**1. Introduction**

The coast of Louisiana in the northern Gulf of Mexico (NGOM) is characterized by semi-enclosed bays with exchange flows of water through multiple inlets, such as Lake Pontchartrain, Calcasieu Lake, Vermillion Bay, and Barataria Bay. They have limited connections with the coastal ocean except through narrow inlets. These are, however, different from inland freshwater lakes or general coastal plain estuaries connected to the coastal ocean through multiple inlets. The NGOM has several major environmental processes that are determined by hydrodynamics, particularly those related to the exchange of water and sediment between estuaries and shelf water, e.g., the significant land loss around lower Mississippi River basin associated with processes that cause erosion and sediment transport [1–6]. Along the NGOM coast, the most regular hydrodynamic motions are the relatively weak tides, which are mainly diurnal oscillations with a maximum tidal range of about 0.6 m [7,8]. Because of weak tides, the effect of weather [9–14] becomes more prominent in moving the sediment through inundation and erosion [15]. As a result, the less predictable weather-induced bay oscillations may cause more significant flood and drain of the micro-tidal system [16–20]. Synoptic weather systems and hurricanes can produce responses in these water bodies affecting the water exchange, which is important to the ecosystem [21–23]. However, there is a lack of in-depth analysis of weather conditions characterizing different weather patterns.

Among the weather influences, hurricane impact can be most dramatic. Because of the low gradient of land and relatively shallow and broad shelf along NGOM, hurricanes can cause significant damages to the micro-tidal coast zone [24–28]. For instance, severe storm surge caused damages in 2005 by Hurricanes Katrina and Rita [29–33], in 2008 by Hurricanes Gustav and Ike [34–37], and in 2017 by Hurricanes Harvey and Irma [38]. Compared with hurricanes, cold front-associated winds can be more frequent, and have a more accumulative e ffect in driving the hydrodynamics. The length of influence of a cold front can reach ~2000–3000 km, much larger [39] than the region of strong wind within a typical hurricane. Previous studies have covered various aspects of weather impact to coastal ocean. For example, Keen [40] used numerical models to predict the waves and currents under cold front passage over Mississippi bight. Keen and Stavn [41] later used observations and numerical models with interaction of atmospheric forcing and hydrodynamics to investigate the optical environment at Santa Rosa Island, Florida during two cold front passages. Water exchange and circulations in the bays and estuaries under meso-scale weather systems like winter storms and cold fronts can be related not only to the circulations in the coastal regions but also to the sediment transport [42] and ecosystems. Sediment transport and distributions on the shallow shelf and in the estuaries of Gulf of Mexico under the influence of cold front passages in winter time are investigated by, e.g., Perez et al. [43] and Kineke et al. [44]. Siadatmousavi et al. [45] studied the wave energy during a cold front and skill assessed a phase-averaged spectral wave model.

There have been many studies on subtidal flow in estuaries [46–48]. In these studies, the wind effects are often discussed as a time series forcing without examining the spatial structure of the weather systems. The subtidal energy in the estuarine circulations caused by cold fronts may be comparable if not larger than that of tides in the area [49]. A recent study [50] investigated the weather-induced exchange flows through multiple inlets of the Barataria Bay in a few months period in 2013, 2014, and 2015 with 51 atmospheric cold fronts passing the Louisiana coast. These events are apparently very common: an analysis [51] covering a period of 40 years identified more than 1600 atmospheric frontal events, with an average of ~41.2 ± 4.7 per year excluding the months between May and August for much weaker activities of this kind. However, no quantitative comparisons are made between the hydrodynamic responses induced by cold fronts and that from hurricanes. Therefore, it is of interest for a comparison between the hydrodynamic responses to these two di fferent weather systems with di fferent scales.

This work will use a calibrated three-dimensional finite volume community model (FVCOM) to simulate water level and flows in Barataria Bay under multi-scale weather systems including cold front and hurricane events. The goals are to (1) compare the hydrodynamic responses to di fferent weather systems (cold front and hurricane), (2) examine water exchange between Barataria Bay and coastal ocean through multiple inlets under the di fferent weather systems, and (3) assess the quasi-steady state balance under di fferent weather systems.

#### **2. Study Site and Data**

Barataria Bay (Figure 1) is a shallow estuary in southeast Louisiana and south of the City of New Orleans. It is bounded by several barrier islands and irregular-shaped wetlands with multiple tidal inlets connecting to the open ocean. The main axis from north to south and from east to west is about 30–40 km. The tidal inlets include Barataria Pass with a width of ~800 m and a maximum depth of 20 m at the mouth, Caminada Pass with ~800 m width, 9 m depth, and a 90 degree turn in channel orientation near the mouth, and the 15 m deep Pass Abel with a width of about 1.9 km. Freshwater is mainly from the manmade Davis Pond Diversion facility with a capacity of about 250 m<sup>3</sup>/s of flux. Water inside the Barataria Bay is very shallow (average depth of 2 m). Erosions in the bay appear to be significant, e.g., there is a 50 m hole [50] northwest of the Barataria Pass, which is the deepest point among all Louisiana lagoons, bays, and estuaries, revealing the significant contribution of non-tidal forcing to the micro-tidal system.

**Figure 1.** Study site (**b**) and model grid (**a**) for FVCOM simulation. Star represents the location of wind observations (NDBC station GISL1); red dot represents the location of ADCP deployment for water level and velocity observations.

Observational data of water level and velocity were obtained from 5 Sontek Argonaut DP SL 500-KHz horizontal acoustic Doppler current profilers (ADCPs, manufacture: SonTek/Xylem Inc., San Diego, USA). Information about the measurements can be found in Li et al. [50]. Wind data are from the National Ocean Service station at Grand Isle (29.265◦ N, 89.958◦ W, Figure 1). The atmospheric forcing for the upper boundary of the hydrodynamic numerical model was obtained from the global Climate Forecast System Reanalysis (CFSR) data (https://climatedataguide.ucar.edu/climate-data/ climate-forecast-system-reanalysis-cfsr).

#### **3. Model Setup and Validation**
