A numerical investigation of eddy-induced chlorophyll bloom in the southeastern tropical Indian Ocean during Indian Ocean Dipole 2006

Transcription

1 Ocean Dynamics (2010) 60: DOI /s A numerical investigation of eddy-induced chlorophyll bloom in the southeastern tropical Indian Ocean during Indian Ocean Dipole 2006 Iskhaq Iskandar & Hideharu Sasaki & Yoshikazu Sasai & Yukio Masumoto & Keisuke Mizuno Received: 7 September 2009 /Accepted: 12 April 2010 /Published online: 2 May 2010 # Springer-Verlag 2010 Abstract An eddy-resolving coupled physical biological model is used to study the effect of cyclonic eddy in enhancing offshore chlorophyll-a (Chl-a) bloom in the southeastern tropical Indian Ocean during boreal summer fall The results demonstrate that the offshore Chl-a blooms are markedly coincident with the high eddy kinetic energy. Moreover, the vertical variations in Chl-a, nitrate, temperature, and mixed-layer depth (MLD) strongly imply that the cyclonic eddies induce surface Chl-a bloom through the injection of nutrient-rich water into the upper layer. Interestingly, we found that the surface bloom only occurs when the deep Chl-a maximum is located within the MLD. On the other hand, the response of subsurface Chl-a Responsible Editor: Jin-Song von Storch I. Iskandar : Y. Sasai : Y. Masumoto : K. Mizuno Research Institute for Global Change, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokosuka, Japan I. Iskandar On leave from Jurusan Fisika, FMIPA, Universitas Sriwijaya, Palembang, Sumatra Selatan, Indonesia H. Sasaki Earth Simulator Center, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokosuka, Japan Y. Masumoto Also at Department of Earth and Planetary Science, Graduate School of Science, The University of Tokyo, Tokyo, Japan I. Iskandar (*) NOAA/PMEL, 7600 Sand Point Way, NE, Seattle, WA 98115, USA Iskhaq.Iskandar@noaa.gov to the eddy pumping is remarkable, although it is hardly observable at the surface. Keywords Chlorophyll-a. Upwelling. Coupled physical biological model. Indian Ocean Dipole. Southeastern tropical Indian Ocean 1 Introduction The southeastern tropical Indian Ocean (SETIO) between the Lesser Sunda Islands and the northwestern Australia represents a water mass crossroad, as several different water masses from the Indian Ocean, the Pacific Ocean, and the Indonesian seas meet in this area (Fieux et al. 1994). The oceanic circulation in the SETIO region is mainly driven by the seasonally varying monsoonal winds (Quadfasel and Cresswell 1992). During the southeast monsoon season, prevailing southeasterly winds drive offshore Ekman transport off south Java and Sumatra resulting in cold sea surface temperature (SST) there. Satellite observed surface chlorophyll-a (Chl-a) shows that this upwelling circulation, through the supply of nutrient-rich subsurface water, elevates Chl-a concentrations (Asanuma et al. 2003; Susanto et al. 2006). The situation reverses for the northwest monsoon season. The SETIO region is also influenced by an interannual air sea coupled climate mode inherent in the tropical Indian Ocean, so-called the Indian Ocean Dipole (IOD) (Saji et al. 1999; Webster et al. 1999; Murtugudde et al. 2000). A positive IOD event is characterized by a pattern of cool SST anomalies in the eastern Indian Ocean and warm anomalies in the west associated with anomalously strong easterly along the equator and southeasterly along the southern coast of Sumatra and Java. The IOD event is

2 732 Ocean Dynamics (2010) 60: seasonally phase-locked, in which it develops during June July, comes to its peak in September October, and terminates in November December. Therefore, the IOD events are accompanied by a strong upwelling off Java and Sumatra, leading to Chl-a blooms (Murtugudde et al. 1999; Susanto and Marra 2005; Iskandar et al. 2009). In addition, this region is also known as a region of strong eddy activity (Feng and Wijffels 2002; Yu and Potemra 2006). It was suggested that baroclinic instability associated with the local instability of the current system in this region (e.g. South Equatorial Current and the Indonesian throughflow) plays a crucial role in generating eddy in the SETIO. Recent study has demonstrated that the eddy activities in the SETIO region are stronger during the positive IOD event (Ogata and Masumoto 2009) Researches carried out in the 1990s have suggested that upwelling of nutrients by cyclonic mesoscale eddies (eddy pumping) could supply the nutrient-rich water into the upper layer (McGillicuddy and Robinson 1997; Oschlies and Arcon 1998). The cyclonic eddy retains water in its core and then upwells nutrients when it forms. Of particular interest for the present study, the role of eddies on the offshore Chl-a bloom off south Java during 2006 has been documented (Iskandar et al. 2009). However, because of the restriction of the satellite data, detailed dynamics of such processes is not clear. The objective of the present study is, therefore, to investigate possible mechanism of the offshore Chl-a bloom in the SETIO during 2006 when the positive IOD event took place in the tropical Indian Ocean (Vinayachandran et al. 2007; Horii et al. 2008) using output from an eddy-resolving coupled physical biological ocean model. The paper is organized as follows. In Section 2, we describe the model used in this study. Section 3 describes the seasonal and interannual variability in Chl-a concentration, and the role of eddy in generating offshore Chl-a bloom during A conclusion is provided in the final section. 2 Physical and biological model 2.1 Model description The physical model is an eddy-resolving ocean model that particularly tune for the Earth Simulator (OFES; Ocean general circulation model For the Earth Simulator). The model covers near-global region extending from 75 S to 75 N, with uniform horizontal resolution of 0.1 in both longitude and latitude. There are 54 levels in vertical, of which 20 levels are located in the upper 200-m depth. A biharmonic scheme is used for the horizontal mixing, whereas the K-Profile Parameterization scheme is adopted for the vertical mixing process. The reader is referred to the work of Masumoto et al. (2004) and Sasaki et al. (2006) for a more complete description of the physical model. The ocean general circulation model is coupled with a biological model of a simple nitrogen-based model, which has four compartments, namely, nitrate, phytoplankton, zooplankton, and detritus (Oschlies 2001). The concentration of the biological tracer is determined as a balance of advection diffusion and source sink terms among those compartments (Sasai et al. 2006). The initial concentrations for the phytoplankton, zooplankton, and detritus are 0.14, 0.014, and 10-4 mmol N m -3, respectively. Note that the concentrations of phytoplankton and zooplankton are decreasing exponentially from the surface with a scale depth of 100 m (Sarmiento et al. 1993). On the other hand, the initial condition for the nitrate field is taken from WOA Model run The physical model is, at first, spun up from a state of rest for a period of 50 years under the climatological mean forcing of NCEP/NCAR reanalysis data. Then, the model is driven by daily mean forcing from 1950 to The surface heat and fresh water flux were specified using bulk formula with atmospheric data obtained from the NCEP/ NCAR reanalysis. In addition, the surface salinity is relaxed to the monthly mean climatology of the WOA98 to include the contribution from the river runoff. Similar relaxation to the climatological monthly temperature and salinity is also applied near the artificial boundary at 75 S and 75 N. To establish the stable condition of the biological system, the biological model is coupled to the physical model, and then, it is integrated for a 5-year period under the climatological mean forcing. The biological fields from the fifth year of climatological simulation are used for the initial condition of the daily coupled simulation. On the other hand, the initial condition for the physical fields is from the end of 1998 of the daily simulation. The coupled model is then integrated from 20 July 1999 to 2006 using daily mean surface wind stress from the QSCAT satellite and the atmospheric daily mean data (heat and salinity fluxes) of NCEP/NCAR reanalysis. In this study, we used the output from January 2002 through December 2006, encompassing strong IOD event in Results 3.1 Seasonal variations First, we describe the seasonal cycle associated with the coastal upwelling off south Java. Understanding the

3 Ocean Dynamics (2010) 60: seasonal cycle is important to properly elucidate the variability at other time scales. Monthly climatology of the observed surface Chl-a obtained from Sea-viewing Wide-Field of view Sensor (SeaWiFS) and that from model output are shown in Figs. 1 and 2, respectively. Note that the climatological fields of the simulations and observations are computed over the period of January 2002 to December The concentration of Chl-a in the model is converted from the model phytoplankton using a constant conversion factor of 1.59 mgchl/mmoln. The observed SST and winds are presented in Fig. 3, whereas model SST and surface currents are shown in Fig. 4. It is shown that both observed (Fig. 1) and model (Fig. 2) Chl-a indicate distinct seasonal variation associated with the Asian Australian monsoon. The Chl-a concentration is high (low) during the southeast (northwest) monsoon season in June October (December March). The southeasterly winds during the southeast monsoon season force offshore Ekman transport associated with upwelling of nutrient-rich subsurface water leading to high Chl-a concentration. In addition, these southeasterly winds drive westward surface currents along the southern coast of Java (Fig. 4), which is known as the South Java Coastal Current (SJCC; Wyrtki 1961; Iskandar et al. 2006). The situation is reversed during the northwest monsoon season. The evolution of high Chl-a concentration is first observed in June in the vicinity of Nusa Tenggara Island chain (Fig. 1c), when the coastal upwelling starts to develop (Fig. 3c). As the southeasterly winds strengthen in the following month, the coastal upwelling also intensified and peaked in August (Figs. 3d and 4d). As a result, Chl-a concentration increased significantly by August and expanded approximately 200 km offshore forming a continuous band of high concentration along the southern coast of Indonesia (Figs. 1d and 2d). By December, the southeasterly winds are weakened and gradually reversed direction to the southeastward (Fig. 3d). The northeasterly winds then force downwelling along the coast leading to high SST (Figs. 3d and 4d) and low Chl-a concentration (Figs. 1d and 2d). In addition, the northeasterly winds drive eastward SJCC along the southern coast of Java (Fig. 4). The Chl-a concentration (SST) remains low (high) during the downwelling season in December March. 3.2 Eddies and high chlorophyll concentrations off south Java As shown in the previous section, high Chl-a concentration near the coast is related to the monsoonal wind variability and upwelling centers. In addition to the high Chl-a concentration near the coast, both observation and model show offshore extension of high Chl-a concentration. Previous studies have shown that Rossby waves, mesoscale eddies, and Ekman pumping play a significant role in generating high Chl-a concentration in the open ocean where the surface layer is nutrient-limited (Oschlies and Arcon 1998; Cipollini et al. 2001). Both observation and a February b April c June d August e October f December Fig. 1 Distribution of monthly climatology of satellite-observed surface Chl-a concentration (mg m -3 ) shown in bimonthly interval

4 734 Ocean Dynamics (2010) 60: a February b April c June d August e October f December Fig. 2 As in Fig. 1 except for model output numerical studies have suggested that there is high eddy activity in the SETIO during the second half of the year (Feng and Wijffels 2002; Yu and Potemra 2006). To investigate the relation between eddy activity and the surface Chl-a distribution in this region, we first calculate the seasonal standard deviation of the surface Chl-a and sea surface height (SSH). It is found that high Chl-a variation in the offshore region is observed during the September November season. The standard deviation of surface Chl-a shows that a relatively large magnitude of a February b April c June d August e October f December Fig. 3 Monthly climatology of the observed winds (vectors, inms -1 ) superimposed on the observed SST (shading, C). The fields are shown in bimonthly interval 1.5 m/s

5 Ocean Dynamics (2010) 60: a February b April c June d August e October f December Fig. 4 As in Fig. 4 except for the simulated surface currents (vectors, in m s-1) and the SST (shading, C) variation in the offshore region is observed south of about 10 S (Fig. 5a b). The Chl-a also shows a distinct location of high Chl-a variation located along the southern coast of Java and the Lesser Sunda Islands. The corresponding SSH variability also shows two regions of large amplifig. 5 Standard deviations of the a observed and b model surface Chl-a (mg m-3) during the September November season. Panels (c) and (d) same as in panel (a) except for the observed and model SSH (cm), respectively. The box shown in panel (c) is the region for the area-averaged plot: Figure 6. Note that the standard deviations for the observations are calculated from weekly data, whereas those for the models are derived from 3-day snapshots tude: in the offshore south of 11 S and along the coastal waveguide (Fig. 5c d). Overall, the observed and model simulation are in a good agreement with one another, although the model SSH variability is larger compared to the observed one. a Obs. Chl-a (log) b Model Chl-a (log) c Obs. SSH d Model SSH

6 736 Ocean Dynamics (2010) 60: To examine the relation between offshore Chl-a bloom and eddy activity, we average quantities in a rectangular region enclosing the high SSH variability (marked by a rectangular in Fig. 5c). The top panel in Fig. 6 shows evolution of the observed surface Chl-a and eddy kinetic energy (EKE) calculated from the geostrophy velocity of the observed SSH. The bottom panel shows the corresponding Chl-a and EKE from model output. In both observation and model, the EKE is maximum during September November, whereas the Chl-a is maximum during July September. However, there are some occasions that the maximum Chl-a concentration co-occurs with the maximum EKE as in August September 2003, September 2004, and August October This may suggest that the eddy activity plays a role in enhancing surface Chl-a concentration in this region. In the next section, we will focus on the co-occurrence of high Chl-a concentration and strong eddy activity during boreal fall Offshore chlorophyll bloom during 2006 To illustrate the offshore bloom in the SETIO off south Java during the boreal fall 2006, we first showed the observed Chl-a concentrations from SeaWiFS and those from our model (Fig. 7a b). Although the most notable feature captured by both SeaWiFS and model is the high concentration of Chl-a associated with the upwelling along the southern coast of Java and Sumatra (Iskandar et al. 2009), another intriguing feature is the offshore bloom south of about 9 S. Owing to cloud cover, however, the satellite image does not cover the complete distribution of the offshore Chl-a bloom. In the model, high concentration Chl-a (log) Chl-a (log) a Observation b Model EKE EKE Chl-a Chl-a Fig. 6 Time series of the EKE (black, cm 2 s -2 ) and surface Chl-a concentration (red, mgm -3 ) averaged over the box shown in Fig. 5 Eddy Kinetic Energy Eddy Kinetic Energy of Chl-a can be found further offshore between 12 S and 14 S (Fig. 7b). Investigation on the observed SSH anomaly (SSHA) demonstrates that the offshore Chl-a bloom is colocated with the presence of cyclonic eddies (Fig. 7c). The upwelling induced by cyclonic eddy (spinning clockwise in the Southern Hemisphere) has a large effect on the input of high nutrients into the surface layer leading to the offshore bloom of Chl-a. We also found that the anticyclonic eddies (spinning anticlockwise) are associated with low Chl-a concentration. Both the cyclonic and anticyclonic eddies are well reproduced by the model (Fig. 7d). The agreement is fairly good in terms of spatial pattern and magnitude. In addition, the model also shows that the cyclonic (anticyclonic) eddy is associated with high (low) Chl-a concentration. The mean EKE calculated from the observed SSHA for a period from May through December 2006 shows a maximum EKE centered at about 11.5 S, 111 E, and a secondary peak near the southern coast of Java around 10 S 113 E (Fig. 7e). The EKE spatial distribution derived from the model surface currents is relatively similar to the observed ones (Fig. 7f). Two regions of high EKE, near the southeastern coast Java and further offshore region along 13 S, are well reproduced by the model, even though the model values are relatively overestimated. In addition, the model also shows relatively high EKE in the south of central Java, where the observed data are not available for this region. To further examine the evolution and the mechanism of the offshore Chl-a bloom, Fig. 8a b show the time longitude diagram of the model surface Chl-a from May 1 to December 30, 2006, averaged over 9 S 10.5 S (hereafter section-a ) and 10.5 S 12.5 S (hereafter section-b ), respectively. High concentration of Chl-a is found between about 100 E 125 E, and it is clearly seen during July October along section-a and during July August along section-b. In particular, along section-a, a high concentration of Chl-a exceeding 0.6 mg m -3 persists throughout July October between about 105 E 120 E. It is also shown that the concentration of Chl-a along section-a is higher than that along section-b. The enhancement of surface Chl-a along both section-a and section-b is propagating westward with a phase speed of about 0.19 m/s. This characteristic of westward propagation is further supported by the time longitude diagrams of the SSHA along section-a and section-b showing similar westward propagating signals at the same phase speed (Fig. 8c d). The phase speed obtained from the time longitude diagram of the surface Chl-a is in a good agreement with those proposed by Feng and Wijffels (2002) and Yu and Potemra (2006). Closer examination of the SSHA shows that the amplitude of SSHA variability

7 Ocean Dynamics (2010) 60: Fig. 7 Observed (a) and model (b) surface Chl-a (mg m-3) distribution during September 19, c Observed SSHA (shaded, cm) and surface geostrophic currents (vector, m s-1) averaged during September 17 23, (d) Model SSHA (shaded, cm) and surface currents (vector, m s-1) during September 19, (e) Observed mean EKE (m2 s-2) estimated from geostrophic currents and (f) model mean EKE (m2 s-2) calculated from the surface currents. The boxes shown in the lower-left panel are the regions chosen for areaaveraged plots: Figs. 9 and a SeaWiFS Chl-a (19/09/2006) b Model Chl-a (19/09/2006) c Obs. SSHA (17-23/09/2006) d Model SSHA (19/09/2006) e Obs. mean EKE (May-Dec) f Model mean EKE (May-Dec) A B a Chla (9S-10.5S) b Chl-a (10.5S-12.5S) c SSHA (9S-10.5S) d SSHA (10.5S-12.5S) Fig. 8 Longitude-time diagrams of model (a b) surface chlorophyll (mg m-3) and (c d) SSHA (cm) averaged over 9 S 10.5 S and S, respectively

8 738 Ocean Dynamics (2010) 60: along section-b is larger than that along section-a, in contrast to that of the surface Chl-a concentration. To further investigate the processes that lead to the offshore Chl-a bloom, we evaluate the EKE, vertical velocity, Chl-a, nitrate, temperature, and mixed-layer depth (MLD) variations averaged over 9 S 10.5 S, 108 E 115 E (box A) and over 10.5 S 12.5 S, 107 E 114 E (box B) (marked by two rectangular boxes in Fig. 7e). Box B is designed to capture the highest eddy activity generated further south, whereas box A is selected to cover the secondary maximum of high eddy activity near the southeastern coast of Java. Figure 9a shows the evolution of surface Chl-a and EKE averaged over the box A. One can clearly see that there are three distinct events of enhancement of surface Chl-a (e.g. in July, September, and October November), which are Fig. 9 Time series of model (a) surface chlorophyll (black curve, mgm -3 ) and EKE (red curve, m 2 s -2 ), (b) vertical velocity at 30-m depth (black curve, m day -1 ) and EKE (red curve, m 2 s -2 ), during May December Time-depth sections for (c) chlorophyll (mg m -3 ), (d) nitrate (mmoln m -3 ), and (e) temperature ( C). Panel (f) same as in panel (d), except for the upper 20-m depth. The white-dashed curve in panels (b) (f) shows the MLD defined as the depth of 0.25 kg m -3 density deviation from the surface. All parameters are averaged over the box-a shown in Fig. 1e Chlorophyll-a Vert. velocity a b c E K E E K E d e f 2006

9 Ocean Dynamics (2010) 60: Fig. 10 As in Fig. 9 except for the parameters averaged over the Box-B shown in Fig. 7e Chlorophyll-a Vert. velocity a b E K E E K E c d e f 2006 associated with strong eddy activities. The EKE first begins to increase on June 26 and continues to do so until July 12. This increase is accompanied by the increase in surface Chl-a up to about 0.6 mgchl m -3 in early July. Then, the EKE begins to decrease on July 18, reaching its minimum value on August 11. The decrease is followed by a decline of surface Chl-a. The EKE, then, again begins to increase on August 17, reaching its maximum value on September 1. The rapid increase of the EKE is also accompanied by an increase in the surface Chl-a concentration. We also noted a moderate increase of the EKE on September 10. This relatively weaker EKE, however, is followed by a sharp increase of the surface Chl-a (maximum value of about 0.57 mgchl m -3 ).

10 740 Ocean Dynamics (2010) 60: Table 1 August October average of the nitrate fluxes into the upper 75 m in box-a and box-b a Region N t Zonal advection Meridional advection Vertical advection Residual Box-A Box-B a Fluxes are measured by mmol N m -2 day -1 over the top of 75 m and averaged at each region. Box-A is capturing region (9 S 10.5 S, 108 E 115 E), whereas box-b is capturing region (10.5 S 12.5 S, 107 E 114 E). The residual term includes the primary production and remineralization. The third occurrence of high EKE in this region begins on October 15, reaching its maximum value of about m 2 s -2 on November 3. In contrast to the previous occurrence of moderate EKE followed by the sharp increase of surface Chl-a, the rapid increase of EKE during November is not accompanied by a clear increase of the surface Chl-a, which is to be discussed next. To examine these blooms in detail, we investigate the evolution of vertical velocity at 30-m depth, subsurface Chl-a, nitrate, and temperature as well as MLD (Fig. 9b e). Apparently, it is shown that the cyclonic eddies are associated with upward vertical velocity (Fig. 9b). In addition, the vertical section of temperature reveal a doming of isotherms in the upper about 100 m (Fig. 9e), a typical structure of a cyclonic eddy in the Southern Hemisphere. There are three doming events (e.g. in July, September, and October November), and all are associated with strong eddy activities. The influence of these cyclonic eddies in pumping nutrient-rich water into the upper layer can be seen in the vertical section of the nutrient (Fig. 9d). The vertical distribution of the nutrient closely tracked the isotherms, and three events of elevated nutrient concentration are coincident with the doming events. The nitrate contour at 1 mmoln m -3 is elevated to about 30 m in July, September, and late October early November. Similarly, the response of Chl-a to the eddy pumping is seen in the vertical distribution of Chl-a (Fig. 9c). There is an increase of subsurface Chl-a during July, September, and October November. During July and September, the deep Chl-a maximum is located above the MLD, whereas during November, it is located below the MLD. This may explain why strong eddy activity in November is not followed by high Chl-a concentration at the surface. It also should be noted that the surface Chl-a concentrations remain relatively high during August ( 0.4 mgchl m -3 )and October ( 0.25 mgchl m -3 ). Considering the location of box A, which is close to the coast, the offshore transport of Chl-a concentration associated with an anomalously strong upwelling during this season (Vinayachandran et al. 2007; Iskandar et al. 2009) could be a possible candidate to explain these a b c d e f Fig. 11 Snapshot of the surface currents (vector, ms -1 ) superimposed on the vertically averaged nitrate concentration in the upper 75 m (mmoln m -3 ) during September October 2006

11 Ocean Dynamics (2010) 60: surface blooms. Strong upward vertical velocity during this period indicates the occurrence of a strong coastal upwelling associated with the IOD event (Fig. 9b). The highest EKE (greater than m 2 s -2 )inbox B occurs in October 1, and the secondary peak reaching m 2 s -2 isfoundinaugust14(fig.10a). EKE is also relatively high during November and December. However, only high EKE during August is followed by enhancements of surface Chl-a concentration up to about 0.3 mgchl m -3. To examine the influence of the eddies on the variability of Chl-a in box B, we evaluate the subsurface evolution of vertical velocity at 30-m depth, Chl-a, nitrate, temperature, and MLD (Fig. 10b e). The variation of vertical velocity and the vertical distributions of Chl-a and nitrate closely follow the isotherms. The shoaling of 24 C from the deeper layer to about 50 m in August, October, November, and December associated with the cyclonic eddy, which induces nutrient-rich water to the upper layer (Fig. 10a, d, and e). The nitrate concentration at the base of the 24 C isotherm increases from 0.2 mmoln m -3 to about 1.0 mmoln m -3 during the events. Moreover, the deep Chl-a maximum layer of concentration 0.4 mgchl m -3 is colocated with the doming of isotherm and nitrate. The increase of surface Chl-a occurrs when the MLD erodes the deep Chl-a maximum as in August September (Fig. 10a and c). The MLD is shallowed from about 50 m in September to about 30 m from October through December, and the deep Chl-a maximum is now located below the MLD (Fig. 10c). This may explain why there is no significant enhancement of surface Chl-a in October, although there is strong cyclonic eddy activity. This is also applied to the relatively strong eddy activity in November December, which is also not followed by the enhancement of surface Chl-a. To examine the offshore Chl-a bloom quantitatively, we have calculated vertical advective flux of nitrate in the upper 75 m at each region during peak bloom period from August through October (Table 1). It is shown that the vertical advection flux plays a dominant role in supporting the supply of nitrate in both box-a and box-b. The secondary contribution to the supply of nitrate in both regions comes from the zonal advection flux. This result further demonstrates the role of eddy in enhancing surface Chl-a bloom during this season. Figure 11 displays snapshot images of the model surface current superimposed on the vertically averaged nitrate in the upper 75 m during September October The cyclonic eddy formed in September near the Lombok Strait. Its core is associated with high nitrate concentration indicating an upward injection of nutrient-rich water (Fig. 11a). The eddy moves southwestward and then moves westward along 12 S (Fig. 11c f). By mid October, the center of the eddy is found around 12 S 110 E (Fig. 11f). We also found relatively weak cyclonic eddies south of the Nusa Tenggara Island chains. These eddies are also associated with relatively high nitrate concentration. The eddies move westward and merge with that formed in the south of Lombok Strait in early October (Fig. 11e). 4 Conclusion The effect of the cyclonic eddy on the surface Chl-a distribution during boreal summer fall 2006 in the SETIO is studied using a coupled physical biological model. During this period, offshore Chl-a bloom is observed and strong eddy activities prevail in this region. The results demonstrate that the cyclonic eddy markedly affects the enhancement of surface Chl-a in this region. The systematic vertical variations in Chl-a, nitrate, temperature, and MLD strongly imply that the cyclonic eddy could induce surface Chl-a bloom. Interestingly, the location of deep Chl-a maximum and the depth of mixed-layer are critical. The surface bloom occurs if the deep Chl-a maximum is located above the MLD. In contrast, the bloom is suppressed when the deep Chl-a maximum is located below the MLD. In addition, intense southeasterly winds during this period also generated favorable conditions for the formation of surface Chl-a bloom near the southeastern coast of Java. Finally, our results also reveal significantly larger subsurface Chl-a enhancement associated with the cyclonic eddy, even if it is hardly detectable at the surface layer. Evidence that nutrient pumping by the cyclonic eddy affects subsurface Chl-a abundance can be seen in the vertical distribution and colocation of high subsurface Chl-a concentration and the doming of isotherms and nitrate. Acknowledgment SeaWiFS chlorophyll data are courtesy of the NOAA CoastWatch Program, NASA's Goddard Space Flight Center. Iskhaq Iskandar would like to thank Prof. Toshio Yamagata for his encouragements and invaluable guidance during the author's early career and for his continued interest in the author's personal and professional growth. The authors would like to thank Dr. A. Ishida for his help in preparing the OFES output. The OFES simulations were conducted on the Earth Simulator under the support of JAMSTEC. Iskhaq Iskandar carried out part of the work while on the Japan Society for the Promotion of Science postdoctoral fellowship. References Asanuma I, Matsumoto K, Okano H, Kawano T, Hendiarti N, Sachoemar SI (2003) Spatial distribution of phytoplankton along the Sunda Islands: the monsoon anomaly in J Geophys Res 108:3202. doi: /1999jc Cipollini P, Cromwell D, Challenor PG, Raffaglio S (2001) Rossby waves detected in global ocean colour data. Geophys Res Lett 28 (2):

12 742 Ocean Dynamics (2010) 60: Feng M, Wijffels S (2002) Intraseasonal variability in the South Equatorial Current of the East Indian Ocean. J Phys Oceanogr 32: Fieux M, Andrie C, Delecluse P, Ilahude AG, Kartavtseff A, Mantisi F, Molcard R, Swallows JC (1994) Measurements within the Pacific Indian oceans throughflow region. Deep Sea Res I 41: Horii T, Hase H, Ueki I, Masumoto Y (2008) Oceanic precondition and evolution of the 2006 Indian Ocean Dipole. Geophys Res Lett 35:L doi: /2007gl Iskandar I, Tosuka T, Sasaki H, Masumoto Y, Yamagata T (2006) Intraseasonal variations of surface and subsurface currents off Java as simulated in a high-resolution ocean general circulation model. J Geophys Res 111:C doi: /2006jc Iskandar I, Rao SA, Tozuka T (2009) Chloropyhll-a bloom along the southern coasts of Java and Sumatra during Int Journal of Rem Sen 30: Masumoto Y, Sasaki H, Kagimoto T, Komori N, Ishida A, Sasai Y, Miyama T, Motoi T, Mitsudera H, Takahashi K, Sakuma H, Yamagata T (2004) A fifty-year eddy-resolving simulation of the world ocean preliminary outcomes of OFES (OGCM for the Earth Simulator). J Earth Simulator 1:35 56 McGillicuddy DJ Jr, Robinson AR (1997) Eddy-induced nutrient supply and new production in the Sargasso Sea. Deep Sea Res, Part I 44: Murtugudde R, Signorini SR, Christian JR, Busalacchi AJ, McClain CR, Picaut J (1999) Ocean color variability of the tropical Indo- Pacific basin observed by SeaWiFS during J Geophys Res 104: Murtugudde R, McCreary JP, Busalacchi AJ (2000) Oceanic processes associated with anomalous events in the Indian Ocean with relevance to J Geophys Res 105: Ogata T, & Masumoto Y (2009) Interactions between messoscale eddy variability and Indian Ocean Dipole events in the southeastern tropical Indian Ocean case study for 1994 and 1997/98. Ocean Dyn. (submitted) Oschlies A (2001) Model-derived estimates of new production: new results point towards lower value. Deep-Sea Res II: Oschlies A, Arcon V (1998) Eddy-induced enhancement of primary productivity in a model of the North Atlantic Ocean. Nature 934: Quadfasel DR, Cresswell G (1992) A note on the seasonal variability in the South Java Current. J Geophys Res 97: Saji NH, Goswami BN, Vinayachandran PN, Yamagata T (1999) A dipole mode in the tropical Indian Ocean. Nature 401: Sarmiento et al (1993) A seasonal three-dimensional ecosystem model of nitrogen cycling in the North Atlantic euphotic zone. Global Biogeochem Cycles 7: Sasai Y, Ishida A, Sasaki H, Kawahara S, Uehara H, Yamanaka Y (2006) A global eddy-resolving coupled physical biological model: physical influences on a marine ecosystem in the North Pacific. Simulation 87: Sasaki H, Sasai Y, Nonaka M, Kawahara S (2006) An eddy-resolving simulation of the quasi-global ocean driven by satellite-observed wind field: pleminary outcomes from physical and biological fields. J Earth Simulator 6:35 49 Susanto RD, Marra J (2005) Efect of the 1997/98 El Niño on chlorophyll a variability along the southern coast of Java and Sumatra. Oceanography 18: Susanto RD, Moore TS II, Marra J (2006) Ocean color variability in the Indonesia Seas during the SeaWiFS era. Geochem, Geophys, Geosys, 7, doi: /2005gc Vinayachandran PN, Kurian J, Neema CP (2007) Indian Ocean response to anomalous conditions in Geophys Res Lett 34: L doi: /2007gl Webster PJ, Moore AM, Loschnigg JP, Leben RR (1999) Coupled ocean atmosphere dynamics in the Indian Ocean during Nature 401: Wyrtki K (1961) Physical oceanography of the Southeast Asian waters: scientific results of marine investigations of the South China Sea and the Gulf of Thailand , NAGA Rep. 2, 195 pp., Scripps Inst. of Oceanogr., La Jolla, Calif Yu Z, Potemra J (2006) Generation mechanism for the intraseasonal variability in the Indo-Australian basin. J Geophys Res 111: C doi: /2005jc003023