The wind-driven circulation over continental shelves has been an area of intensive research over the last three decades. This research has provided us with a much clearer understanding of the dynamics of the alongshelf component of the flow and particularly the role of coastal-trapped waves [e.g., Brink, 1991; Huyer, 1990]. However, the wind-driven cross-shelf circulation remains poorly understood. The cross-shelf flow is usually weaker than the alongshelf flow, however, it can be important in redistributing water properties, nutrients, sediments, pollutants and other constituents because cross-shelf variations in these tracers tend to be larger than alongshelf variations.
The classical conceptual picture of the wind-driven
cross-shelf circulation during coastal upwelling is shown in Figure 3.
There is a transport to the right (in the
northern hemisphere) of the alongshelf wind stress in the surface boundary
layer due to the Coriolis acceleration associated with the earth's
rotation [ Ekman, 1905]. In this simple two-dimensional
schematic there is a corresponding onshore transport in the interior and
bottom boundary layers. This cross-shelf circulation closes over the inner
shelf region where the surface and bottom boundary layers merge [e.g.,
Allen et al., 1995]. This is the circulation pattern
typically associated with coastal upwelling which brings nutrients from
the deeper water to the surface where they can be utilized by
phytoplankton. Studies of the volume heat budget in several coastal
regions are consistent with this simple conceptual picture [e.g.,
Dever and Lentz, 1994] in the sense that a wind-driven
offshore (onshore) flux of heat leads to a decrease (increase) in the
spatially-averaged temperature over the shelf.
Numerical modeling studies have shown that with stratification, more realistic topography, and complex mixing the simple conceptual picture in Figure 3 is still relevant, though the details of the flow are more complicated [e.g. Allen et al., 1995; Keen and Glenn, 1994; Zamudio and Lopez, 1994]. Recent efforts to simulate observed circulations using two-dimensional numerical models have had mixed success. Zamudio and Lopez [1994] included a time-variable alongshelf pressure gradient derived from coastal-trapped wave theory in the two-dimensional shelf circulation model of Chen and Wang [1990] to simulate observations taken over the northern California shelf. Federiuk and Allen [1995] used a two-dimensional numerical model to simulate the observed circulation over the Oregon shelf. In both studies, the model results reproduced the observed alongshelf velocity and its vertical structure, but were less successful at simulating the cross-shelf circulation, particularly below the surface boundary layer. Federiuk and Allen [1995] pointed out that the imperfect agreement between model results and observations may be due to the three-dimensionality of the flow field. Another factor which may contribute to the lack of agreement is the sensitivity of the cross-shelf circulation to the form of the mixing parameterization [ Allen et al., 1995; Lentz, 1995].
Besides simulating observations, numerical model
studies are providing
insight into how the cross-shelf circulation and the shelf dynamics may
depend on factors such as stratification and topography. In a study of
coastal upwelling using a two-dimensional numerical model,
Allen et al. [1995] found, for example, that more of the onshore return
flow was concentrated in the bottom boundary layer on a wide shelf than on
a steep, narrow shelf. Numerical modeling studies also can identify and
clarify important processes that may be poorly resolved by observations.
For example, Allen et al. [1995] found that a region of
strong cross-shelf gradients in velocity and density, i.e. a front, forms
in model simulations of coastal upwelling and that the dynamics of the
frontal region are complex. Their model results suggest the frontal region
is unstable and hence may be an important region of mixing. In another
example, Federiuk and Allen [1995] found that
near-inertial motions have a complex structure in simulations of upwelling
over the Oregon shelf with realistic bottom bathymetry and stratification.
These types of results may provide an important basis for interpreting
previous and future observations.
The remainder of this section briefly highlights some of the recent progress in understanding each of the four regions in the cross-shelf circulation shown in Figure 3: the surface boundary layer; the interior; the bottom boundary layer; and the inner shelf. Though the discussion focuses on each region in isolation, it should be kept in mind that they are in fact tightly coupled.