The Impact of Horizontal Sea Surface Temperature Gradients on Long Island Sound Sea Breezes

Eric Sinsky1, Kelly Lombardo1, James B Edson1 and Michael M Whitney2, (1)University of Connecticut, Marine Sciences, Groton, CT, United States, (2)University of Connecticut, Department of Marine Sciences, Groton, CT, United States
Abstract:
Sea breeze is a mesoscale process that requires accurate estimates of land-sea temperature gradients for realistic forecasts. These forecasts are of interest to the oceanography community due to its influence on upwelling and estuarine circulation. Accurate representation of the SST is an important component to these forecasts. Specifically, the importance of properly represented horizontal SST gradients offshore of complex coastlines should be emphasized. A more realistic numerical representation of the horizontal SST field is expected to improve sea breeze forecasts. This work explores how horizontal SST gradients impact the dynamics of sea breeze circulations.
The 8 July and 21 August 2013 sea breeze events are simulated using the Weather and Research Forecasting (WRF) model in the Long Island Sound (LIS) region. All simulations are initialized using the 32km North American Regional Reanalysis (NARR) for atmospheric conditions. To illustrate the impact of SST resolution on sea breeze forecasts, sensitivity experiments are performed varying the SST product used to initialize the model. The control experiment uses a spatially uniform SST of 22˚C. This will be compared to simulations initialized with the NARR 32km spatially varying SST (moderate SST gradient) and those initialized with the 1km G1SST (strong SST gradient).
Comparing the moderate SST gradient experiment to the control during the 21 August event, the SST is 0.75˚C cooler in eastern LIS and 0.75˚C warmer in western LIS. This creates a stronger and weaker land-sea 2m temperature gradient in eastern and western LIS, respectively, compared to the control. As a result, the sea breeze front in the moderate SST experiment propagates more slowly inland in the western part of the domain and more quickly inland in the eastern part of the domain. Similar patterns are observed comparing the strong SST gradient experiment to the control, though the deviations from the control are greater.