West Java Local Government. Neraca kualitas lingkungan hidup daerah (NKLD) tahun 2000. 3rd ed. Bandung: West Java Local Government; 2000. Ginting NT, Rahayu S. Potensi alami dan perubahan kualitas air di Pulau Jawa. Research Report. Bandung: PUSLITBANG Pengairan, Departemen Pekerjaan Umum; 1989 (in Bahasa). Jennerjahn TC, Ittekkot V, Klöpper S, Adi S, Nugroho SP, Sudiana N, Yusmal A, Prihartanto, Gaye-Haake B. Biogeochemistry of a tropical

552

38.

39. 40. 41. 42. 43. 44. 45.

Alan F. Koropitan and Motoyoshi Ikeda / Procedia Environmental Sciences 33 (2016) 532 – 552

river alected by human activities in its catchment: Brantas River estuary and coastal waters of Madura Strait, Java, Indonesia. Estuarine. Coastal and Shelf Science 2004;60:503-14. Hadi S, Ningsih NS, Pranowo WS, Achmad H, Ramadhan MA, Fauzie D, Sunendar H, Muliadi, Huda A, Sodikin K, Ali H, Cempaka R, Asparini M, Gusman AR, Berlianty D. Pengumpulan data dan informasi untuk MCMA Propinsi Kalimantan Tengah. Research Report. Bandung: Pusat Penelitian Kelautan-LPPM-ITB; 2002 (in Bahasa). Haraguchi et al 2007 Orr J, Najjar R, Sabine C, Joos F. Abiotic-HOWTO. Internal OCMIP Report, pp. 25. Gif-sur-Yvette, France: LSCE/CEA Saclay; 1999. Galloway JN, Aber JD, Erisman JW, Seitzinger SP, Howarth RW, Cowling EB, Cosby BJ. The nitrogen cascade. Bioscience 2003;53(4):341-56. Dickson AG, Sabine CL, Christian JR. Guide to best practices for ocean CO2 measurements. PICES Special Publication; 2007:3:191 pp. Steele JH. Environmental control of photosynthesis in the sea. Limnology and Oceanography 1962;7:137–50. Wanninkhof R, McGillis WR. A cubic relationship between air–sea CO2 exchange and wind speed. Geophysical Research Letter 1999;26(13):1889–92. Aldrian E, Susanto RD. Identification of three dominant rainfall regions within Indonesia and their relationship to seas surface temperature. International Journal of Climatology 2003;23:1435-52.