Geophysical Research Letters

Marine ecosystem simulation in the eastern tropical Pacific with a global eddy resolving coupled physical-biological model

Authors


Abstract

[1] We have simulated the seasonal variability of marine biology in the eastern tropical Pacific using a global eddy-resolving coupled physical-biological model. Using high-resolution satellite wind fields, the model reproduces the seasonal variability of surface chlorophyll influenced by the meso-scale eddies and upwelling associated with the strong offshore wind jets. In winter, upwelling generated by the wind jets in the Gulfs of Tehuantepec, Papagayo, and Panama brings up cold and nitrate-rich waters from subsurface layer, where the tropical spring bloom occurs and is transported offshore. In summer, the intertropical convergence zone moves northward, and these jets weaken. The Costa Rica Dome develops with wind fields west of the Gulf of Papagayo. The dome in the open ocean supports high chlorophyll by the nutrient supply with upwelling. The westward expansion of surface chlorophyll of dome is response to the thermocline variation with the westward propagation of Rossby waves.

1. Introduction

[2] The eastern tropical Pacific Ocean is one of the most complex oceanic regimes in terms of the variety of physical processes. The region features numerous eddies generated as the northeasterly winds from the Gulf of Mexico and the Caribbean Sea are funneled through narrow passes in the mountains of Central America. Distinct eddy generation regions off Central America are the Gulf of Tehuantepec, the Gulf of Papagayo, and the Gulf of Panama. The eastern tropical Pacific also features the Costa Rica Dome, which is centered at 9°N, 90°W off the Gulf of Papagayo and is 300–500 km in diameter [Cromwell, 1958; Fiedler, 2002]. The Costa Rica Dome is an oceanic upwelling center where the thermocline ascends to very near the sea-surface. Such thermal domes are attributed to several dynamical mechanisms [e.g., McCreary et al., 1989; Umatani and Yamagata, 1991; Xie et al., 2005].

[3] The biological environment in the eastern tropical Pacific is characterized by physical features and atmospheric forcing, including a thermocline, the intertropical convergence zone (ITCZ), coastal wind jets, coastal and equatorial upwelling, the Costa Rica Dome, and meso-scale eddies. Ocean color satellite captures the chlorophyll pattern influenced by the meso-scale phenomena along the west coast of Central America, and in the Costa Rica Dome [e.g., Müller-Karger and Fuentes-Yaco, 2000; Fiedler, 2002]. Along the west coast of Central America, the strong wind jets over the Gulfs of Tehuantepec, Papagayo, and Panama during winter drive oceanic vertical mixing and Ekman pumping. Upwelling near the coast bring colder and nutrient-rich waters from lower layers to the surface, where the blooms occur and are transported offshore in large filaments created by the wind jets [e.g., Müller-Karger and Fuentes-Yaco, 2000; Willett et al., 2006]. The seasonal variability in chlorophyll is related to the thermocline depth [e.g., Müller-Karger and Fuentes-Yaco, 2000; Fiedler, 2002]. In spring and summer, a regional maximum in chlorophyll concentration is present in surface waters over the Costa Rica Dome. Open-ocean upwelling brings colder and nutrient-rich waters close to surface layer, where the high biological productivity is maintained [Fiedler, 2002].

[4] Our eddy-resolving (0.1°) ocean general circulation model represents realistically meso-scale phenomena, including separation of boundary current, and meso-scale eddy generation [Masumoto et al., 2004]. Sasaki and Nonaka [2006] demonstrated a local air-sea interaction associated with Hawaiian Lee Countercurrent using a same ocean model forced by QuikSCAT (QSCAT) wind field. Ocean circulation system in the eastern tropical Pacific is affected by the orographic wind field. Capet et al. [2004] demonstrated the upwelling sensitivity to coastal wind profiles using a regional ocean model for the California Coast. The structure of near-shore wind strongly influences physical and biological environments in coastal regions. In this study, we simulate the seasonal variability of the marine ecosystem using an eddy-resolving coupled physical-biological ocean model forced by QSCAT wind field. We investigate the response of the marine ecosystem to meso-scale eddies and upwelling generated by the strong offshore winds. We focus on the response of marine biology to the variability of thermocline depth.

2. Model Description

[5] We used an eddy-resolving coupled physical-biological ocean model. The physical model is an eddy-resolving ocean model for the Earth Simulator (OFES) [Masumoto et al., 2004]. The domain of this model covers from 75°S to 75°N. The horizontal grid spacing is 1/10° and the vertical levels are 54. The model is integrated for 50 years under the climatological mean forcing (NCEP/NCAR reanalysis data) from the observed temperature and salinity fields (WOA98) without motion. After integration of 50-year periods, the model is driven by the daily mean forcing of NCEP/NCAR reanalysis from 1950 to 1998 [Kalnay et al., 1996].

[6] The marine ecosystem model is a simple nitrogen-based Nitrate, Phytoplankton, Zooplankton, Detritus (NPZD) pelagic model [Oschlies, 2001]. The evolution of any biological tracer concentration is governed by an advective-diffusive equation [Sasai et al., 2006]. The initial nitrate field is taken from the WOA98. The initial phytoplankton and zooplankton concentrations are set to 0.14 mmol N m−3 and 0.014 mmol N m−3 at the surface, respectively, decreasing exponentially with a scale depth of 100 m [Sarmiento et al., 1993]. Detritus is initialized to 10−4 mmol N m−3 everywhere. To establish the stable pattern of biological system, the biological model coupled with the physical model is integrated over a 5-year period under climatological mean forcing. The biological fields at the end of the 5 years are used as the initial condition. The initial condition of physical fields is from the end of 1998. The coupled model is driven by the daily mean surface wind stress data of QSCAT satellite and atmospheric daily mean data (heat and salinity fluxes) of NCEP/NCAR reanalysis [Kalnay et al., 1996] from 1999 to 2004. The QSCAT data is constructed by weighted mean method with horizontal resolution of 1° [Kutsuwada, 1998] and is provided from J-OFURO data set [Kubota et al., 2002]. Results are presented for the 11th coupled year (atmospheric forcing year is 2004).

3. Results

[7] The seasonal pattern of physical and biological fields is shown in Figure 1. In March, distribution of satellite chlorophyll concentration is high along the coast of Central America, especially, in the Gulfs of Tehuantepec, Papagayo, and Panama under the strong wind jets. These strong wind jets produced as winter northerlies over the Gulf of Mexico, and the northeast trade winds from the Caribbean, are funneled through the narrow mountain gaps [Chelton et al., 2000; Chelton et al., 2004]. The simulated chlorophyll distribution has represented the same pattern of SeaWiFS, especially, high chlorophyll (>1.0 mg chl m−3) distribution off the three gulfs. Off the Gulfs of Tehuantepec and Papagayo, the high chlorophyll area extends from the coast to the open ocean. Off the Gulf of Panama, the high chlorophyll extends to the southwest along the coast. The region of low chlorophyll concentration (<0.2 mg chl m−3) between the Gulfs of Papagayo and Panama is also clearly reproduced. The coastal eddies generated by the wind jets entrain and trap high nutrients, and carry them offshore when they propagate westward. The wind field produces wind mixing beneath the jets, and Ekman pumping on either side of the jets [Willett et al., 2006]. In the model, cold (<25°C) and high nitrate (>20.0 mmol N m−3) waters are lifted due to upwelling and vertical mixing with the development of mixed layer, and the spring bloom occurs off the three gulfs. In September, SeaWiFS captures high chlorophyll concentrations in the open ocean off the Gulf of Papagayo and along the coast of Central America. High chlorophyll concentration in the open ocean is in the shallow thermocline. The wind field is reverse from northeasterly in March to southwesterly in September with the migration of the ITCZ. The thermocline shoaling via upwelling has a large effect on input of high nutrients into the surface layer. The model clearly represents the high chlorophyll distribution in the Costa Rica Dome. The thermocline depth in the Costa Rica Dome is very shallow (20 m) by upwelling. The cold (<25°C) and high nitrate (>30.0 mmol N m−3) waters lift from subsurface layers to surface layer and high chlorophyll is maintained at the surface layer. However, in the south of 5°N, the simulated chlorophyll is higher than the SeaWiFS. The model may not represent high nitrate-low chlorophyll condition in the equatorial region [e.g., Pennington et al., 2006] because a biological model does not include micronutrients and iron limitation. Over the Costa Rica Dome, the simulated chlorophyll is a slightly higher than the SeaWiFS because the upwelling effect is too strong in the model (Figure 1c).

Figure 1.

Distribution of monthly climatological mean surface chlorophyll concentration (mg m−3) from (a) SeaWiFS with QSCAT wind stress (vectors in N m−2), (b) model, (c) simulated sea surface temperature (°C) and horizontal velocity at 25 m depth (vectors in cm s−1), and (d) mean nitrate concentration (mmol N m−3) upper 75 m depth and thermocline depth (contours in m, 20°C isotherm depth). SeaWiFS data was obtained from NASA GSFC's Ocean Color Web Site (http://oceancolor.gsfc.nasa.gov). SeaWiFS product is from September 1997 to March 2007. Simulated chlorophyll concentrations in mg chl m−3 computed from nitrogen via a constant ratio of 1.59 g chlorophyll per mol nitrogen. The QSCAT wind and model data are from 2000 to 2004.

[8] The full seasonal cycle of SeaWiFS chlorophyll concentration, simulated surface chlorophyll concentration, and mean nitrate concentration upper 75 m depth with the thermocline depth (20°C isotherm) in 2004 along 10°N, a latitude that cuts through the Costa Rica Dome, is shown in Figure 2. The seasonal variability of simulated chlorophyll is same pattern as the SeaWiFS chlorophyll (Figures 2a and 2b). The chlorophyll concentration is high east of 90°W and is low west of 90°W during first half of the year. The high chlorophyll area appears in the center of 90°W during second half of the year. The simulated chlorophyll concentration, however, is lower than the SeaWiFS chlorophyll along the coast during June–January and is higher in the open ocean during November–January. The seasonal variability of simulated chlorophyll responds to the variations of nitrate concentration and thermocline depth (Figure 2c). The thermocline shoaling by upwelling lifts up the nutricline and the nutrient supply to the surface layer increases. The thermocline-shoaling region (<30 m) is consistent with the distribution of high nitrate concentration (>20.0 mmol N m−3). The seasonal variability of thermocline depth behaves differently east and west of 90°W. East of 90°W, the thermocline is shallow by wind-driven upwelling in winter. The near-shore thermocline deepens during July–December as the winds weaken. West of 90°W, the seasonal cycle of thermocline is an opposite phase. The thermocline response displays a clear westward propagation indicative of Rossby waves [Xie et al., 2005], and the biological fields in the model show the same pattern in response to the variation of thermocline depth. The westerly expansion of surface high chlorophyll in the dome in summer responds to the westward propagation of thermocline shoaling.

Figure 2.

Long-time sections along 10°N: (a) 8 day SeaWiFS chlorophyll (mg m−3) in 2004; (b) simulated surface chlorophyll (mg chl m−3); (c) mean nitrate concentration (mmol N m−3) upper 75 m depth and thermocline depth (m, 20°C isotherm depth) in 2004. The temporal resolution of model is 3 day snapshots.

[9] Nutrient supply by upwelling has a large effect on the temporal pattern of surface chlorophyll concentration. Figure 3 shows the monthly mean distribution of simulated chlorophyll concentration and nitrate concentration with the vertical velocity along 10°N in 2004. During February–April, high chlorophyll (>1.0 mg chl m−3) appears between 89°W and 85°W in both the SeaWiFS and model. The strong wind jet causes Ekman upwelling on the southeast side of the jet off the Gulf of Papagayo, and the upwelling brings cold and high nitrate waters to the surface (Figures 1, 2, and 3). The maximum chlorophyll in the model expands to 50 m depth between 89°W and 85°W in March (Figure 3). The strong upwelling near the coast brings the high nitrate water to the surface layer, and the high nitrate water is transported to open ocean. The maximum chlorophyll distribution is mainly controlled by the input of high nitrate water from the subsurface layer. During May–June, the Costa Rica Dome separates from the Gulf of Papagayo and shifts to 90°W [Fiedler, 2002]. The center of the dome is cold and high nitrate waters relative to surrounding waters. The thermocline is shallow (20 m) by the upwelling and lifts up cold and high nitrate waters. Surface chlorophyll is high with the input of nitrate-rich water. In June, maximum chlorophyll between 91°W and 87°W in the surface layer appears along 10°N in Figure 3. The nitrate-rich water (20.0 mmol N m−3) is supplied in the same location by the upwelling. During July–September, distribution of high chlorophyll extends to the west because the supply area of nitrate-rich water (between 93°W and 87°W) extends with the spread of shoaling thermocline (20 m) in the Costa Rica Dome (Figures 1, 2, and 3). The high chlorophyll in the Costa Rica Dome is maintained by the supply of high nitrate water (>20.0 mmol N m−3) from the subsurface layer. In September, the high chlorophyll area around 95°W in the model is also shown in Figure 3. The simulated high chlorophyll depends on the supply of high nitrate water with the upwelling. During November–January, the Costa Rica Dome deepens and decreases in size when ITCZ moves south and strong trade winds blow over the dome [Fiedler, 2002]. The supply of high nitrate water to the surface is reduced (thermocline deepens) and high chlorophyll area is smaller than that in September. However, the simulated chlorophyll in the open ocean is higher than the SeaWiFS chlorophyll during November–January.

Figure 3.

Vertical distribution of monthly mean (a) chlorophyll concentration (mg chl m−3) and (b) nitrate concentration (mmol N m−3) and vertical velocity (vectors in cm s−1) in 2004 along 10°N in the model. Vertical velocity is multiplied by 104.

[10] The simulated high chlorophyll concentration (>0.1 mg chl m−3) is upper 75 m depth in Figure 3. The supply of high nitrate water to the surface with the upwelling affects the surface chlorophyll concentration (Figure 3). The temporal pattern of surface chlorophyll concentration is consistent with the pattern of average 0–75 m nitrate concentration (Figure 2). We have calculated three month means of vertical advective flux of nitrate (mmol N m−2 day−1) into the upper 75 m for the biological production in two regions (Gulf of Papagayo and Costa Rica Dome) along 10°N (Table 1). Nitrate fluxes are integrated over the top of 75 m and are averaged at each region. At the Gulf of Papagayo, the supply of nitrate by the vertical advection flux (30.28 mmol N m−2 day−1) is dominant in January–March. The primary production (17.13 mmol N m−2 day−1) is supported primarily by the extra nitrate input the upper 75 m by vertical advection. In June, the Costa Rica Dome is released from the coast and shifts to 90°W. The open-sea upwelling is dominant in this region. From April to September, the nitrate flux into the upper 75 m is controlled by the vertical advection (30.73 and 33.17 mmol N m−2 day−1). The primary production (19.76 and 24.75 mmol N m−2 day−1) in the Costa Rica Dome is supported primarily by the vertical advective flux. When the vertical advection flux is negative (nitrate output) or weak, horizontal advection and diffusion fluxes supply nitrate into the upper 75 m for biological production.

Table 1. Simulated Three-Month Means of Nitrate Fluxes Into the Upper 75 m in Two Regions Along 10°N in 2004a
RegionNtV.Adv.PPRH.Adv. and D
  • a

    Fluxes measured by mmol N m−2 day−1. Nitrate fluxes are integrated over the top of 75 m and are averaged at each region. The Gulf of Papagayo region is between 87°W and 85°W. The Costa Rica Dome region is between 90°W and 89°W. Nt is time variation of nitrate, V.Adv. is vertical advection, PP is primary production, and R is remineralization. Horizontal advection and diffusion fluxes (H.Adv. and D) are the difference between Nt and the three listed fluxes. Positive value is nitrate input into the upper 75 m and negative value is nitrate output.

Jan.–Mar.
Gulf of Papagayo16.6130.28−17.131.112.35
Costa Rica Dome−3.23−16.66−13.811.0726.17
 
Apr.–Jun.
Gulf of Papagayo−12.221.01−19.791.954.61
Costa Rica Dome16.4530.73−19.761.803.68
 
Jul.–Sep.
Gulf of Papagayo−3.625.07−12.580.653.24
Costa Rica Dome1.8133.17−24.753.42−9.85
 
Oct.–Dec.
Gulf of Papagayo1.20−13.80−13.670.6826.90
Costa Rica Dome−14.13−20.45−22.012.9425.39

4. Conclusion

[11] The model reproduces the seasonal variability of phytoplankton bloom in the eastern tropical Pacific, especially, the Gulfs of Tehuantepec, Papagayo, and Panama and the Costa Rica Dome. During January–March, Ekman upwelling generated by the strong offshore wind jets brings up the nitrate-rich water, where distinct surface blooms occur in the three gulfs. High chlorophyll areas in the Gulfs of Tehuantepec and Papagayo in the model extend as far as 400–600 km offshore. Off the Gulf of Papagayo, the supply of nitrate-rich water by the coastal upwelling is dominant. During June–September, high chlorophyll in the Costa Rica Dome is supported by the nutrient supply from subsurface layer by the open-ocean upwelling. High chlorophyll area of the dome extends to west from June to September because the nutrient supply area is response to the westerly expansion of thermocline shoaling with the westward propagation of Rossby waves.

[12] Using a satellite observation wind data, the model reproduces the seasonal variability of marine biology in the complicated physical environment (e.g., eastern tropical Pacific). Coarse-resolution atmospheric models are unable to resolve orographic wind fields induced by the mountains of island and continents (e.g., Central America). For accurate simulation with global eddy-resolving coupled physical-biological model, it is necessary to use high-resolution wind field such as from satellite observation (QSCAT).

Acknowledgments

[13] We thank Y. Masumoto, T. Kagimoto, and S. Kawahara for their support. The QSCAT product of J-OFURO was obtained from K. Kutsuwada. Two anonymous reviewers made helpful comments on the manuscript. OFES simulations were conducted on the Earth Simulator under support of JAMSTEC. This work was partly supported by CREST, JST.

Ancillary