• Allen, S. E., , and X. Durrieu de Madron, 2009: A review of the role of submarine canyons in deep-ocean exchange with the shelf. Ocean Sci., 5 , 607620.

    • Search Google Scholar
    • Export Citation
  • Banas, N. S., , P. MacCready, , and B. M. Hickey, 2009: The Columbia River plume as cross-shelf exporter and along-coast barrier. Cont. Shelf Res., 29 , 292301.

    • Search Google Scholar
    • Export Citation
  • Baptista, A. M., and Coauthors, 2005: A cross-scale model for 3D baroclinic circulation in estuary-plume-shelf systems: II. Application to the Columbia River. Cont. Shelf Res., 25 , 935972.

    • Search Google Scholar
    • Export Citation
  • Barron, C. N., , A. B. Kara, , P. J. Martin, , R. C. Rhodes, , and L. F. Smedstad, 2006: Formulation, implementation and examination of vertical coordinate choices in the Global Navy Coastal Ocean Model (NCOM). Ocean Modell., 11 , 347375.

    • Search Google Scholar
    • Export Citation
  • Barth, J. A., 2003: Anomalous southward advection during 2002 in the northern California Current: Evidence from Lagrangian surface drifters. Geophys. Res. Lett., 30 , 8024. doi:10.1029/2003GL017511.

    • Search Google Scholar
    • Export Citation
  • Barth, J. A., , S. D. Pierce, , and R. L. Smith, 2000: A separating coastal upwelling jet at Cape Blanco, Oregon and its connection to the California Current System. Deep-Sea Res. II, 47 , 783810.

    • Search Google Scholar
    • Export Citation
  • Barth, J. A., and Coauthors, 2007: Delayed upwelling alters nearshore coastal ocean ecosystems in the northern California Current. Proc. Natl. Acad. Sci. USA, 104 , 37193724.

    • Search Google Scholar
    • Export Citation
  • Beckmann, A., , and D. B. Haidvogel, 1993: Numerical simulation of flow around a tall isolated seamount. Part I: Problem formulation and model accuracy. J. Phys. Oceanogr., 23 , 17361753.

    • Search Google Scholar
    • Export Citation
  • Bosley, K. L., , J. W. Lavelle, , R. D. Brodeur, , W. W. Wakefield, , R. L. Emmett, , E. T. Baker, , and K. M. Rehmke, 2004: Biological and physical processes in and around Astoria submarine Canyon, Oregon, USA. J. Mar. Syst., 50 , 2137.

    • Search Google Scholar
    • Export Citation
  • Burk, S. D., , T. Haack, , and R. M. Samelson, 1999: Mesoscale simulation of supercritical, subcritical, and transcritical flow along coastal topography. J. Atmos. Sci., 56 , 27802795.

    • Search Google Scholar
    • Export Citation
  • Chan, F., , J. A. Barth, , J. Lubchenco, , A. Kirincich, , H. Weeks, , W. T. Peterson, , and B. A. Menge, 2008: Emergence of anoxia in the California Current large marine ecosystem. Science, 319 , 920.

    • Search Google Scholar
    • Export Citation
  • Chhak, K., , and E. DiLorenzo, 2007: Decadal variations in the California Current upwelling cells. Geophys. Res. Lett., 34 , L14604. doi:10.1029/2007GL030203.

    • Search Google Scholar
    • Export Citation
  • Crawford, W. R., , and R. K. Dewey, 1989: Turbulence and mixing: Sources of nutrients on the Vancouver Island continental shelf. Atmos.–Ocean, 27 , 428442.

    • Search Google Scholar
    • Export Citation
  • Freeland, H. J., , G. Gatien, , A. Huyer, , and R. L. Smith, 2003: Cold halocline in the northern California Current: An invasion of subarctic water. Geophys. Res. Lett., 30 , 1141. doi:10.1029/2002GLO16663.

    • Search Google Scholar
    • Export Citation
  • Garfield, N., , C. A. Collins, , R. G. Paquette, , and E. Carter, 1999: Lagrangian exploration of the California Undercurrent, 1992–95. J. Phys. Oceanogr., 29 , 560583.

    • Search Google Scholar
    • Export Citation
  • Gay, P. S., , and T. K. Chereskin, 2009: Mean structure and seasonal variability of the poleward undercurrent off southern California. J. Geophys. Res., 114 , C02007. doi:10.1029/2008JC004886.

    • Search Google Scholar
    • Export Citation
  • Gould, J., and Coauthors, 2004: Argo profiling floats bring new era of in situ ocean observations. Eos, Trans. Amer. Geophys. Union, 85 (179) 190191.

    • Search Google Scholar
    • Export Citation
  • Grantham, B. A., , F. Chan, , K. J. Nielsen, , D. S. Fox, , J. A. Barth, , A. Huyer, , J. Lubchenco, , and B. A. Menge, 2004: Upwelling-driven nearshore hypoxia signals ecosystem and oceanographic changes in the northeast Pacific. Nature, 429 , 749754.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., 1997: The response of a steep-sided, narrow canyon to time-variable wind forcing. J. Phys. Oceanogr., 27 , 697726.

  • Hickey, B. M., , and N. S. Banas, 2008: Why is the northern end of the California Current system so productive? Oceanography, 21 , 90107.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., , S. Geier, , N. Kachel, , and A. MacFadyen, 2005: A bi-directional river plume: The Columbia in summer. Cont. Shelf Res., 25 , 16311656.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., , A. MacFadyen, , W. Cochlan, , R. Kudela, , K. Bruland, , and C. Trick, 2006: Evolution of chemical, biological, and physical water properties in the northern California Current in 2005: Remote or local wind forcing? Geophys. Res. Lett., 33 , L22S02. doi:10.1029/2006GL026782.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., , R. McCabe, , S. Geier, , E. Dever, , and N. Kachel, 2009: Three interacting freshwater plumes in the northern California Current system. J. Geophys. Res., 114 , C00B03. doi:10.1029/2008JC004907.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., and Coauthors, 2010: River influences on shelf ecosystems: Introduction and synthesis. J. Geophys. Res., 115 , C00B17. doi:10.1029/2009JC005452.

    • Search Google Scholar
    • Export Citation
  • Huyer, A., 1983: Coastal upwelling in the California Current system. Prog. Oceanogr., 12 , 259284.

  • Huyer, A., , J. H. Fleischbein, , J. Keister, , P. M. Kosro, , N. Perlin, , R. L. Smith, , and P. A. Wheeler, 2005: Two coastal upwelling domains in the northern California Current system. J. Mar. Res., 63 , 901929.

    • Search Google Scholar
    • Export Citation
  • Huyer, A., , P. A. Wheeler, , P. T. Strub, , R. L. Smith, , R. Letelier, , and P. M. Kosro, 2007: The Newport Line off Oregon—Studies in the north east Pacific. Prog. Oceanogr., 75 , 126160. doi:10.1016/j.pocean.2007.08.003.

    • Search Google Scholar
    • Export Citation
  • IFREMER, 2002: QuikSCAT scatterometer mean wind field products user manual. IFREMER C2-MUT-W-04-IF, 47 pp.

  • Kosro, P. M., 2003: Enhanced southward flow over the Oregon shelf in 2002: A conduit for subarctic water. Geophys. Res. Lett., 30 , 8023. doi:10.1029/2003GL017436.

    • Search Google Scholar
    • Export Citation
  • Kosro, P. M., , W. T. Peterson, , B. M. Hickey, , R. K. Shearman, , and S. D. Pierce, 2006: Physical versus biological spring transition: 2005. Geophys. Res. Lett., 33 , L22S03. doi:10.1029/2006GL027072.

    • Search Google Scholar
    • Export Citation
  • Kudela, R. M., , W. P. Cochlan, , T. D. Peterson, , and C. G. Trick, 2006: Impacts on phytoplankton biomass and productivity in the Pacific Northwest during the warm ocean conditions of 2005. Geophys. Res. Lett., 33 , L22S06. doi:10.1029/2006GL026772.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., , P. MacCready, , and B. M. Hickey, 2009a: Columbia River plume patterns in summer 2004 as revealed by a hindcast coastal ocean circulation model. Geophys. Res. Lett., 36 , L02601. doi:10.1029/2008GL036447.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., , P. MacCready, , B. M. Hickey, , E. P. Dever, , P. M. Kosro, , and N. S. Banas, 2009b: Evaluation of a coastal ocean circulation model for the Columbia River plume in summer 2004. J. Geophys. Res., 114 , C00B04. doi:10.1029/2008JC004929.

    • Search Google Scholar
    • Export Citation
  • MacCready, P., , N. S. Banas, , B. M. Hickey, , E. P. Dever, , and Y. Liu, 2009: A model study of tide- and wind-induced mixing in the Columbia River estuary and plume. Cont. Shelf Res., 29 , 278291.

    • Search Google Scholar
    • Export Citation
  • MacFadyen, A., , B. M. Hickey, , and W. P. Cochlan, 2008: Influences of the Juan de Fuca Eddy on circulation, nutrients, and phytoplankton production in the northern California Current system. J. Geophys. Res., 113 , C08008. doi:10.1029/2007JC004412.

    • Search Google Scholar
    • Export Citation
  • Mackas, D. L., , K. L. Denman, , and A. F. Bennett, 1987: Least squares multiple tracer analysis of water mass composition. J. Geophys. Res., 92 , 29072918.

    • Search Google Scholar
    • Export Citation
  • Mackas, D. L., , W. T. Peterson, , M. D. Ohman, , and B. E. Lavaniegos, 2006: Zooplankton anomalies in the California Current system before and during the warm ocean conditions of 2005. Geophys. Res. Lett., 33 , L22S07. doi:10.1029/2006GL027930.

    • Search Google Scholar
    • Export Citation
  • Mellor, G. L., , and T. Yamada, 1982: Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys. Space Phys., 20 , 851875.

    • Search Google Scholar
    • Export Citation
  • Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87 , 343360.

  • Penven, P., , P. Marchesiello, , L. Debreu, , and J. Lefèvre, 2008: Software tools for pre- and post-processing of oceanic regional simulations. Environ. Model. Softw., 22 , 117122.

    • Search Google Scholar
    • Export Citation
  • Perlin, N., , E. D. Skyllingstad, , R. M. Samelson, , and P. L. Barbour, 2007: Numerical simulations of air–sea coupling during coastal upwelling. J. Phys. Oceanogr., 37 , 20812093.

    • Search Google Scholar
    • Export Citation
  • Pierce, S. D., , R. L. Smith, , P. M. Kosro, , J. A. Barth, , and C. D. Wilson, 2000: Continuity of the poleward undercurrent along the eastern boundary of the mid-latitude North Pacific. Deep-Sea Res. II, 47 , 811829.

    • Search Google Scholar
    • Export Citation
  • Pierce, S. D., , J. A. Barth, , R. E. Thomas, , and G. W. Fleischer, 2006: Anomalous warm July 2005 in the northern California Current: Historical context and the significance of cumulative wind stress. Geophys. Res. Lett., 33 , L22S04. doi:10.1029/2006GL027149.

    • Search Google Scholar
    • Export Citation
  • Schwing, F. B., , N. A. Bond, , S. J. Bograd, , T. Mitchell, , M. A. Alexander, , and N. Mantua, 2006: Delayed coastal upwelling along the U.S. west coast in 2005: A historical perspective. Geophys. Res. Lett., 33 , L22S01. doi:10.1029/2006GL026911.

    • Search Google Scholar
    • Export Citation
  • Shchepetkin, A. F., , and J. C. McWilliams, 2003: A method for computing horizontal pressure-gradient force in an oceanic model with a nonaligned vertical coordinate. J. Geophys. Res., 108 , 3090. doi:10.1029/2001JC001047.

    • Search Google Scholar
    • Export Citation
  • Shchepetkin, A. F., , and J. C. McWilliams, 2005: The Regional Ocean Modeling System: A split-explicit, free-surface, topography following coordinates ocean model. Ocean Modell., 9 , 347404.

    • Search Google Scholar
    • Export Citation
  • Smith, S. D., 1988: Coefficients for sea surface wind stress, heat flux, and wind profiles as a function of wind speed and temperature. J. Geophys. Res., 93 , 1546715472.

    • Search Google Scholar
    • Export Citation
  • Smith, W. H. F., , and D. T. Sandwell, 1997: Global sea floor topography from satellite altimetry and ship depth soundings. Science, 277 , 19561962.

    • Search Google Scholar
    • Export Citation
  • Springer, S. R., , R. M. Samelson, , J. S. Allen, , G. D. Egbert, , A. L. Kurapov, , R. N. Miller, , and J. C. Kindle, 2009: A nested model of the Oregon Coastal Transition Zone: Simulations and comparisons with observations during the 2001 upwelling season. J. Geophys. Res., 114 , C02010. doi:10.1029/2008JC004863.

    • Search Google Scholar
    • Export Citation
  • Stegmann, P. M., , and F. Schwing, 2007: Demographics of mesoscale eddies in the California Current. Geophys. Res. Lett., 34 , L14602. doi:10.1029/2007GL029504.

    • Search Google Scholar
    • Export Citation
  • Strub, P. T., , and C. James, 2003: Altimeter estimates of anomalous transports into the northern California Current during 2000–2002. Geophys. Res. Lett., 30 , 8025. doi:10.1029/2003GL017513.

    • Search Google Scholar
    • Export Citation
  • Strub, P. T., , J. S. Allen, , A. Huyer, , R. L. Smith, , and R. C. Beardsley, 1987: Seasonal cycles of currents, temperatures, winds, and sea level over the northeast Pacific continental shelf: 35°N to 48°N. J. Geophys. Res., 92 , 15071526.

    • Search Google Scholar
    • Export Citation
  • Thomas, A., , and P. Brickley, 2006: Satellite measurements of chlorophyll distribution during spring 2005 in the California Current. Geophys. Res. Lett., 33 , L22S05. doi:10.1029/2006GL026588.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Model domain and bathymetry (m). Locations of the Columbia River mouth (CR), South Beach (SB), Cape Blanco (CB), Crescent City Line (CresCy), NH Line, and NH10 mooring are indicated, along with the positions of the RISE moorings (Rc, Rs) and the NDBC buoy (gray triangle). The six analysis subdomains are also shown: the “upwelling region” (shaded, inshore of the 200-m isobath), 3 adjacent regions (NW, W, and SW), and 2 offshore regions (ON, OS). The plot is scaled so that the zonal and meridional distances are equal.

  • View in gallery

    (a) Cumulative meridional wind stress vs time (months) for a year beginning on 1 January and averaged over the upwelling region in Fig. 1. Columbia River discharge (b) volume flux and (c) temperature. Values are shown for the year 2005 (black lines) and for the climatology (gray).

  • View in gallery

    Observed ADCP (obs) and model (mod) velocity vectors (cm s−1, eastward along positive time axis) at (a) 10 and (b) 68 m at the NH10 location (Fig. 1) vs time (months) for the year 2005. The scale for the model velocities, relative to the displaced 0 line, is the same at each depth as that for the corresponding observed velocities.

  • View in gallery

    (a) SST at the position of NDBC buoy 46050 (see Fig. 1), and (b) adjusted sea level anomaly (relative to a mean annual cycle, represented by a single fitted annual harmonic) at South Beach (see Fig. 1) vs time (months) during 2005: observed (gray lines) and modeled values (black).

  • View in gallery

    Temporal means (May–October 2005) of (a),(c) SST (°C) and (b),(d) sea level anomaly (SLA, cm) from the (a),(b) model and (c),(d) observations. The model SLA shown in (b) is the difference from the annual mean.

  • View in gallery

    Observed (thin lines, shorter time series) and model (thick lines, longer time series) for (top to bottom) temperature T, salinity S, and velocity (u, υ) at 5- and 20-m depths at the positions of moorings (a) Rc and (b) Rs (Fig. 1) vs time (months) for May–October 2005.

  • View in gallery

    (a) Ekman volume transport (Sv, 1 Sv ≡ 106 m3 s−1) into the upwelling region through the western boundary (see Fig. 1) vs time (months) for the year 2005 (black line) and climatological (gray) simulations. Net volume transport (Sv) vs time (months) for the (b) year 2005 and (c) climatological simulations inward through the western (W, gray line), outward through the northern plus southern [−(N + S), thick black], and outward through the northern [(−N), thin black] boundary of the upwelling region shown in Fig. 1.

  • View in gallery

    Model mean mid-July–September (top) near-surface and (bottom) near-bottom velocities (cm s−1; arrows proportional to magnitude) in and near the upwelling region (Fig. 1) for the (a) year 2005 and (b) climatological simulations. At each point, these near-surface and near-bottom velocities are vertical averages over the uppermost 80 m and the lowermost 50 m, respectively.

  • View in gallery

    Model mean mid-July–September (left) meridional and (right) zonal velocity (cm s−1; 0 contour shown in black) vs longitude (°W) and depth (m) at the NH Line (Fig. 1) for the (a),(b) year 2005 and (c),(d) climatological simulations. The location near 124.7°W and 250-m depth of the model poleward undercurrent core, represented by the cross-sectional maximum northward velocity with the indicated mean and temporal standard deviation, is shown in (a) and (c). The 200-m isobath (gray line) defines the western boundary of the upwelling region (Fig. 1). Markers within this region indicate physical final particle positions for the NH Line trajectory analysis described in section 3b, with the symbols indicating the corresponding source regions. In (b) and (d), markers for source regions are plotted at final positions only for source regions supplying at least 10% of the corresponding particles at that position.

  • View in gallery

    The 6-month trajectories from March through 29 Sep 2005 vs latitude (°N) and longitude (°W), for 201 particles with final positions on 29 Sep 2005 evenly distributed at 2-m depth across the upwelling region (Fig. 1). Trajectory color indicates (a) the final latitude (°N) for the corresponding particle and (b) the instantaneous particle depth (m) along the trajectory. The region shown covers the NW, W, SW, and upwelling subdomains defined in Fig. 1.

  • View in gallery

    Mean paths for the (a),(b) year 2005 and (c),(d) climatological simulations, computed as described in the text, vs latitude (°N) and longitude (°W): (left) north and (right) south of Cape Blanco (CB). The means were computed from the 6231 particle paths discussed in section 3b, for which final positions were specified in August. Color indicates along-path depth (m). Mean paths are shown separately for particles upwelling north and south of CB. The subdomains defined in Fig. 1—according to which the trajectories are grouped for this analysis—are also shown along with the 200- and 1000-m isobaths. Numbers adjacent to the paths indicate the percentage of the particles represented by the corresponding mean path; only those paths originating outside the upwelling region and representing at least 1% of the total number of particles are shown so the percentages shown in (a),(b) and (c),(d) do not sum to 100.

  • View in gallery

    (a)–(d) Trajectory statistics vs final position latitude (°N) and longitude (°W) for the 62% of particles with specified final positions in the upwelling region in August that originated outside of the model domain in the year 2005 simulation. Shown at each corresponding final position are (a) the percentage of particles at that location that originated outside of the model’s domain, (b) the boundary from which these particles originated (northern or southern; black if both), and the mean (c) depth (m) and (d) time (days prior to date of specified final position) at which the particles entered from the corresponding boundary. (e)–(h) As in (a)–(d), but for the climatological simulation.

  • View in gallery

    (top) Trajectory statistics vs initial position (in February, 6 months prior to the corresponding final time in August) vs latitude (°N) and longitude (°W), for the 38% of particles with specified final positions in the upwelling region in August that originated offshore inside of the model domain in the year 2005 simulation. Shown at each corresponding initial position are the final latitude (°N, left) and initial depth (m, right) of particles at that initial location. (bottom) As in the top panels, but for the climatological simulation.

  • View in gallery

    Histograms (%) of particle initial positions and depths, for particles with final positions (top) north and (next row) south of Cape Blanco, vs the analysis subdomain (Fig. 1), for the 38% of particles with specified final positions in August and initial positions offshore in the year 2005 simulation (as in Fig. 13). Depth bins are 50 m, and the inset numbers (right panels) are the total number of particles represented in each corresponding pair of histograms. (bottom four panels) As in the top panels, but for the climatological simulation.

  • View in gallery

    (top) Upwelling displacement quantities for the 38% of particles with specified final positions in August and initial positions offshore in the year 2005 simulation (as in Fig. 13). For particles (left to right): number of paths and the corresponding mean rate of vertical motion for those that cross or are found at depths of 150 ± 5 m depth vs latitude (°N) and longitude (°W); numbers for approximately 14 km × 14 km cells that cross the indicated isobaths; and the numbers of paths that cross the indicated isobaths vs the final latitude (°N) of those crossing the 1000-m isobath at the given latitude and their corresponding depths (m). (bottom) As in the top panels, but for the climatological simulation.

  • View in gallery

    Along-path mean (top to bottom) temperature (θp), salinity (Sp), and depth (zp) deviation vs time tp, where tp = 0 is the time at which the particle’s final position in the upwelling region is specified for all of the particles with specified final positions in August in the (a) year 2005 and (b) climatological simulations. Along-path means were computed separately for particles entering from the northern (blue) and southern (red) boundaries and for those originating within the model domain (black). The along-path means for particles originating outside the domain were computed for periods during which at least 50% of the corresponding sets of particles had entered the domain. The temperature, salinity, and depth deviations are defined relative to the initial path mean at the time step prior to the first plotted time. The numbers in the top panels indicate the number and percentage of particles in each of the three groups. Error bars correspond to the standard deviations computed every 6 days for each group, relative to the corresponding ensemble mean.

  • View in gallery

    Along-path means computed similarly to the quantities in Fig. 16, but for diagnosed along-path values of the (top) heat and (bottom) salinity diffusion terms (κtTz)z (°C s−1) and (κsSz)z (psu s−1) where T and S are temperature and salinity, κt and κs are heat and salt diffusivities, and subscripts indicate derivatives with respect to the vertical coordinate z.

  • View in gallery

    Monthly mean TS diagrams for the NH (gray dots) and Crescent City (black) Lines. For (top) year 2005 and (bottom) climatological simulations, during the period June–September and inshore of the 200-m isobath. Note the change in axis scales between the July and August diagrams. Contours of the potential density anomaly are also shown.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 4 4 2
PDF Downloads 3 3 2

A Numerical Modeling Study of the Upwelling Source Waters along the Oregon Coast during 2005

View More View Less
  • 1 College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon
© Get Permissions
Full access

Abstract

Regional ocean circulation along the Oregon coast is studied numerically for forcing fields derived from year 2005 and climatological-mean conditions. The primary object is to study directly the Lagrangian pathways by which fluid arrives in the Oregon upwelling zone. Roughly half of the upwelling fluid is found to arrive in the regional domain from alongshelf source points, primarily north but also south of the Oregon upwelling zone, while the other half ultimately arrives from points offshore and west of the zone. For both the year 2005 and the climatological simulations, different regimes of dominant alongshelf source waters are found, with Cape Blanco being a dividing point for northern versus southern sources, and with the water parcels coming primarily from depths below 100 m. For the offshore sources, most upwelling fluid originates from depths between 150 and 250 m during 2005, but from within the upper 150 m for the climatological simulations. In both cases there are specific regions along the shelf of enhanced vertical motion, which appear to be associated with topographic features such as submarine banks and canyons. A perhaps surprising result is the apparently small role played by the poleward undercurrent as a direct, immediate source of upwelling fluid; on the seasonal time scale considered here, most trajectories are found to move southward over the slope and shelf, with weaker northward motion only farther offshore, in the deep interior. In 2005 the water parcels cover distances longer than in the climatology, and the meridional exchange along their paths is more vigorous. These differences are likely associated in part with the presence of mesoscale vortical structures seaward of the shelf, which are largely absent in the climatological simulation, perhaps because their formation may depend on the large-amplitude, short-time-scale wind variations and reversals that occur in the year 2005 wind field but not in the multiyear mean climatological wind field.

* Current affiliation: Ciencias Oceánicas y Atmosféricas Aplicadas, CICATA-IPN, Altamira, Tamaulipas, Mexico

Corresponding author address: David Rivas, CICATA-IPN, Km. 14.5 Carr. Tampico - Puerto de Altamira, 89600 Altamira, Tamaulipas, Mexico. Email: drivasc@ipn.mx

Abstract

Regional ocean circulation along the Oregon coast is studied numerically for forcing fields derived from year 2005 and climatological-mean conditions. The primary object is to study directly the Lagrangian pathways by which fluid arrives in the Oregon upwelling zone. Roughly half of the upwelling fluid is found to arrive in the regional domain from alongshelf source points, primarily north but also south of the Oregon upwelling zone, while the other half ultimately arrives from points offshore and west of the zone. For both the year 2005 and the climatological simulations, different regimes of dominant alongshelf source waters are found, with Cape Blanco being a dividing point for northern versus southern sources, and with the water parcels coming primarily from depths below 100 m. For the offshore sources, most upwelling fluid originates from depths between 150 and 250 m during 2005, but from within the upper 150 m for the climatological simulations. In both cases there are specific regions along the shelf of enhanced vertical motion, which appear to be associated with topographic features such as submarine banks and canyons. A perhaps surprising result is the apparently small role played by the poleward undercurrent as a direct, immediate source of upwelling fluid; on the seasonal time scale considered here, most trajectories are found to move southward over the slope and shelf, with weaker northward motion only farther offshore, in the deep interior. In 2005 the water parcels cover distances longer than in the climatology, and the meridional exchange along their paths is more vigorous. These differences are likely associated in part with the presence of mesoscale vortical structures seaward of the shelf, which are largely absent in the climatological simulation, perhaps because their formation may depend on the large-amplitude, short-time-scale wind variations and reversals that occur in the year 2005 wind field but not in the multiyear mean climatological wind field.

* Current affiliation: Ciencias Oceánicas y Atmosféricas Aplicadas, CICATA-IPN, Altamira, Tamaulipas, Mexico

Corresponding author address: David Rivas, CICATA-IPN, Km. 14.5 Carr. Tampico - Puerto de Altamira, 89600 Altamira, Tamaulipas, Mexico. Email: drivasc@ipn.mx

1. Introduction

Wind-driven upwelling is one of the most important processes affecting the coastal circulation off Oregon, and a critical factor in regional biological productivity and ecosystem structure. The upwelling season begins with the “spring transition,” when northward winds switch to southward, upwelling-favorable winds, and extends through the summer, typically until late fall (Huyer 1983). The timing of the spring transition and the strength of the upwelling winds both show significant interannual variability (Schwing et al. 2006). The purpose of this paper is to study numerically the upwelling circulation off Oregon during the year 2005, with particular focus on identifying the source of the upwelling waters, and to compare the results from year 2005 with those obtained for climatological forcing.

The year 2005 is of particular interest because the timing of the spring transition in 2005 was anomalous, with the upwelling season starting about 5 weeks later than normal (Pierce et al. 2006; Schwing et al. 2006; Barth et al. 2007). Indeed, the 2005 upwelling season in the northern California Current had one of the most delayed onsets in the last 40 yr (Schwing et al. 2006). This implied a delayed provision of inorganic nutrients to the upper waters of the coastal ocean, resulting in large, temporary decreases in biomass and productivity (Kudela et al. 2006; Thomas and Brickley 2006; Hickey et al. 2006; Mackas et al. 2006). The response to climatological forcing is examined here for comparison with the anomalous year 2005 forcing. Significant differences between these two cases are found in the paths of the upwelling waters and in other aspects of the circulation.

The rest of the paper is organized as follows. Section 2 describes the model setup and its forcing and boundary data, including brief comparisons between model outputs and observations, and a brief discussion of previous, related numerical modeling results. In section 3 the results from a Lagrangian analysis of the upwelling water is presented; results for the year 2005 are compared with those for a simulation with climatological mean wind forcing, and volume transports into the inner Oregon shelf are also discussed. In section 4 the implications of this analysis are discussed. Section 5 summarizes the main results.

2. Model setup and validation

a. Setup

For these simulations, a hydrostatic primitive equation model with a terrain-following vertical coordinate was used, namely, version 3.0 of the Regional Ocean Modeling System (ROMS; Shchepetkin and McWilliams 2005). The model was configured on a spherical-coordinate domain extending from 41° to 48°N and from 132.0° to 123.7°W (see Fig. 1), with the Coriolis parameter varying with latitude. The horizontal resolution was ∼1/30° (∼3 km), resulting in a horizontal grid of 296 × 250 points. The vertical resolution consisted of 31 layers defined by 32 sigma levels with enhanced resolution near the lower and upper boundaries (specified by the stretching parameters θs = 4.0 and θb = 0.9). The model grid was prepared by using software described by Penven et al. (2008), with topography from the 2′ gridded elevations–bathymetry for the world (ETOPO2; Smith and Sandwell 1997) dataset. Along the coast, the minimum water depth was fixed at 10 m. Bottom slopes were smoothed to meet the r-factor criterion of 0.20 to prevent horizontal pressure gradient errors (see Beckmann and Haidvogel 1993).

This model includes a splined density Jacobian scheme for pressure gradient calculations (Shchepetkin and McWilliams 2003), a third-order upstream-biased scheme for horizontal advection of momentum and tracers (temperature and salinity), and a fourth-order-centered scheme for vertical advection of tracers. Subgrid-scale mixing is parameterized by the Mellor–Yamada level 2.5 model (Mellor and Yamada 1982) in the vertical direction, and by harmonic diffusivity and viscosity, both with coefficients of 20 m2 s−1, in the horizontal direction. Lateral diffusion of tracers (temperature and salinity) is restricted to isopycnal surfaces, whereas that of momentum is restricted to geopotential (constant depth) surfaces.

The model’s eastern boundary corresponds to the coastline, where no-normal flow and free-slip conditions are applied. The western, northern, and southern boundaries are open. The implementation of the open boundary conditions is similar to that of Springer et al. (2009). Radiation plus nudging conditions are imposed for the free surface, depth-dependent horizontal momentum and tracers (temperature and salinity), and a Flather condition is used for depth-averaged momentum. Nudging (relaxation) time scales are 6 days for both passive (outflow) and active (inflow) open boundary conditions. Nudging to data was also needed in the interior of the model’s domain in order to prevent boundary-related spurious signals; this nudging was applied to the tracer fields (temperature and salinity) in the first 20 grid points from each open boundary, with scales increasing linearly from 6 days at the boundary to 60 days at the interior. In addition, the model included a sponge layer extending throughout 40 points toward the interior, within which the viscosity and the diffusivity increased (as cosine-shaped functions) from their interior values (mentioned above) to 200 and 100 m2 s−1, respectively, at the boundaries.

Two different basic numerical experiments were carried out: one using wind forcing, riverine input, and boundary data for year 2005, and another using climatological mean forcing, riverine input, and boundary data. In both cases, a climatological surface heat flux was applied. These experiments are described in detail below, along with the observational data used to assess the performance of the model prior to the Lagrangian analysis.

b. Year 2005 simulation

For the year 2005 simulation, the model was integrated for the full year using wind forcing and river discharge data for 2005. Surface heat fluxes, which for the shelf circulation are of secondary importance relative to the wind stress forcing, were taken from the climatological datasets described by Penven et al. (2008). At the model’s open boundaries, monthly mean fields of velocity, sea level, temperature, and salinity for 2005 from the ⅛° version of the Navy Coastal Ocean Model (NCOM; Barron et al. 2006) were used. The model was initialized on 1 January 2005, using the corresponding fields from NCOM.

The daily varying wind stress forcing for year 2005 was obtained from a gridded Quick Scatterometer (QuikSCAT) product. The ½°-resolution data were linearly interpolated to the model grid; for the model grid points closest to the coastline, the values were extrapolated by replacing them with the nearest interpolated values. Because orographic, air–sea interaction, and other processes may modify the wind stress substantially adjacent to the coast (e.g., Burk et al. 1999; Perlin et al. 2007), there may be significant local errors in these fields; unfortunately, in general, no estimate of the near-coast wind field is available that is demonstrably more accurate than the QuikSCAT measurements. A point comparison was made with the wind stress computed, following Smith (1988), from observations at the surface buoy described in section 2d(1), which is located approximately 21 km from the coast (see Fig. 1). The resulting complex correlation and phase were 0.86 and −6.8°, respectively; the agreement was better for the meridional component, which had a correlation of 0.91 and an RMS difference of 4.4 × 10−2 Pa, than for the zonal component, which had a correlation of 0.67 and an RMS difference of 4.1 × 10−2 Pa.

The anomalous character of the year 2005 wind forcing is evident from a comparison of the cumulative series of the meridional wind stress, averaged over the central Oregon shelf region, for year 2005 and the reference climatology (Fig. 2a). In this representation, the spring transition is indicated by the maximum cumulative stress value, which is followed by a decreasing trend during the upwelling season. In year 2005, this transition occurred on 22–23 May, about 5 weeks later than in the reference climatology, when it occurs on 18–19 April. It was not until 12 July 2005 that strong, persistent southward winds began (Barth et al. 2007).

The Columbia River inflow was partially represented in the model, through an effective freshwater input equal to one-third of the observed discharge. In the model, the river was included as a vertically distributed point source at 46.2°N along the coast (Fig. 1). The inflow vertical profile had its volume flux decreasing roughly linearly from a maximum at the surface to zero at ∼20 m, as suggested, for example, by high-resolution modeling of the estuarine exchange during upwelling-favorable wind conditions (e.g., Figs. 3c–d in MacCready et al. 2009). The inflow temperature (Fig. 2c) and salinity were taken to be vertically uniform. The inflow salinity was set to 20 psu, typical of observed mean daily values at the mouth of the estuary (e.g., Baptista et al. 2005), but the total volume input was kept equal to the observed upstream discharge. This effectively reduces the total freshwater input to one-third of its observed value, but retains the correct total volume input, since no ocean water is allowed to enter the estuary. The observed discharge and temperature during the year 2005 were obtained from U.S. Geological Survey (USGS) gauging station 14246900, at Beaver Army Station, 86 km upstream of the mouth of the Columbia River. The maximum discharge during 2005 was in the second half of March, and the minimum was in mid-September (Fig. 2b).

c. Climatological simulation

For the climatological reference simulation, the model was integrated for a full year, with the same model configuration as the year 2005 simulation described above, but with climatological forcing and boundary data. For the climatological wind stress, the daily 9-yr mean from 1 November 1999 through 31 October 2008 of the QuikSCAT product was used, and the climatological surface heat fluxes were again applied. For the Columbia River discharge and temperature, the daily 14-yr mean for 1993–2006 of USGS observations was used (Figs. 2b and 2c); as before, the inflow salinity was set to 20 psu, so the effective freshwater input was one-third the observed. The total discharge for the year 2005 was lower than the climatological total, but the range of daily values was similar. At the boundaries, the climatological temperature and salinity fields described by Penven et al. (2008) were used. The corresponding velocity values were taken as the geostrophic flow relative to 1000 dB, while the free-surface conditions were derived from the dynamic topography anomaly relative to 500 dB; this inadvertent discrepancy in the reference levels resulted in a slight departure from geostrophic balance in the open-boundary conditions for the climatological simulation, but, relative to the behavior for the year 2005 simulation, there was no indication of additional, significant complications at the boundary. The model was initialized by using fields interpolated to 1 January from the monthly climatological fields of Penven et al. (2008).

d. Comparison with observations and previous simulations

1) Observed data

A number of diverse observed datasets from year 2005 were used to assess the basic performance of the model, prior to the Lagrangian analysis. Among the in situ measurements available for year 2005 were current-profile observations from a long-term mooring located on the Oregon shelf at 44°38.8′N, 124°18.3′W (see Fig. 1), at 81-m water nominal depth (see Kosro 2003; Kosro et al. 2006). Other mooring observations were available from the River Influences on Shelf Ecosystems (RISE) program (see Hickey et al. 2010), from May to October 2005 near areas influenced by the Columbia River plume (see Fig. 1). These moorings included measurements of velocity, temperature, and salinity at different levels.

Additional in situ observations during 2005 include sea surface temperature series from National Data Buoy Center (NDBC; information online at http://www.ndbc.noaa.gov/maps/Northwest.shtml) buoy 46050 (see Fig. 1) and coastal sea level from a tide gauge at nearby South Beach, at station 592 (44°37.5′N, 124°02.6′W; see Fig. 1) provided by the University of Hawaii Sea Level Center (UHSLC; information online at http://uhslc.soest.hawaii.edu/uhslc/datai.html). The sea level data were adjusted using local atmospheric sea level pressure, following Strub et al. (1987). The atmospheric pressure was taken from the North American Regional Reanalysis (NARR; see Mesinger et al. 2006), which is in close agreement (correlation of 0.99 and RMS error of 1.1 hPa) with that recorded at buoy 46050; the reanalysis pressure was used because it covers the full year 2005, in contrast with the buoy 46050 observations, which are missing prior to mid-April.

Remote observations are also included. These include sea surface temperature from the monthly averaged level-3 standard mapped 11-μm images available at the Aqua–Moderate Resolution Imaging Spectroradiometer (MODIS) Web site (http://oceancolor.gsfc.nasa.gov/cgi/l3), which have a resolution of 4 km. Images for day and night were taken and averaged. Sea level anomaly is also included; this was obtained from the multiple-satellite [Jason-1, the Environmental Satellite (Envisat), the Geosat Follow-On (GFO), and the Ocean Topography Experiment (TOPEX)/Poseidon] merged gridded maps available in the Archiving, Validation, and Interpretation of Satellite Oceanographic Data (AVISO) Web site (http://www.aviso.oceanobs.com). These data have a spatial resolution of ¼° and a temporal resolution of 7 days.

2) Shelf circulation

To assess the performance of the model, results of the year 2005 simulation were compared with the observations described above. These comparisons focus primarily on the shelf circulation during the April–September upwelling season, which is most significant for the subsequent analysis.

Horizontal velocity at the Northwest Association of Networked Ocean Observing Systems/Center for Coastal Margin Observation and Prediction (NANOOS/CMOP) NH10 mooring from the 2005 simulation is in reasonable agreement with the observations (Fig. 3). At 10-m depth the standard deviations of the observed and the modeled series are 24.3 and 20.0 cm s−1, respectively; whereas, at 68-m depth these are 16.0 cm s−1 for the observations and 11.1 cm s−1 for the model. The corresponding complex correlation and the phase between the model and the observations at these depths are r10m = 0.76, ϕ10m = 1.6°, and the correlation and phase are r68m = 0.77 and ϕ68m = −6.2°. If the annual cycle in the form of a single annual period harmonic is first removed, the resulting correlation and phase values are r10m = 0.63, ϕ10m = 1.8°, and the r68m = 0.72, ϕ68m = −7.8°.

During summer, there is southward flow in the surface-intensified coastal jet associated with the upwelling front. Interestingly, the formation of this jet occurred in mid-April, about 1 month before the onset of local upwelling-favorable winds; the role of remote forcing in this early jet formation will be explored elsewhere (D. Rivas and R. M. Samelson 2010, unpublished manuscript). The comparison with the NH10 mooring data suggests that the model tends to overestimate the strength of this southward flow during the July–September period (Fig. 3a), which may represent errors in the representation of either the magnitude or the position of the coastal jet.

At shelf buoy 46050 (Fig. 4a), the model reproduces the dominant observed variability in near-surface temperature, showing simultaneity in the cold water signal associated with the upwelling, but overestimating this in late August and early September. According to Pierce et al. (2006), year 2005 was anomalously warm compared with previous years, with temperatures in mid-July up to 5°C warmer than the 1962–68 mean along the NH Line (see Fig. 1); similar calculations in our model show a slightly smaller anomaly of 4°C (not shown). Note that point comparisons of model and observed sea surface temperature (SST) are complicated by the large spatial gradients, which can lead to large SST differences from relatively small circulation errors; differences in heat fluxes and mixed layer depth can also contribute.

The model sea level at the South Beach tidal gauge (Fig. 4b) is in reasonable agreement with the observed adjusted sea level, with a correlation of 0.77, but with a somewhat smaller standard deviation of 5.8 cm, compared with 7.9 cm for the observations. Most of the large-amplitude sea level events present in the observations are reproduced by the model.

3) Large-scale circulation

The May–October 2005 mean fields of sea surface temperature and sea level anomaly (with respect to the annual mean) show a general similarity to the corresponding observed fields (Fig. 5), and to similar fields for the year 2001 presented by Springer et al. (2009, their Fig. 4) for the year 2001. The present model and that of Springer et al. (2009) both show excessively low temperatures along the Washington shelf, possibly related to an overestimated southward inner-shelf flow [see section 2d(2)]. The salinity in both models is generally similar, with the present model being slightly fresher in the inner shelf south of Cape Blanco (not shown). Comparison with three available offshore vertical hydrographic profiles during mid-June from Argo floats (e.g., Gould et al. 2004) suggests that the model may underestimate the thicknesses of the offshore mixed layer and thermocline by a factor of 2 or 3, perhaps leading to offshore model temperatures about 1°–2°C higher than in the observations at the surface. The offshore halocline appears to be better reproduced, with salinities about 0.1 psu fresher than the observed profiles (not shown).

The model sea level anomaly is qualitatively similar to the observations (Figs. 5b and 5d), and also to that found by Springer et al. (2009) for the year 2001. Persistent mesoscale features (e.g., Stegmann and Schwing 2007) are found seaward of the continental shelf, with some rough correspondence in location between the model and observations, with gradients stronger in the observations than the model in the southern half of the domain.

4) Columbia River plume

Between May and June 2005, the monthly mean Columbia River plume position switched from a typical mean winter flow northward along the coast, to a summer regime, in which the monthly mean flow is typically directed southwestward and a minor branch flows northward (see Hickey et al. 2005, 2009; Liu et al. 2009a). This switch occurred about 1 month later than normal, and the summer regime persisted through October. As in the case of the sea surface temperature (Fig. 4a), comparisons with mooring observations in the vicinity of the Columbia River mouth during this period show good simultaneity at the start of the upwelling signal (Fig. 6), with correlations generally greater than 0.6 for temperature, salinity, and velocity. There are, however, significant differences between the model and the observed series, especially in the surface salinity.

To provide more details about the comparisons mentioned above, here we provide some basic statistics; notice that the correlations (and phases, in the case of the velocities) were calculated after removing a partial annual cycle [represented by a 365-day-period harmonic, as in section 2d(2)]. For temperature and salinity, the largest differences occur at the river mouth near 5-m depth (Fig. 6a), where the model fields are colder and saltier than the observations, with RMS errors of 2.3°C and 3.4 psu, respectively, and correlations of 0.79 for temperature and only 0.12 for salinity; better agreement is found at 20-m depth, with RMS errors of 1.0°C and 0.4 psu, respectively, and correlations of 0.78 for temperature and 0.69 for salinity. For the velocities, the largest differences are found at 20-m depth, where the complex correlation amplitude and phase are 0.61 and −17.5°, compared to 0.64 and −18.2° at 5-m depth. South of the river mouth, the model fields still tend to be too warm and salty, but the differences are smaller, with RMS errors of 1.8°C and 1.6 psu, and correlations of 0.71 for temperature and only 0.17 for salinity (Fig. 6b). At 20-m depth, the differences are essentially the same as at the river mouth, with RMS errors of 1.7°C and 0.4 psu, and correlations of 0.71 for temperature and 0.68 for salinity. The velocities show better agreement than those for the river mouth, with correlation and phase of, respectively, 0.77 and −3.8° at 5-m depth, and 0.78 and 5.8° at 20-m depth.

The differences described above are likely due in part to the underestimated freshwater input specified in the model. They suggest also that a more sophisticated representation of the estuary (MacCready et al. 2009; Liu et al. 2009b) may be necessary, for example, to represent the entrainment of bottom shelf water into the plume via tidal mixing during upwelling (Hickey and Banas 2008; Hickey et al. 2009; MacCready et al. 2009). With the exception of the surface (5 m deep) salinity, this comparison suggests that the present, simplified representation of the river-induced circulation in the region should be adequate enough for the present study, in which the focus is primarily on the shelf-scale upwelling circulation.

3. Upwelling source waters

a. Mean circulation

This analysis focuses on upwelling into a coastal region along the Oregon shelf (Fig. 1, shaded region), which is hereafter referred to as the upwelling region. This region, which covers most of the Oregon shelf, is limited to the west approximately by the 200-m isobath, and to the south and to the north by the latitudes of 41.5° and 45.5°N, respectively. A brief summary of the volume transport patterns is given first, to provide context for the Lagrangian analysis that follows.

Offshore Ekman transport through the western boundary of this region was generally stronger in year 2005 than in the climatology (Fig. 7a); even twice as large during some periods (e.g., late July). In contrast with the climatology, during the 2005 upwelling season some wind reversals occurred during May and June. Associated with these wind reversals in mid-May and mid-June are reversals in the velocity series shown in Fig. 3a, during which the model and observed surface current reversed to northward for a few days.

The offshore Ekman transport is partially compensated by onshore flow through the western boundary of the region, by means of net exchanges to offshore. The net exchange through the western boundary of the upwelling region (the 200-m isobath) is compensated almost completely by exchange through its northern boundary (at 45.5°N), since the southern boundary (of the upwelling region) contributes only a small fraction to the net volume exchange (Figs. 7b and 7c). There is a correlation between the vertically integrated net transport across the western boundary of the upwelling region and the across-shelf Ekman transport, with the former transport being about 29% of the Ekman transport in year 2005, and about 58% in the climatology. This suggests that, in addition to the balance between the exchange through the Ekman layer and the across-shelf exchange beneath it, there is an excess in volume supplied by exchanges north of the upwelling region that is roughly proportional to the Ekman transport.

Near-surface and near-bottom mean velocities for mid-July–September show in more detail how the transports occur during the upwelling season (Fig. 8). Near-surface net incoming transport into the upwelling region is driven by the southward surface-intensified jet in the north, but part of this gained volume is lost offshore near Heceta Bank around 44.3°N, where a portion of the jet deflects from the shelf. The deflected flow turns back onshore south of the bank and then turns southward to make a significant contribution to the southward main flow. The rest of the volume leaves the upwelling region near Cape Blanco, where the remaining jet separates from the shelf. These patterns are consistent with observational results (Barth et al. 2000). Near the bottom, a coherent poleward undercurrent (>5 cm s−1) develops below 150-m depths, as has been reported in observational results (e.g., Pierce et al. 2000; see also Gay and Chereskin 2009). This northward flow supplies some of the volume to the upwelling region, especially south of Cape Blanco and south of Heceta Bank, but as will be shown in the next sections, it apparently plays only a small role as an immediate, direct source of the upwelling waters.

The mid-July–September mean velocity along the NH Line (Fig. 9) provides more information about the flow over the central Oregon shelf and slope. The coastal jet is significantly more intense and thicker in year 2005 than in the reference climatology. In both simulations the cross-shelf component is significant; the near-bottom onshore flow is restricted to a shallow layer in year 2005, whereas it occupies most of the water column in the climatology. Basic characteristics of the observed poleward undercurrent are reproduced in both simulations, at depths around 250 m. For the year 2005 simulation, the magnitude of the poleward core is comparable to the instantaneous values reported by Pierce et al. (2000) from observations during summer 1995. In the climatology, the poleward core is weaker, but its magnitude is consistent with the 1997–2003 mean reported by Huyer et al. (2007).

b. Lagrangian analysis

1) Overview

The paths of the upwelling source waters may be analyzed, from a Lagrangian point of view, by computing the paths of particles that are passively advected by the model velocity field. The approach taken here was to initialize groups of particles along the Oregon shelf at regular intervals during the upwelling season, and then advect these particles backward in time for 180 days, in order to determine the locations of the sources of the upwelling waters. To prevent confusion in terminology, the paths are described here from the standard, physical, forward-in-time perspective, in which the particle paths originate at their initial positions and terminate, at a later time, at their final positions. Computationally, the paths were obtained instead by backward-in-time integrations, which originated at the particle final positions, and terminated at the particle initial positions.

The physical final positions of the particles, which were prescribed as the initial conditions for the backward integrations, were chosen to be evenly distributed horizontally at 2-m depth within the upwelling region, with a total of 201 particles per release. For such sets of prescribed particle final positions, 6-month backward-in-time integrations were carried out daily during the period from July through September, giving a total of 18 492 particle paths. This was done for both the year 2005 and the reference climatology. The model’s three-dimensional velocity u = (u, υ, w) was used for the advection, with a standard bilinear interpolation to the instantaneous particle positions. Although the small errors resulting from the interpolation could have large effects on individual trajectories, for example, in the vicinity of stagnation points with horizontal divergences, their mean effects on the large number of trajectories computed here should be relatively small. The particle-path time integration used an Euler scheme and a time step of 3 h, with no diffusion.

Note that the presence of diffusion and mixing in the model introduces an inevitable ambiguity into any attempt to analyze Lagrangian motion of water parcels, since such mixing can destroy the identity of the parcels. The present approach offers the least ambiguity, in which the resolved model velocity is used directly for the advection, and the ambiguities and complexities of attempting to include a representation of mixing are avoided, but can be misleading. To supplement this approach, rates of change of the water properties along Lagrangian pathways were computed, and these are discussed below.

In addition to the trajectory integrations described above, we carried out similar Lagrangian calculations for particles with physical final positions specified on vertical cross sections at the NH and Crescent City latitudes (see Fig. 1), on a regular spatial grid with 2.5-km separations in the horizontal and 15-m separations in the vertical. These experiments were done to further address the dependence of source regions on the final positions of the upwelling particles, as will be discussed below.

For the 6-month trajectory time scale and the regional circulation model considered here, many particle paths are found to originate outside the computational domain. That is, the backward integrations from the final positions result in trajectories that leave the domain in backward time. This pattern of behavior is apparent, for example, for the particles with final positions specified in the upwelling domain on 29 September 2005 (Fig. 10), the date of minimum cumulative wind stress (see Fig. 2a). The southward advection that carries many of these particles into the domain from the north is apparently associated with deeper levels of motion in the coastal upwelling jet that forms along the mid- and outer shelf.

Notice that the alongshelf paths present some offshore deflection in the vicinity of the Columbia River’s mouth, affected by recirculation features around the riverine plume; this is consistent with previous advective results that show that the river and its induced motions favor the cross-shelf transport (Hickey and Banas 2008; Banas et al. 2009). Those features may be underestimated in our model, due to the diminished freshwater input (see section 2b), so the offshore deflection near the river should probably be larger. However, advection paths of particles released close to the surface at 48°N over the continental shelf by Banas et al. (2009, their Fig. 3) are generally consistent with our results, as they are deflected offshore near the river’s mouth after roughly following the 100-m isobath, but many then pass around the bulge and turn toward the shelf, ultimately reaching the central Oregon shelf. Also, the alongshelf paths of particular interest in our experiments are found mainly offshore of the 100-m isobath and below 50-m depth; hence, they are less likely to be influenced by the surface plume. Therefore, we believe that if the freshwater discharge were increased by a factor of 3, the upwelling particle path results would likely not be dramatically affected.

More direct pathways from offshore also exist for the set shown in Fig. 10, especially around Heceta Bank (∼44.3°N), where paths cross onto the shelf from the west. There are also zones where the particles follow roughly circular paths, for example, near 46.0°N and 126.5°W, which suggest the existence of vortical structures close to the shelf break.

In the following discussion, we present separately the particle paths originating at or near the northern or southern boundaries of the model domain, which generally represent alongshelf sources of upwelling water, and the paths coming from locations within the domain, which generally represent offshore sources. Although this grouping is, to some degree, arbitrary and dependent on the choice of computational domain, it is necessary in the analysis to distinguish between particles that enter from outside the computational domain, for which full 180-day trajectories are either not available or are affected directly by the conditions imposed along the open boundaries, and those that are contained inside the central computational domain for the full 180 days. The distinction also proves useful for establishing some aspects of the temporal and spatial scales for the particle displacements, as well as emphasizing motion patterns. To avoid or at least reduce artificial boundary effects that the nudging and sponge layers (see section 2a) could cause in the particle motions, paths that originated within 0.2° of the three open boundaries were grouped with those that originated at the edge of the domain.

The following discussion focuses on the results for particles with final positions specified in August. Of the full set of particle paths with final positions in August for year 2005 (a total of 6231 particles released during this period), 58% enter the domain from the north, 32% originate offshore, 4% enter the domain from the south, and 6% originate within the upwelling region. For the climatological simulation, the corresponding numbers are 35%, 49%, 5%, and 11%, respectively.

The general structure of the particle paths can be summarized by considering the mean particle paths (Fig. 11). These mean paths are computed by averaging the 180-day-long paths depicting one of the patterns (groups) described in the next sections. Those paths coming from outside of the domain have generally different lengths, depending essentially on the distance from the corresponding model’s zonal boundary to the upwelling position along the Oregon coast. The mean paths from these length-varying paths were computed by averaging the zonal positions of each trajectory along the latitudinal coordinate:
i1520-0485-41-1-88-e1
where yupw is the latitude at which the particles upwell along the Oregon coast, ybound is the latitude of the corresponding zonal boundary (southern or northern), and yp is the latitude along the path; angle brackets indicate an average. In the computation of the coordinate y*, each particle path is normalized by the distance from the boundary the corresponding particle enters the domain and its final position within the upwelling region; notice that y* is dimensional and its values extend from approximately 〈yupw〉 to ybound. There are significant differences among the mean paths (Fig. 11; see also Fig. 10). Those paths coming from the north show the regular pattern of the mean position of the coastal jet. Those paths coming from offshore show a meandering pattern associated with the mesoscale features prevalent seaward of the shelf. Those paths coming from the south show a veering from the northward to the onshore direction, seaward of the shelf break. Results for the particles with final positions specified in September are qualitatively similar to those for August.

2) Alongshelf sources

Particles with paths originating outside of the model domain and arriving at their final positions in August are found throughout the entire upwelling zone (Fig. 12). A remarkable result is the clear regime separation at Cape Blanco (∼42.8°N). South of this region, a significant percentage [∼(10%–60%)] of the 31 particles released in each position during August) of the upwelled waters come from southern locations beyond the model’s domain, whereas north of it, most of the particles (percentage > 50%) come from remote northern locations. This regime separation is even clearer for the climatological simulation; for year 2005, there is a significant incursion of northern particles into the region south of Cape Blanco.

In the northern part of the upwelling region, these paths originate at levels below 50–100-m depth, taking about 30–70 days to arrive at their final positions after leaving the northern model boundary (Figs. 12c and 12d), following the coastal jet that enters the upwelling region and then flows offshore near Heceta Bank (see Fig. 8). In the central part, the paths originate at deeper levels, below 80 m, taking 80–100 days to arrive from the northern boundary, advected by the part of the coastal jet that remains attached to the shelf. In the southern part the particles come from 50–100-m depth and take longer, 90–160 days, to arrive from the southern boundary, as they are advected by a weaker flow. For the climatological simulations, the particle depths along the trajectories arriving at the upwelling region are restricted to the upper 100 m, with some intrusion from 100 to 150 m near the coast at the central part. In the northern part the particles take 30–120 days to arrive, slightly longer than in year 2005, due to the weaker and shallower jet. In the central part they mostly take longer than 90 days, and in the southern part they take 30–60 days.

The results from the particle paths calculated for Crescent City and NH latitudes are consistent with the results described above. Particles along the Crescent City Line come mostly from the south, but in the year 2005 they come also from the north; this pattern is the same for the whole water column (not shown). Particles in the NH Line present some different patterns in the deep layers with respect to the surface layers. In the year 2005, the water column is dominated by particles coming from the north above 100-m depth, where practically all the particles come from those locations (Fig. 9a); below 100-m depth, particles of different origin are also present. Interestingly, in these depths there is some evidence that particles can come from south of the model’s domain, probably advected by the slope undercurrent (Fig. 9a; see also Fig. 8). Note however that these particles have relatively deep final positions and have not upwelled into the surface layer, so this pathway cannot be clearly identified as a source of upwelling water. In the climatology, only the 50–60-m-thick surface layer is dominated by particles coming from northern locations (Fig. 9c); below this level, the particles do not come from the north but they come from offshore locations, which will be discussed in the following section.

3) Offshore sources

In addition to the alongshelf sources associated with paths arriving from the northern and southern boundaries, there are offshore sources illustrated by the many paths that arrive from locations directly west of the upwelling zone. For the analysis of this group of particles, it is useful to define five complementary geographical regions within the model domain (Fig. 1). Three of these regions are adjacent to the upwelling region and extend 300 km offshore of the 200-m isobath; the southernmost and northernmost of these reach the coast south of 41.5°N and north of 45.5°N, respectively. Westward of these 3 regions, the last 2 regions extend offshore up to the western boundary of the model’s domain.

Six months before they upwell within the Oregon shelf in August, most of the offshore-source particles are found no farther than 300 km offshore of the upwelling region (Fig. 13), in both the year 2005 and climatological simulations; however, there are significant differences in the particle paths between one simulation and the other. Year 2005 is characterized by a vigorous exchange in the meridional and vertical directions during the movement from offshore into the upwelling zone. Many particles originally in the north upwell in the south and vice versa, with initial depths spread over the upper 400 m, and some initial levels deeper than 600 m. In the climatology, on the other hand, the particles generally travel with minor latitude changes and mostly originate within the upper 200 m.

Another qualitative difference between the simulations is the presence in year 2005 of several empty patches adjacent to the shelf and slope, around which the particles apparently follow well-defined paths, for example, in the region southwest of Cape Blanco (Fig. 13). These structures are apparently associated with persistent mesoscale eddies that originate seaward of the 1000-m isobath. These vortical structures seem to enhance the meridional movement of the water parcels, through circulation around the vortex cores as they approach to the shelf. In the climatological simulation such structures are not clearly apparent, suggesting that they may arise in part from the isolated strong wind-forcing events that are present in the year 2005 simulation but not in the smoother, multiyear average forcing for the climatological simulation.

Figure 14 summarizes the statistics of the results for the offshore-source particles in Fig. 13, showing the initial positions of the particles 6 months before they arrive at their final positions in the upwelling zone (i.e., the positions shown in Fig. 13). North of Cape Blanco, most of the offshore-source particles come from the west and northwest regions (W and NW), from depths between 100 and 250 m for year 2005 and within the upper 150 m for the climatological simulation. There is also some contribution from the far-offshore region (ON), but only for year 2005. South of Cape Blanco, the offshore-source particles come from all regions except ON for the year 2005, but primarily W and SW for the climatological simulation. Most of these particles originate from depths of 100–250 m for the year 2005, with 32% of the particles coming from shallower levels, while for the climatology, most originate instead in the upper 100 m.

The results for the additional particles with final positions on the NH and Crescent City cross sections are again consistent with the results described above. However, there are a few details for the case of the NH Line that are worth pointing out here. In the year 2005, about 20% of the total particles of the section have an offshore origin, which are concentrated within a shallow layer close to the bottom (Fig. 9b). This percentage is almost equally divided into particles coming from the W, NW, and OS regions, from depths 50–110 m below their final positions along the section. In the climatology, on the other hand, about 63% of the total particles of the section come from offshore locations, practically all of them concentrated below 70 m-depth (Fig. 9d); indeed, in these deeper levels no alongshelf-source particles are found (see Fig. 9c). The percentage of the offshore-source particles is roughly divided into two-thirds of the particles coming from region W and about one-third of them coming from region NW, from depths 40–60 m below their final positions along the section. Interestingly, the offshore-originated particle positions in both simulations coincide with the thickness of the onshore flow in the sections, consistent with the typical notion of a cross-shore upwelling cell. But as mentioned before, the particles have not upwelled into the surface layer; hence they are not clearly identified as part of the upwelling water.

4) Vertical motion, tracer conservation, and TS properties

There are specific regions along the shelf where much of the vertical motion occurs in the offshore-source upwelling pathways. For example, in the year 2005, many particles pass upward through the 150-m-depth level between 45° and 47°N, and inshore of the 200-m isobath (Fig. 15). Similarly, onshore motion across the 1000-m isobath in the northern half of the domain is concentrated near 46.5°N (Fig. 15). Particles in the southern half of the domain are often above 150 m by the time they cross the 1000-m isobath, but show similar latitudinal localizations in vertical and cross-isobath motion (Fig. 15). In the climatological simulation, the particles at all latitudes generally are at depths of less than 150 m when they cross the 1000-m isobath, but the latitudinal distribution of the crossing latitudes remains nonuniform. The enhanced vertical displacements during the year 2005 may in part be associated with the vortical structures over the outer shelf and slope, apparent as blank regions in Fig. 13, and these may also enhance the cross-isobath exchanges. There are also “shadow regions,” where the cross-slope motion nearly vanishes, surrounding the regions of enhanced exchange (Fig. 15); these appear to be associated with topographic features. In the climatological simulation, a clearly defined region of convergence of particles crossing the 1000-m isobath is centered at about 43.8°N, near Heceta Bank, and it is mostly in this region that the particles come from levels below 150 m. The particles crossing the 1000-m isobath south of 43.5°N generally maintain their latitude as they move onshore, whereas those crossing the 1000-m isobath north of this latitude spread along nearly the full meridional extent of the upwelling zone. There is also enhanced convergence of particles around the narrow canyon located near 46.4°N. These topographically modulated pathways are less clear in the year 2005 than in the climatological simulation, perhaps because of the additional stirring induced by the mesoscale vortical structures for the year 2005.

To assess the importance of turbulent diffusion along particle paths, which is in effect neglected in the trajectory calculations, the turbulent heat and salt diffusion terms were diagnosed for all the particle paths, including both the along-shelf (Fig. 12) and offshore (Fig. 14) sources. The changes in temperature and salinity, relative to the initial values, were computed along each path as a function of time tp, with tp = 0 at the final time (Fig. 16). The patterns are qualitatively similar for the climatology and for the year 2005, differing essentially in the magnitude of the changes. The greatest changes of the tracers occur during the last 10–30 days of the paths, near tp = 0, where the particles reach shallower depths and are affected by energetic mixing over the shelf and in the surface boundary layer (Fig. 17). For early times (large values of |tp|), relatively minor changes in the offshore-source temperature and salinity occur, which must be associated with small mixing at middepths in offshore locations. Interestingly, the variance of the changes in the three groups of particles (along-shelf source and offshore-source particles) are different, with the particles from the northern boundary showing the greatest differences (even when their vertical displacements are smaller), and those from the southern boundary showing the smallest differences. This is probably because the displacements of the northerly parcels are essentially along the coast; hence, they are more exposed to the coastal mixing, in contrast with the southerly parcels, as their displacements present a larger across-shore component. Those parcels originating within the domain have the greatest vertical displacement, with a tracer modification of magnitude intermediate between the southerly and the northerly parcels. Although the tracer standard deviations along the paths are substantial (Fig. 16), suggesting significant levels of mixing all along the paths, the mean changes are small prior to the latter segments, when mean depths decrease rapidly and upwelling into the turbulent surface layer occurs. This nearly adiabatic behavior of the mean along-path tracer values suggests that the qualitative picture of the sources of upwelling water that is provided by the effectively adiabatic trajectory calculations should be relatively robust.

An additional Lagrangian element of the simulated circulation, only indirectly related to the upwelling motion but still of general interest, is the motion of the river plume, which can be traced from its salinity signature (Fig. 18). In June, the influence of the warmer and fresher water from the Columbia River is evident at both the NH and the Crescent City sections for the year 2005, but not for the climatological simulation. In July, the riverine water is barely observed at either section from the climatological simulation, as it has been pushed offshore by the upwelling circulation, but it is still present for year 2005. The presence of riverine water in July at the Crescent City Line during 2005, and its absence in the climatology, may be associated with the wind reversals, which are strong in year 2005 and practically absent in the climatology (see Fig. 7a). These reversals force intermittent strong onshore Ekman flow in a shallow surface layer and may, thus, support the maintenance of a riverine influence over the shelf. During August and September, the thermohaline characteristics at both lines are entirely those of the upwelled waters. The greater scatter during August–September for the year 2005 simulation suggests a more vigorous mixing in the upwelling region and also stronger entrainment of waters from different offshore locations, relative to the climatological simulation. By the last stages of the upwelling season, August–September, the climatology shows a very clear distinction between the water masses corresponding to each section. In year 2005, on the other hand, the two distinct water masses are observed, but also water with intermediate characteristics. This is consistent with the Lagrangian analysis (Fig. 12), which shows that south of Cape Blanco, in several coastal locations, the upwelled water can come either from the north or from the south, in year 2005.

4. Discussion

The analysis in the preceding sections addresses, from numerical simulations, the origin of the upwelling source waters along the Oregon coast during year 2005 and during a mean climatological year. Two major results of the analysis are that roughly half of the upwelling waters in the Oregon coastal zone arrive from alongshelf locations well north or south of the zone and that Cape Blanco is a dividing point for northern versus southern alongshelf sources. The results also show that there are preferred locations along the shelf for vertical and cross-shore motion.

The identification of a dividing point near Cape Blanco for northern versus southern sources of upwelling water is consistent with results from hydrographic and biochemical samples along the NH (44.7°N) and Crescent City (42°N) Lines during five summers (1998–2000 and 2002–03), as reported by Huyer et al. (2005). These surveys showed systematic differences in water properties between the two lines, with Crescent City having a more saline, cooler, denser, and thicker mixed layer; higher nutrient concentrations in the photic zone; and higher phytoplankton biomass. Huyer et al. (2005) cite the reduced influence of the Columbia River discharge on properties at the Crescent City line as one of the factors responsible for these significant differences. This reduced riverine influence is clearly observed in the model TS diagrams (Fig. 18), where the riverine water is always more evident at the NH Line than at the Crescent City Line. Indeed, such a riverine influence at this latter location is essentially absent from the climatological simulation. We attribute this absence to the lack of reversals in upwelling-favorable winds in the climatological mean (Fig. 7a), which results in uninterrupted southwestward and offshore motion of the river plume.

As shown in the previous sections, the model poleward undercurrent has only a small direct contribution to the upwelling source waters (Figs. 11 and 14). It is important to recognize that this conclusion holds only on the limited temporal and spatial scales of the simulations: for example, on longer time scales and larger space scales, flow from the undercurrent could well contribute to the shallow northern upwelling source found here, from which fluid enters the regional domain at depths that are above, and inshore of, the core of the model poleward undercurrent. In addition, the simulations may be consistent with an undercurrent source for many of the model particles with specified final positions at near-bottom levels over the shelf. Watermass analyses for the British Columbia coast during summer (Mackas et al. 1987; MacFadyen et al. 2008) have shown that ∼75% of the water below 50-m depth has thermohaline characteristics attributed to the “California Undercurrent Core” (described typically by T = 6.9°C, S = 33.9 psu, and σt = 26.6 kg m−3). Our results show that in the climatological simulation about two-thirds of the advected particles below 70-m depth in central Oregon (Figs. 9c and 9d), come directly from the west, from depths below 150 m. In the anomalous year 2005 simulation, on the other hand, this occurs for only about 12% of the near-bottom waters [almost equally divided in western and offshore southwestern origins, see section 3b(3)]. Thus, under the assumption that these western particles represent parcels carrying mostly undercurrent-like water, our climatological results would be in reasonable agreement with the estimated thermohaline conditions reported for the British Columbia coast (Mackas et al. 1987; MacFadyen et al. 2008), whereas our year 2005 results would be closer to the wintertime conditions of ∼6% of such water masses (MacFadyen et al. 2008).

The trajectories of upwelled parcels that come from offshore show differences between year 2005 and the climatological simulation, but in both cases there are specific regions along the shelf where the vertical motion in the upwelling pathways is enhanced, associated with topographic features such as submarine banks and canyons (e.g., around Heceta Bank). This is in agreement with the notion that a portion of the water present in the shelf is derived from upwelling along shelf-break canyons (e.g., Hickey 1997; Allen and Durrieu de Madron 2009), a process that favors the nutrient supply from deeper levels (Crawford and Dewey 1989; see also Bosley et al. 2004; Hickey and Banas 2008).

The comparison of the circulation response for year 2005 and for the climatological wind forcing allows insights into the influence of large-scale atmospheric conditions on the regional ocean circulation. Year 2005 is characterized by a delayed, shorter but stronger, upwelling season (Fig. 7), relative to the climatology, as indicated by the Ekman transport along the Oregon coastal upwelling zone east of the 200-m isobath and between 41.5° and 45.5°N, which was about 33% greater during the period July–September 2005 than in the climatological mean. In year 2005, about one-third to half of the net surface-layer transport lost to offshore from this zone, is supplied by inflow from, primarily, the northern boundary via the coastal jet; the remainder is compensated for by onshore flow at depth between these latitudes. During 2005, the water parcels cover longer distances in the 6-month-long advection period than those in the climatological simulation (Fig. 14), with a much larger number of particles coming from locations more than 300 km away from the upwelling region. In year 2005 about 30% of the parcels come from levels between 150- and 250-m depth, whereas in the climatological simulation they are mostly found in the first 150 m. Also, the meridional exchange is more vigorous in year 2005, as the initial positions of particles entering the upwelling zone from offshore have a broader meridional distribution (Fig. 13), with their subsequent paths onshore sometimes including roughly circular patterns. These patterns seem to be associated with mesoscale eddy structures, which are broadly similar to features that have been observed to form seaward of the shelf (Stegmann and Schwing 2007) or from the California Undercurrent (Garfield et al. 1999). Such eddy structures are not clearly observed in the climatology, which suggests that short-time-scale, large-amplitude wind fluctuations or reversals may be associated with their formation. Regardless of their formation mechanism, these eddies seem to favor the meridional transport and shallowing of the water parcels, and have a strong effect on their pathways into the upwelling zone.

The present Lagrangian results are in some aspects consistent with the results of Chhak and DiLorenzo (2007), who argue on the basis of passive-tracer results from a coarser-resolution, 20-km-grid circulation model in a larger domain that during a “warm” phase (post-mid-1970s) of the Pacific decadal oscillation (PDO), like that for year 2005 (information online at http://jisao.washington.edu/pdo/), much of the water mass upwelled along the Oregon and northern California shelves originates from offshore regions and from the north, within the upper 100 m, indicating a shallow upwelling cell and a strong influence of lateral advection. On the other hand, our results show a somewhat deeper upwelling cell with a significant percentage of waters coming from levels below 100 m. It is possible that this difference may be associated with the presence of the mesoscale vortical structures, which would not be well resolved in the coarser-resolution model.

We anticipate that the present analysis of the paths of the upwelling source waters, as well as their dependence on atmospheric conditions that may vary interannually, will be useful as a step toward improved understanding of coastal phenomena such as upwelling-driven hypoxia events. Since the unprecedented development of severe hypoxia along the central Oregon shelf during 2002 (Grantham et al. 2004), an intensification of these conditions, including the novel appearance of water-column anoxia during 2006, has been observed (Chan et al. 2008). The development of the abnormally low dissolved oxygen levels appears to be a response to the anomalously strong flow of nutrient-rich, subarctic water into the upwelling zone (Grantham et al. 2004; Freeland et al. 2003; Strub and James 2003; Barth 2003; Kosro 2003), an essentially Lagrangian phenomenon that emphasizes the importance of determining the sources of upwelling waters and how those sources depend on external forcing and other conditions.

5. Conclusions

We have presented numerical simulations and associated Lagrangian particle-tracking results for circulation in the Oregon coastal upwelling regime based on surface forcing from the year 2005 and from a mean climatological year. It was found that roughly half of the upwelling waters in the Oregon coastal zone arrive from alongshelf locations well north or south of the zone, and additionally that Cape Blanco is a dividing point for northern versus southern alongshelf sources. The results also show that there are preferred locations along the shelf for vertical and cross-shore motion. A perhaps surprising result is the apparently small role played by the model poleward undercurrent as an immediate, direct source of upwelling fluid; on the seasonal time scale considered here, most upwelling fluid parcels over the slope and shelf are found to move southward, with weaker northward motion only farther offshore, in the deep interior. On longer temporal and larger spatial scales, the simulations are not necessarily inconsistent with the traditional hypothesis that the poleward undercurrent may supply much of the water that ultimately upwells onto the shelf.

These results and others discussed above represent a step toward a quantitative understanding of the locations and controls of the source waters for the upwelling circulation in the Oregon coastal zone. Their validity relies on the veracity of the circulation model simulations, as well as on more specific details of the Lagrangian calculation. Comparison with available data indicated that the model generally reproduced basic aspects of the observed flow, but existing data provide only a limited constraint on the model solution and are largely absent in many regions covered by the Lagrangian trajectories. Anticipated enhancements of coastal ocean observing systems, and advances in data assimilation modeling, may provide stronger constraints over expanded regions, allowing future improvements and extensions to the present study. Other elements deserving additional attention include the interpretation and representation of Lagrangian particle trajectories in the presence of mixing by unresolved, subgrid-scale motions. In the present approach, the trajectories were computed following the resolved velocity field, with no effects of diffusion, under the assumption that the use of a large ensemble of trajectories will give representative information on the mean Lagrangian motions, accompanied by a posteriori diagnostics of changes in TS properties along the derived trajectories.

The basic dynamics of coastal upwelling, in which offshore Ekman transport leads to a coastal divergence that must be compensated by upward motion of deep fluid, have been understood for over a century. It is a testament to the difficulty and subtlety of the full problem, however, that our understanding of the pathways by which the deep fluid reaches the surface remains severely limited. Knowledge of these pathways, and of the source properties of upwelling water and the mechanisms that control the locations of these sources, are and will be essential to understanding the basic elements of the response of the coastal ocean to changing large-scale conditions, including such society relevant phenomena as the onset and development of coastal hypoxia and anoxia. Future work will require extensions both to larger scales, beyond the spatial limits of the present model domain, and to smaller scales, with more detailed representation and examination of the role of along-path turbulent mixing and diffusion on upwelling water parcels.

Acknowledgments

We thank the reviewers for their critical comments and suggestions to an earlier version of this manuscript. Comments, suggestions, and/or some guidance from Scott Springer, Sangil Kim, Andrey Koch, and Alexander Kurapov enriched this work and are thankfully acknowledged. Brief comments by Neil Banas and Parker MacCready were also useful. We also appreciate the help provided by Paul Turner in the preprocessing of the NCOM outputs, and by Rodrigo Durán in the recalculation of some advection experiments. This study was funded by the National Science Foundation (NSF) Science and Technology Center for Coastal Margin Observation and Prediction (CMOP), NSF Award 0424602. Mooring NH10 data were provided by Mike Kosro. RISE mooring data were provided by Ed Dever. The QuikSCAT product was available through the French Research Institute for Exploitation of the Sea (IFREMER) Web site (http://www.ifremer.fr/cersat/en/data/download/gridded/mwfqscat.htm). The altimeter product was produced by Ssalto/Duacs and distributed by Archiving, Validation, and Interpretation of Satellite Oceanographic data (AVISO), with support from Cnes.

REFERENCES

  • Allen, S. E., , and X. Durrieu de Madron, 2009: A review of the role of submarine canyons in deep-ocean exchange with the shelf. Ocean Sci., 5 , 607620.

    • Search Google Scholar
    • Export Citation
  • Banas, N. S., , P. MacCready, , and B. M. Hickey, 2009: The Columbia River plume as cross-shelf exporter and along-coast barrier. Cont. Shelf Res., 29 , 292301.

    • Search Google Scholar
    • Export Citation
  • Baptista, A. M., and Coauthors, 2005: A cross-scale model for 3D baroclinic circulation in estuary-plume-shelf systems: II. Application to the Columbia River. Cont. Shelf Res., 25 , 935972.

    • Search Google Scholar
    • Export Citation
  • Barron, C. N., , A. B. Kara, , P. J. Martin, , R. C. Rhodes, , and L. F. Smedstad, 2006: Formulation, implementation and examination of vertical coordinate choices in the Global Navy Coastal Ocean Model (NCOM). Ocean Modell., 11 , 347375.

    • Search Google Scholar
    • Export Citation
  • Barth, J. A., 2003: Anomalous southward advection during 2002 in the northern California Current: Evidence from Lagrangian surface drifters. Geophys. Res. Lett., 30 , 8024. doi:10.1029/2003GL017511.

    • Search Google Scholar
    • Export Citation
  • Barth, J. A., , S. D. Pierce, , and R. L. Smith, 2000: A separating coastal upwelling jet at Cape Blanco, Oregon and its connection to the California Current System. Deep-Sea Res. II, 47 , 783810.

    • Search Google Scholar
    • Export Citation
  • Barth, J. A., and Coauthors, 2007: Delayed upwelling alters nearshore coastal ocean ecosystems in the northern California Current. Proc. Natl. Acad. Sci. USA, 104 , 37193724.

    • Search Google Scholar
    • Export Citation
  • Beckmann, A., , and D. B. Haidvogel, 1993: Numerical simulation of flow around a tall isolated seamount. Part I: Problem formulation and model accuracy. J. Phys. Oceanogr., 23 , 17361753.

    • Search Google Scholar
    • Export Citation
  • Bosley, K. L., , J. W. Lavelle, , R. D. Brodeur, , W. W. Wakefield, , R. L. Emmett, , E. T. Baker, , and K. M. Rehmke, 2004: Biological and physical processes in and around Astoria submarine Canyon, Oregon, USA. J. Mar. Syst., 50 , 2137.

    • Search Google Scholar
    • Export Citation
  • Burk, S. D., , T. Haack, , and R. M. Samelson, 1999: Mesoscale simulation of supercritical, subcritical, and transcritical flow along coastal topography. J. Atmos. Sci., 56 , 27802795.

    • Search Google Scholar
    • Export Citation
  • Chan, F., , J. A. Barth, , J. Lubchenco, , A. Kirincich, , H. Weeks, , W. T. Peterson, , and B. A. Menge, 2008: Emergence of anoxia in the California Current large marine ecosystem. Science, 319 , 920.

    • Search Google Scholar
    • Export Citation
  • Chhak, K., , and E. DiLorenzo, 2007: Decadal variations in the California Current upwelling cells. Geophys. Res. Lett., 34 , L14604. doi:10.1029/2007GL030203.

    • Search Google Scholar
    • Export Citation
  • Crawford, W. R., , and R. K. Dewey, 1989: Turbulence and mixing: Sources of nutrients on the Vancouver Island continental shelf. Atmos.–Ocean, 27 , 428442.

    • Search Google Scholar
    • Export Citation
  • Freeland, H. J., , G. Gatien, , A. Huyer, , and R. L. Smith, 2003: Cold halocline in the northern California Current: An invasion of subarctic water. Geophys. Res. Lett., 30 , 1141. doi:10.1029/2002GLO16663.

    • Search Google Scholar
    • Export Citation
  • Garfield, N., , C. A. Collins, , R. G. Paquette, , and E. Carter, 1999: Lagrangian exploration of the California Undercurrent, 1992–95. J. Phys. Oceanogr., 29 , 560583.

    • Search Google Scholar
    • Export Citation
  • Gay, P. S., , and T. K. Chereskin, 2009: Mean structure and seasonal variability of the poleward undercurrent off southern California. J. Geophys. Res., 114 , C02007. doi:10.1029/2008JC004886.

    • Search Google Scholar
    • Export Citation
  • Gould, J., and Coauthors, 2004: Argo profiling floats bring new era of in situ ocean observations. Eos, Trans. Amer. Geophys. Union, 85 (179) 190191.

    • Search Google Scholar
    • Export Citation
  • Grantham, B. A., , F. Chan, , K. J. Nielsen, , D. S. Fox, , J. A. Barth, , A. Huyer, , J. Lubchenco, , and B. A. Menge, 2004: Upwelling-driven nearshore hypoxia signals ecosystem and oceanographic changes in the northeast Pacific. Nature, 429 , 749754.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., 1997: The response of a steep-sided, narrow canyon to time-variable wind forcing. J. Phys. Oceanogr., 27 , 697726.

  • Hickey, B. M., , and N. S. Banas, 2008: Why is the northern end of the California Current system so productive? Oceanography, 21 , 90107.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., , S. Geier, , N. Kachel, , and A. MacFadyen, 2005: A bi-directional river plume: The Columbia in summer. Cont. Shelf Res., 25 , 16311656.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., , A. MacFadyen, , W. Cochlan, , R. Kudela, , K. Bruland, , and C. Trick, 2006: Evolution of chemical, biological, and physical water properties in the northern California Current in 2005: Remote or local wind forcing? Geophys. Res. Lett., 33 , L22S02. doi:10.1029/2006GL026782.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., , R. McCabe, , S. Geier, , E. Dever, , and N. Kachel, 2009: Three interacting freshwater plumes in the northern California Current system. J. Geophys. Res., 114 , C00B03. doi:10.1029/2008JC004907.

    • Search Google Scholar
    • Export Citation
  • Hickey, B. M., and Coauthors, 2010: River influences on shelf ecosystems: Introduction and synthesis. J. Geophys. Res., 115 , C00B17. doi:10.1029/2009JC005452.

    • Search Google Scholar
    • Export Citation
  • Huyer, A., 1983: Coastal upwelling in the California Current system. Prog. Oceanogr., 12 , 259284.

  • Huyer, A., , J. H. Fleischbein, , J. Keister, , P. M. Kosro, , N. Perlin, , R. L. Smith, , and P. A. Wheeler, 2005: Two coastal upwelling domains in the northern California Current system. J. Mar. Res., 63 , 901929.

    • Search Google Scholar
    • Export Citation
  • Huyer, A., , P. A. Wheeler, , P. T. Strub, , R. L. Smith, , R. Letelier, , and P. M. Kosro, 2007: The Newport Line off Oregon—Studies in the north east Pacific. Prog. Oceanogr., 75 , 126160. doi:10.1016/j.pocean.2007.08.003.

    • Search Google Scholar
    • Export Citation
  • IFREMER, 2002: QuikSCAT scatterometer mean wind field products user manual. IFREMER C2-MUT-W-04-IF, 47 pp.

  • Kosro, P. M., 2003: Enhanced southward flow over the Oregon shelf in 2002: A conduit for subarctic water. Geophys. Res. Lett., 30 , 8023. doi:10.1029/2003GL017436.

    • Search Google Scholar
    • Export Citation
  • Kosro, P. M., , W. T. Peterson, , B. M. Hickey, , R. K. Shearman, , and S. D. Pierce, 2006: Physical versus biological spring transition: 2005. Geophys. Res. Lett., 33 , L22S03. doi:10.1029/2006GL027072.

    • Search Google Scholar
    • Export Citation
  • Kudela, R. M., , W. P. Cochlan, , T. D. Peterson, , and C. G. Trick, 2006: Impacts on phytoplankton biomass and productivity in the Pacific Northwest during the warm ocean conditions of 2005. Geophys. Res. Lett., 33 , L22S06. doi:10.1029/2006GL026772.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., , P. MacCready, , and B. M. Hickey, 2009a: Columbia River plume patterns in summer 2004 as revealed by a hindcast coastal ocean circulation model. Geophys. Res. Lett., 36 , L02601. doi:10.1029/2008GL036447.

    • Search Google Scholar
    • Export Citation
  • Liu, Y., , P. MacCready, , B. M. Hickey, , E. P. Dever, , P. M. Kosro, , and N. S. Banas, 2009b: Evaluation of a coastal ocean circulation model for the Columbia River plume in summer 2004. J. Geophys. Res., 114 , C00B04. doi:10.1029/2008JC004929.

    • Search Google Scholar
    • Export Citation
  • MacCready, P., , N. S. Banas, , B. M. Hickey, , E. P. Dever, , and Y. Liu, 2009: A model study of tide- and wind-induced mixing in the Columbia River estuary and plume. Cont. Shelf Res., 29 , 278291.

    • Search Google Scholar
    • Export Citation
  • MacFadyen, A., , B. M. Hickey, , and W. P. Cochlan, 2008: Influences of the Juan de Fuca Eddy on circulation, nutrients, and phytoplankton production in the northern California Current system. J. Geophys. Res., 113 , C08008. doi:10.1029/2007JC004412.

    • Search Google Scholar
    • Export Citation
  • Mackas, D. L., , K. L. Denman, , and A. F. Bennett, 1987: Least squares multiple tracer analysis of water mass composition. J. Geophys. Res., 92 , 29072918.

    • Search Google Scholar
    • Export Citation
  • Mackas, D. L., , W. T. Peterson, , M. D. Ohman, , and B. E. Lavaniegos, 2006: Zooplankton anomalies in the California Current system before and during the warm ocean conditions of 2005. Geophys. Res. Lett., 33 , L22S07. doi:10.1029/2006GL027930.

    • Search Google Scholar
    • Export Citation
  • Mellor, G. L., , and T. Yamada, 1982: Development of a turbulence closure model for geophysical fluid problems. Rev. Geophys. Space Phys., 20 , 851875.

    • Search Google Scholar
    • Export Citation
  • Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87 , 343360.

  • Penven, P., , P. Marchesiello, , L. Debreu, , and J. Lefèvre, 2008: Software tools for pre- and post-processing of oceanic regional simulations. Environ. Model. Softw., 22 , 117122.

    • Search Google Scholar
    • Export Citation
  • Perlin, N., , E. D. Skyllingstad, , R. M. Samelson, , and P. L. Barbour, 2007: Numerical simulations of air–sea coupling during coastal upwelling. J. Phys. Oceanogr., 37 , 20812093.

    • Search Google Scholar
    • Export Citation
  • Pierce, S. D., , R. L. Smith, , P. M. Kosro, , J. A. Barth, , and C. D. Wilson, 2000: Continuity of the poleward undercurrent along the eastern boundary of the mid-latitude North Pacific. Deep-Sea Res. II, 47 , 811829.

    • Search Google Scholar
    • Export Citation
  • Pierce, S. D., , J. A. Barth, , R. E. Thomas, , and G. W. Fleischer, 2006: Anomalous warm July 2005 in the northern California Current: Historical context and the significance of cumulative wind stress. Geophys. Res. Lett., 33 , L22S04. doi:10.1029/2006GL027149.

    • Search Google Scholar
    • Export Citation
  • Schwing, F. B., , N. A. Bond, , S. J. Bograd, , T. Mitchell, , M. A. Alexander, , and N. Mantua, 2006: Delayed coastal upwelling along the U.S. west coast in 2005: A historical perspective. Geophys. Res. Lett., 33 , L22S01. doi:10.1029/2006GL026911.

    • Search Google Scholar
    • Export Citation
  • Shchepetkin, A. F., , and J. C. McWilliams, 2003: A method for computing horizontal pressure-gradient force in an oceanic model with a nonaligned vertical coordinate. J. Geophys. Res., 108 , 3090. doi:10.1029/2001JC001047.

    • Search Google Scholar
    • Export Citation
  • Shchepetkin, A. F., , and J. C. McWilliams, 2005: The Regional Ocean Modeling System: A split-explicit, free-surface, topography following coordinates ocean model. Ocean Modell., 9 , 347404.

    • Search Google Scholar
    • Export Citation
  • Smith, S. D., 1988: Coefficients for sea surface wind stress, heat flux, and wind profiles as a function of wind speed and temperature. J. Geophys. Res., 93 , 1546715472.

    • Search Google Scholar
    • Export Citation
  • Smith, W. H. F., , and D. T. Sandwell, 1997: Global sea floor topography from satellite altimetry and ship depth soundings. Science, 277 , 19561962.

    • Search Google Scholar
    • Export Citation
  • Springer, S. R., , R. M. Samelson, , J. S. Allen, , G. D. Egbert, , A. L. Kurapov, , R. N. Miller, , and J. C. Kindle, 2009: A nested model of the Oregon Coastal Transition Zone: Simulations and comparisons with observations during the 2001 upwelling season. J. Geophys. Res., 114 , C02010. doi:10.1029/2008JC004863.

    • Search Google Scholar
    • Export Citation
  • Stegmann, P. M., , and F. Schwing, 2007: Demographics of mesoscale eddies in the California Current. Geophys. Res. Lett., 34 , L14602. doi:10.1029/2007GL029504.

    • Search Google Scholar
    • Export Citation
  • Strub, P. T., , and C. James, 2003: Altimeter estimates of anomalous transports into the northern California Current during 2000–2002. Geophys. Res. Lett., 30 , 8025. doi:10.1029/2003GL017513.

    • Search Google Scholar
    • Export Citation
  • Strub, P. T., , J. S. Allen, , A. Huyer, , R. L. Smith, , and R. C. Beardsley, 1987: Seasonal cycles of currents, temperatures, winds, and sea level over the northeast Pacific continental shelf: 35°N to 48°N. J. Geophys. Res., 92 , 15071526.

    • Search Google Scholar
    • Export Citation
  • Thomas, A., , and P. Brickley, 2006: Satellite measurements of chlorophyll distribution during spring 2005 in the California Current. Geophys. Res. Lett., 33 , L22S05. doi:10.1029/2006GL026588.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

Model domain and bathymetry (m). Locations of the Columbia River mouth (CR), South Beach (SB), Cape Blanco (CB), Crescent City Line (CresCy), NH Line, and NH10 mooring are indicated, along with the positions of the RISE moorings (Rc, Rs) and the NDBC buoy (gray triangle). The six analysis subdomains are also shown: the “upwelling region” (shaded, inshore of the 200-m isobath), 3 adjacent regions (NW, W, and SW), and 2 offshore regions (ON, OS). The plot is scaled so that the zonal and meridional distances are equal.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 2.
Fig. 2.

(a) Cumulative meridional wind stress vs time (months) for a year beginning on 1 January and averaged over the upwelling region in Fig. 1. Columbia River discharge (b) volume flux and (c) temperature. Values are shown for the year 2005 (black lines) and for the climatology (gray).

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 3.
Fig. 3.

Observed ADCP (obs) and model (mod) velocity vectors (cm s−1, eastward along positive time axis) at (a) 10 and (b) 68 m at the NH10 location (Fig. 1) vs time (months) for the year 2005. The scale for the model velocities, relative to the displaced 0 line, is the same at each depth as that for the corresponding observed velocities.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 4.
Fig. 4.

(a) SST at the position of NDBC buoy 46050 (see Fig. 1), and (b) adjusted sea level anomaly (relative to a mean annual cycle, represented by a single fitted annual harmonic) at South Beach (see Fig. 1) vs time (months) during 2005: observed (gray lines) and modeled values (black).

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 5.
Fig. 5.

Temporal means (May–October 2005) of (a),(c) SST (°C) and (b),(d) sea level anomaly (SLA, cm) from the (a),(b) model and (c),(d) observations. The model SLA shown in (b) is the difference from the annual mean.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 6.
Fig. 6.

Observed (thin lines, shorter time series) and model (thick lines, longer time series) for (top to bottom) temperature T, salinity S, and velocity (u, υ) at 5- and 20-m depths at the positions of moorings (a) Rc and (b) Rs (Fig. 1) vs time (months) for May–October 2005.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 7.
Fig. 7.

(a) Ekman volume transport (Sv, 1 Sv ≡ 106 m3 s−1) into the upwelling region through the western boundary (see Fig. 1) vs time (months) for the year 2005 (black line) and climatological (gray) simulations. Net volume transport (Sv) vs time (months) for the (b) year 2005 and (c) climatological simulations inward through the western (W, gray line), outward through the northern plus southern [−(N + S), thick black], and outward through the northern [(−N), thin black] boundary of the upwelling region shown in Fig. 1.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 8.
Fig. 8.

Model mean mid-July–September (top) near-surface and (bottom) near-bottom velocities (cm s−1; arrows proportional to magnitude) in and near the upwelling region (Fig. 1) for the (a) year 2005 and (b) climatological simulations. At each point, these near-surface and near-bottom velocities are vertical averages over the uppermost 80 m and the lowermost 50 m, respectively.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 9.
Fig. 9.

Model mean mid-July–September (left) meridional and (right) zonal velocity (cm s−1; 0 contour shown in black) vs longitude (°W) and depth (m) at the NH Line (Fig. 1) for the (a),(b) year 2005 and (c),(d) climatological simulations. The location near 124.7°W and 250-m depth of the model poleward undercurrent core, represented by the cross-sectional maximum northward velocity with the indicated mean and temporal standard deviation, is shown in (a) and (c). The 200-m isobath (gray line) defines the western boundary of the upwelling region (Fig. 1). Markers within this region indicate physical final particle positions for the NH Line trajectory analysis described in section 3b, with the symbols indicating the corresponding source regions. In (b) and (d), markers for source regions are plotted at final positions only for source regions supplying at least 10% of the corresponding particles at that position.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 10.
Fig. 10.

The 6-month trajectories from March through 29 Sep 2005 vs latitude (°N) and longitude (°W), for 201 particles with final positions on 29 Sep 2005 evenly distributed at 2-m depth across the upwelling region (Fig. 1). Trajectory color indicates (a) the final latitude (°N) for the corresponding particle and (b) the instantaneous particle depth (m) along the trajectory. The region shown covers the NW, W, SW, and upwelling subdomains defined in Fig. 1.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 11.
Fig. 11.

Mean paths for the (a),(b) year 2005 and (c),(d) climatological simulations, computed as described in the text, vs latitude (°N) and longitude (°W): (left) north and (right) south of Cape Blanco (CB). The means were computed from the 6231 particle paths discussed in section 3b, for which final positions were specified in August. Color indicates along-path depth (m). Mean paths are shown separately for particles upwelling north and south of CB. The subdomains defined in Fig. 1—according to which the trajectories are grouped for this analysis—are also shown along with the 200- and 1000-m isobaths. Numbers adjacent to the paths indicate the percentage of the particles represented by the corresponding mean path; only those paths originating outside the upwelling region and representing at least 1% of the total number of particles are shown so the percentages shown in (a),(b) and (c),(d) do not sum to 100.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 12.
Fig. 12.

(a)–(d) Trajectory statistics vs final position latitude (°N) and longitude (°W) for the 62% of particles with specified final positions in the upwelling region in August that originated outside of the model domain in the year 2005 simulation. Shown at each corresponding final position are (a) the percentage of particles at that location that originated outside of the model’s domain, (b) the boundary from which these particles originated (northern or southern; black if both), and the mean (c) depth (m) and (d) time (days prior to date of specified final position) at which the particles entered from the corresponding boundary. (e)–(h) As in (a)–(d), but for the climatological simulation.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 13.
Fig. 13.

(top) Trajectory statistics vs initial position (in February, 6 months prior to the corresponding final time in August) vs latitude (°N) and longitude (°W), for the 38% of particles with specified final positions in the upwelling region in August that originated offshore inside of the model domain in the year 2005 simulation. Shown at each corresponding initial position are the final latitude (°N, left) and initial depth (m, right) of particles at that initial location. (bottom) As in the top panels, but for the climatological simulation.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 14.
Fig. 14.

Histograms (%) of particle initial positions and depths, for particles with final positions (top) north and (next row) south of Cape Blanco, vs the analysis subdomain (Fig. 1), for the 38% of particles with specified final positions in August and initial positions offshore in the year 2005 simulation (as in Fig. 13). Depth bins are 50 m, and the inset numbers (right panels) are the total number of particles represented in each corresponding pair of histograms. (bottom four panels) As in the top panels, but for the climatological simulation.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 15.
Fig. 15.

(top) Upwelling displacement quantities for the 38% of particles with specified final positions in August and initial positions offshore in the year 2005 simulation (as in Fig. 13). For particles (left to right): number of paths and the corresponding mean rate of vertical motion for those that cross or are found at depths of 150 ± 5 m depth vs latitude (°N) and longitude (°W); numbers for approximately 14 km × 14 km cells that cross the indicated isobaths; and the numbers of paths that cross the indicated isobaths vs the final latitude (°N) of those crossing the 1000-m isobath at the given latitude and their corresponding depths (m). (bottom) As in the top panels, but for the climatological simulation.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 16.
Fig. 16.

Along-path mean (top to bottom) temperature (θp), salinity (Sp), and depth (zp) deviation vs time tp, where tp = 0 is the time at which the particle’s final position in the upwelling region is specified for all of the particles with specified final positions in August in the (a) year 2005 and (b) climatological simulations. Along-path means were computed separately for particles entering from the northern (blue) and southern (red) boundaries and for those originating within the model domain (black). The along-path means for particles originating outside the domain were computed for periods during which at least 50% of the corresponding sets of particles had entered the domain. The temperature, salinity, and depth deviations are defined relative to the initial path mean at the time step prior to the first plotted time. The numbers in the top panels indicate the number and percentage of particles in each of the three groups. Error bars correspond to the standard deviations computed every 6 days for each group, relative to the corresponding ensemble mean.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 17.
Fig. 17.

Along-path means computed similarly to the quantities in Fig. 16, but for diagnosed along-path values of the (top) heat and (bottom) salinity diffusion terms (κtTz)z (°C s−1) and (κsSz)z (psu s−1) where T and S are temperature and salinity, κt and κs are heat and salt diffusivities, and subscripts indicate derivatives with respect to the vertical coordinate z.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Fig. 18.
Fig. 18.

Monthly mean TS diagrams for the NH (gray dots) and Crescent City (black) Lines. For (top) year 2005 and (bottom) climatological simulations, during the period June–September and inshore of the 200-m isobath. Note the change in axis scales between the July and August diagrams. Contours of the potential density anomaly are also shown.

Citation: Journal of Physical Oceanography 41, 1; 10.1175/2010JPO4327.1

Save