Wind-driven Coastal Generation of Annual Mesoscale Eddy Activity in the California Current

View More View Less
  • 1 Scrripps Institution of Oceanography, University of California, San Diego, La Jolla, California
  • | 2 Centro de Investigació Científica y Educación superior de Ensenada
  • | 3 NOAA, NOS, Rockville, Maryland
© Get Permissions Rent on DeepDyve
Restricted access

Abstract

Two candidate sources for the generation of mesoscale eddy activity in the California Current are local baroclinic instability and/or the wind stress adjacent to the coast. The latter constitutes remote forcing, with eddy activity propagating westward from the coast into the California Current via Rossby wave dynamics. In this study, two wind-driven models are utilized to test the relative significance of these two sources. One is an eddy resolving quasigeostrophic (QG) model, with the ability to represent baroclinic instability but not the coastal response to winds. The other is a 1½-layer primitive equation (PE) model with the ability to represent the coastal response to winds but not baroclinic instability. Both models have the same spatial grid (i.e., approximately 20 km) and are driven by the same coarse-grid wind-stress forcing fields over the same one-year time period (i.e., January 1987 to December 1987). This period is chosen because of the availability of Geosat altimetric sea-level observations with which to verify these models. Earlier, White and colleagues analyzed these same altimetric sea-level observations, finding dominant mesoscale eddy activity occurring on wavelength scales of 400–800 km and period scales of 6–12 months. This mesoscale variability propagates to the west at 2–5 cm s−1, faster at lower latitude, consistent with Rossby wave dynamics. Moreover, the eddy variance is largest next to the coast, maximum between Monterey and Cape Mendocino, and southwest of Point Conception. The PE model is able to simulate qualitatively this distribution of the eddy variance as it appears in altimetric sea level, yielding significant coherence and phase between model and observed sea-level residuals along 1ongitude/time matrices at 30°N and 40°N. The QG model, on the other hand, is found incapable simulating the main features of this distribution of eddy variance. The reason for the agreement between the PE model and the satellite altimetric sea-level observations is that the dominant source of mesoscale eddy activity on these time and space scales is the wind forcing adjacent to the coast, modified by both Rossby and Kelvin wave dynamics.

Abstract

Two candidate sources for the generation of mesoscale eddy activity in the California Current are local baroclinic instability and/or the wind stress adjacent to the coast. The latter constitutes remote forcing, with eddy activity propagating westward from the coast into the California Current via Rossby wave dynamics. In this study, two wind-driven models are utilized to test the relative significance of these two sources. One is an eddy resolving quasigeostrophic (QG) model, with the ability to represent baroclinic instability but not the coastal response to winds. The other is a 1½-layer primitive equation (PE) model with the ability to represent the coastal response to winds but not baroclinic instability. Both models have the same spatial grid (i.e., approximately 20 km) and are driven by the same coarse-grid wind-stress forcing fields over the same one-year time period (i.e., January 1987 to December 1987). This period is chosen because of the availability of Geosat altimetric sea-level observations with which to verify these models. Earlier, White and colleagues analyzed these same altimetric sea-level observations, finding dominant mesoscale eddy activity occurring on wavelength scales of 400–800 km and period scales of 6–12 months. This mesoscale variability propagates to the west at 2–5 cm s−1, faster at lower latitude, consistent with Rossby wave dynamics. Moreover, the eddy variance is largest next to the coast, maximum between Monterey and Cape Mendocino, and southwest of Point Conception. The PE model is able to simulate qualitatively this distribution of the eddy variance as it appears in altimetric sea level, yielding significant coherence and phase between model and observed sea-level residuals along 1ongitude/time matrices at 30°N and 40°N. The QG model, on the other hand, is found incapable simulating the main features of this distribution of eddy variance. The reason for the agreement between the PE model and the satellite altimetric sea-level observations is that the dominant source of mesoscale eddy activity on these time and space scales is the wind forcing adjacent to the coast, modified by both Rossby and Kelvin wave dynamics.

Save