PHILIppINE STRAITS DYNAMICS EXPERIMENT Multiscale Physical and Biological Dynamics in the Philippine Archipelago Predictions and Processes a 12°30’N b 15°N 28.5 12°00’N 28.2 27.9 14°N 27.6 27.3 11°30’N 27.0 13°N 26.7 26.4 28.5 26.1 11°00’N 12°N 28.2 25.8 27.9 120°00’E 120°30’E 121°00’E 121°30’E 122°00’E 27.6 11°N 27.3 c 29.7 20°N 27.0 29.0 26.7 28.3 10°N 26.4 27.6 26.1 25.8 26.9 26.2 118°E 119°E 120°E 121°E 122°E 123°E 124°E 25.5 15°N BY PIERRE F.J. LERMUSIAUX, 24.8 PATRICK J. HALEY JR., 24.0 WAYNE G. LESLIE, 23.4 ARPIT AgARWAL, OLEG G. LogUTov, AND LISA J. BURTon d 28 10°N -50 26 24 -100 22 -150 20 -1 18 40 cm s -200 16 5°N 14 -250 06Feb 11Feb 16Feb 21Feb 26Feb 03Mar 70 Oceanography | Vol.24, No.1 115°E 120°E 125°E 130°E ABSTRACT. The Philippine Archipelago is remarkable because of its complex of which are known to be among the geometry, with multiple islands and passages, and its multiscale dynamics, from the strongest in the world (e.g., Apel et al., large-scale open-ocean and atmospheric forcing, to the strong tides and internal 1985). The purpose of the present study waves in narrow straits and at steep shelfbreaks. We employ our multiresolution is to describe and reveal such regional modeling system to predict and study multiscale dynamics in the region, without ocean features as estimated by a multi- the use of any synoptic in situ data, so as to evaluate modeling capabilities when resolution, tidally driven ocean model only sparse remotely sensed sea surface height is available for assimilation. We focus for the February and March 2009 period, on the February to March 2009 period, compare our simulation results to ocean without any in situ data assimilation. observations, and utilize our simulations to quantify and discover oceanic features in Our ocean science focus is on biogeo- the region. The findings include: the physical drivers for the biogeochemical features; chemical fields and circulation features, the diverse circulation features in each sub-sea and their variations on multiple scales; transport balances for the Sulu Sea and the flow fields within the major straits and their variability; the transports to and from flow fields in the corresponding straits, the Sulu Sea and the corresponding balances; and finally, the multiscale mechanisms and, finally, formation mechanisms for involved in the formation of the deep Sulu Sea water. the deep Sulu Sea water. The goals of the Philippine Straits InTRODUCTION surface and subsurface water masses are Dynamics Experiment (PhilEx; Gordon, The Philippine Archipelago is a advected to the archipelago, where they 2009; Lermusiaux et al., 2009; Gordon fascinating multiscale ocean region. interact and every so often mix to form et al., 2011) were to enhance our Its geometry is very complex, with new water properties. Due to Earth’s understanding of physical and biogeo- multiple straits, islands, steep shelf- rotation, and the ocean’s stratification chemical processes and features arising breaks, and coastal features, leading and complex bathymetry, mesoscale in and around straits, and to improve to partially interconnected seas and features are created, often with spatially our capability to predict the spatial and basins (Figure 1). At depth, bathymetric inhomogeneous Rossby radii of defor- temporal variability of these regions. barriers form the boundaries of a mation. The surface atmospheric fluxes A specific objective of the modeling number of semi-enclosed seas. On the are also multiscale, including interannual research was to evaluate the capability of east, the western Pacific, including the variations, monsoon regimes, weather tuned modeling systems to estimate the North Equatorial Current, Kuroshio, events, and topographic wind jets (May circulation features and processes using and Mindanao Current dynamically et al., 2011; Pullen et al., 2011). Bottom only historical data sets for initializa- force these multiply connected domains. forcing also occurs, for example, in deep tion and only remotely sensed data for On the north-northwest, they are waters that are known to be affected by assimilation, for example, satellite sea forced by the South China Sea and its hydrothermal vents (e.g., Gamo et al., surface temperature (SST), height (SSH), coastal currents, eddies, and jets. The 2007). Finally, and as importantly, baro- and color (SSC). The applied motivation interactions of these forcings at lateral tropic tides, often out of phase in the of this approach is to simulate the very boundaries with complex geometry different basins (Logutov, 2008), strongly frequent operational situation where no drive abundant flow features with varied affect flows, especially in shallower synoptic in situ data can be collected, temporal and spatial scales (Broecker regions and straits. Due to the area’s and remotely sensed data are the only et al., 1986; Metzger and Hurlburt, 1996; variable stratification, rotation, and steep synoptic information available. The Gordon et al., 2011) and multiple feed- topographies, they drive a wealth of scientific motivation is to evaluate the backs to the lateral forcing seas. Several internal tides, waves, and solitons, some intrinsic capabilities of models, specifi- cally to determine if some dynamics Opposite page. MIT Multidisciplinary Simulation, Estimation, and Assimilation System (MSEAS) esti- can be simulated without using any mates of: (a-c) 25-m temperature at 0430Z on February 17, 2009 from three implicit two-way nested synoptic in situ data. simulations at 1-km, 3-km, and 9-km resolutions, and (d) a time series of temperature profiles at the Sulu Sea entrance to Sibutu Passage. Features are simulated at multiple scales, including the North For our PhilEx simulations, we Equatorial Current, mesoscale eddies, jets, filaments, and internal tides and waves. employ the MIT Multidisciplinary Oceanography | March 2011 71 a Figure 1. Spherical-grid Simulation, Estimation, and Assimilation domains in a telescoping System (MSEAS Group, 2010). It zoom configuration for our multiscale simulations in includes a free-surface hydrostatic the Philippine Archipelago primitive-equation physical ocean model overlaid on our estimate of bathymetry (in m, same developed for multiscale dynamics, color bar for all panels). resolving very shallow regions with Our bathymetry combines strong tides, steep bathymetries, and V12.1 (2009) of the Smith and Sandwell (1997) the deep ocean. The system is capable topography with hydro- of multiresolution simulations over graphic and bathymetric complex geometries with implicit ship data. (a) Archipelago 9-km-resolution domain schemes for telescoping nesting (Haley with nested Mindoro (3 km) and Lermusiaux, 2010) and has an and Mindanao (3 km) option for stochastic subgrid-scale domains. (b) Mindoro Strait 3-km and 1-km domains. representations (Lermusiaux, 2006). The (c) Mindanao/Surigao Strait physical model is coupled to multiple 3-km and 1-km domains. b Straits and local features biological models (Besiktepe et al., are identified by one or 2003; Tian et al., 2004) and acoustic two-letter abbreviations. models (Lam et al., 2009; Lermusiaux Alphabetically these are: B – Balabac Strait et al., 2010). The ocean physics is Cw – Cuyo West Passage forced with high-resolution barotropic D – Dipolog (Mindanao) tides, estimated using nested coastal Strait IB – Illigan Bay inversions (Logutov and Lermusiaux, M – Mindoro Strait 2008). Due to the complex multicon- P – Panay Sill nected sea domains, all ocean fields SB – San Bernardino Strait Si – Sibutu Passage are initialized here with new objective Su – Surigao Strait mapping schemes developed specifi- T – Tablas Strait V – Verde Island Passage cally for PhilEx, using fast marching Ta – Tapiantana Strait methods (Agarwal, 2009; Agarwal and Z – Zamboanga Strait Lermusiaux, 2010). SST is only used in the initial conditions. There is no c in situ data assimilation (Lermusiaux, 1999, 2002, 2007; Lermusiaux et al., 2000); the only synoptic data used are the sparse satellite SSH observations, providing weak corrections every four days to a week. During the two-month Intensive Observational Period of February to March 2009 (IOP-09), the MSEAS system was employed in real time, issuing daily physical-biological fore- casts. Dynamical descriptions and adaptive sampling guidance were also provided every three to four days 72 Oceanography | Vol.24, No.1 (Lermusiaux et al., 2009). Fields were focusing on their most novel compo- tidal forcing, parameterizations for compared to data sets from ships and nents: biogeochemical ocean predictions river input and subgrid-scale processes, gliders when available. with region-specific biological state feature models (Gangopadhyay et al., The present work is partly inspired initializations and multiresolution tidal 2003), and a suite of coupled biological by our experience in coastal regions predictions. We then describe a subset (NPZ) models. In the present IOP-09 with complex geometries (Haley and of our re-analysis modeling results and multiresolution simulations, we use the Lermusiaux, 2010), especially with steep compare them to ocean observations. two-way nesting scheme fully implicit shelfbreaks and straits such as the Sicily We report the major circulation features in space and time (i.e., such that within Strait (Lermusiaux, 1999; Lermusiaux estimated and discuss our estimates each time step, updated field values and Robinson, 2001), Massachusetts Bay of transports to and from the Sulu Sea are exchanged across scales and nested and Stellwagen Bank (Besiktepe et al., as well as the flow fields in the corre- domains as soon as they become avail- 2003), Middle Atlantic Bight shelfbreak sponding straits. Finally, we examine the able). This method differs from explicit (Lermusiaux, 1999), Monterey Bay shelf- multiscale formation mechanisms for the nesting schemes that exchange coarse break (Haley et al., 2009), and Taiwan deep Sulu Sea water.