Simple linear models, which do a good job of representing the observed
characteristics of the alongshelf flow, fail to reproduce the
characteristics of the interior cross-shelf flow [e.g.,
Chapman, 1987; Brink et al., 1987]. The models suggest
the cross-shelf flow should be weaker and have larger correlation scales
than are observed. In fact, correlation scales for observed cross-shelf
currents are short, typically less than the separation between moorings
(
10--30 km). The reason for these short correlation scales is not
known. Brink et al. [1994] hypothesized that the short
correlation scales in the interior cross-shelf velocity were due to short
scales in the wind field. They tested this hypothesis using a linear
model and found that inclusion of shorter scale variability in the wind
field still yielded cross-shelf currents that were an order of magnitude
smaller than those observed. Several other possible explanations exist,
including the impact of topographic irregularities and the impingement of
offshore flows onto the shelf (see Section 2).