ABSTRACT

1
The Structure of Sea Breezes Over New Jersey
Coastal Area H. Pan, R. Avissar, S. M. Glenn
and D. A. Haidvogel Center for Environmental
Prediction and Institute of Marine and Coastal
Sciences, Rutgers University, New Brunswick, New
Jersey, USA.
ABSTRACT Irregular coast lines and terrain,
heterogeneous land cover, and sea surface
temperature (SST) affect the structure and
intensity of sea breezes in complicated ways. Due
mostly to limits in computational resources,
operational weather forecasting models do not
resolve mesoscale atmospheric processes like sea
breeze very well. Ocean models forced with the
forecasts that do not correctly reproduce these
processes may not be able to correctly simulate
observed marine coastal flows. In July 1999, the
Regional Atmospheric Modeling System (RAMS) was
used operationally, with a grid size of 4 x 4
km2, to forecast coastal weather in the northeast
US. Using an extended network of observations
(including buoys, radars, sodars, and
meteorological towers), we demonstrate that the
RAMS is able to simulate the detailed features of
the sea breezes that develop on the New Jersey
coast. We find that sea breezes in this region
can be decoupled from synoptic winds. When
southerly winds prevail, significant positive
wind curls form above the ocean near the shore.
On the other hand, northerly winds result in
negative wind curls. The intensity and the
spatial scale of these wind curls depend on the
magnitude of the synoptic wind and the sea
breezes. The maximum spatial extent of these
curls can be as large as 50 km off-shore. These
wind curls can have a significant impact on
coastal ocean processes, especially coastal
flows.
CONFIGURATION OF ATMOSPHERIC MODULE The
grid-nesting technique was adopted for the
atmospheric module, and the RAMS domains,
together with the domain of oceanic module, are
shown in Fig. 1. The fine grid of 4 x 4 km2
resolution is nested in the coarse grid of 16 x
16 km2 resolution, and the LEO-15 and CPS field
areas are located near the centers of both
atmospheric domains. Major options for the
atmospheric module are listed as follows

• 3D, non-hydrostatic, compressible dynamic
  equations, thermodynamic equations, and a set of
  microphysical equations
• 2-way interactive grid nesting
• Initial Condition (NCEP Eta reanalysis field)
• Boundary conditions
• Lateral (NCEP Eta and AVN)
• Top (rigid lid with Raleigh friction absorption
  layer)
• Bottom (over land Kessler and Tremback, and
  Avissar and Pielke parameterizations over water
  prescribed most recent SST)
• Turbulence (Mellor-Yamada)
• Microphysics (Kuo, and Tremback
  parameterizations)
• Radiation (Chen and Cotton)

INTRODUCTION In 1999, a Coastal Predictive
Skill (CPS) experiment was conducted during the
summer at the Rutgers University Long-term
Ecosystem Observatory (LEO-15) in the New York
Bight to evaluate a relocatable,
data-assimilative, coastal ocean forecast model
coupled to a multi-platform, adaptive sampling
network in a data- and process-rich environment
(CPS field area shown in Fig.1). The
a)
c)
b)
d)
Fig.2 Location of meteorological sensors (a) and
model/data comparisons at (b) Long Island Buoy
(c) Tuckerton and (d) Delaware Bay Buoy.
Fig. 1 Domains of atmospheric and oceanic
modules.
RESULTS AND DISCUSSION Comparisons between the
different model simulations and observations
shows that the high-resolution RAMS module is
better at reproducing the surface air
temperature, wind speed and wind direction most
of the time (Fig. 2). During July 1999, the
dominant winds are alongshore (southerly) with an
obvious diurnal onshore and offshore variation.
These variations are mostly due to the influence
of sea breezes. The RAMS simulation clearly
demonstrates these variations, but Eta and COAMPS
forecasts, which are output only every 6 hours,
poorly reflect this feature.
coastal forecasting modeling system consists of
two modules, the Regional Atmospheric Modeling
System (RAMS) (Pielke et al., 1992) and the
Regional Oceanic Modeling System (ROMS)
(Haidvogel et al., 2000), which are coupled
together at the air-sea interface. By using the
latest parallel versions of these two modules,
one-way coupling was realized in July 1999 with
the atmospheric module passing the surface
meteorological fields to the oceanic module every
30 minutes. A multilayer soil model was used over
land, and sea surface temperature was prescribed
using satellite imagery over water. The
high-resolution atmospheric module makes it
possible to simulate many meso-scale atmospheric
processes that are important for coastal oceanic
processes but poorly or hardly resolved by other
weather products. One of these processes is the
sea breeze, a prominent feature in the summer
climatology in the New Jersey / New York area.
Fig.3 RAMS simulated surface wind and air
temperature at (top to bottom) 06, 12, 18, and
24 UT on July 18, 1999.