Editorial Type:
Article
Article Type:
Research Article

Dynamics of the Block Island Sound Estuarine Plume

Qianqian Liu Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island

Search for other papers by Qianqian Liu in
Current site
Google Scholar
PubMed Close
,
Lewis M. Rothstein Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island

Search for other papers by Lewis M. Rothstein in
Current site
Google Scholar
PubMed Close
, and
Yiyong Luo Physical Oceanography Laboratory, Ocean University of China, Qingdao, China

Search for other papers by Yiyong Luo in
Current site
Google Scholar
PubMed Close
Print Publication:
01 May 2016
DOI:
https://doi.org/10.1175/JPO-D-15-0099.1
Page(s):
1633–1656
Full access

Abstract

Buoyant discharge of freshwater from Long Island Sound (LIS) forms a seasonal buoyant plume outside Block Island Sound (BIS) between the coast of Long Island and the denser shelf waters. The plume’s seasonal variability and its response to tides, winds, and surface heating are investigated through a series of process-oriented experiments using the Regional Ocean Modeling System (ROMS). Results show the importance of river discharge, wind directions, and surface heating in the seasonal variation of the BIS buoyant plume. In winter and spring, the plume is intermediate with a large surface offshore extension detached from the bottom. From winter to spring, the river discharge increases; meanwhile, upwelling-favorable winds keep dominating. They compete with the increase of surface heating and generate a broader buoyant plume in spring than in winter. In summer, the plume is bottom advected with most of its width in contact with the bottom and is featured with the steepest isopycnals and narrowest plume, which is driven by a combination of strong insolation, weak buoyant discharge from LIS, and feeble winds. In fall, although the river discharge is comparable to that in winter, the upwelling-favorable wind is relatively weaker, corresponding to a narrower intermediate plume.

Current affiliation: School of Marine Sciences, University of Maine, Orono, Maine.

Corresponding author address: Qianqian Liu, Graduate School of Oceanography, University of Rhode Island, 215 South Ferry Road, Narragansett, RI 02882. E-mail: liuqianqian0622@gmail.com

Abstract

Buoyant discharge of freshwater from Long Island Sound (LIS) forms a seasonal buoyant plume outside Block Island Sound (BIS) between the coast of Long Island and the denser shelf waters. The plume’s seasonal variability and its response to tides, winds, and surface heating are investigated through a series of process-oriented experiments using the Regional Ocean Modeling System (ROMS). Results show the importance of river discharge, wind directions, and surface heating in the seasonal variation of the BIS buoyant plume. In winter and spring, the plume is intermediate with a large surface offshore extension detached from the bottom. From winter to spring, the river discharge increases; meanwhile, upwelling-favorable winds keep dominating. They compete with the increase of surface heating and generate a broader buoyant plume in spring than in winter. In summer, the plume is bottom advected with most of its width in contact with the bottom and is featured with the steepest isopycnals and narrowest plume, which is driven by a combination of strong insolation, weak buoyant discharge from LIS, and feeble winds. In fall, although the river discharge is comparable to that in winter, the upwelling-favorable wind is relatively weaker, corresponding to a narrower intermediate plume.

Current affiliation: School of Marine Sciences, University of Maine, Orono, Maine.

Corresponding author address: Qianqian Liu, Graduate School of Oceanography, University of Rhode Island, 215 South Ferry Road, Narragansett, RI 02882. E-mail: liuqianqian0622@gmail.com
Keywords: Geographic location/entity; Estuaries; Circulation/ Dynamics; Buoyancy; Coastal flows; Fronts; Models and modeling; Numerical analysis/modeling; Variability; Seasonal variability

1. Introduction

Block Island Sound (BIS) is a strait on the inner shelf of the southern New England shelf. It separates Block Island (BI) from the coast of Rhode Island and connects Long Island Sound (LIS) and Rhode Island Sound (RIS; Fig. 1), working as the most important passage of freshwater that originates from the Connecticut River.

Fig. 1.
Fig. 1.

Bathymetry and geographic features around the Rhode Island Sound. Place names: Connecticut River (CR), Long Island Sound (LIS), Rhode Island Sound (RIS), Block Island Sound (BIS), Narragansett Bay (NB), Buzzards Bay (BB), and Vineyard Sound (VS). The thick red line is the section used to describe the buoyant plume to the southwest of BIS, and the thick black lines are the sections referenced in Table 2, as discussed further in section 5.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

According to Yankovsky and Chapman (1997), buoyant plumes generated by relatively fresh waters and offshore denser waters are characterized by advective processes that generate three dynamically distinctive plumes: 1) a surface-advected plume that is isolated from any interaction with the bottom, 2) a bottom-advected plume that is controlled by advection in the bottom boundary layer (Lentz and Helfrich 2002), and 3) an intermediate plume whose dynamics are a combination of 1 and 2. The historical development of estuary plume models are well summarized by Horner-Devine et al. (2015) and O’Donnell (2010). The Connecticut River, as the purest surface-advected plume that can be found (Garvine 1974), takes ~70% of the buoyancy (i.e., freshwater) source from LIS and entrains that out of BIS (Latimer et al. 2014; O’Donnell et al. 2014); during that process a downshelf buoyant coastal current is generated along the southern shore of Long Island, and a bottom-advected plume front is generated to the south of BIS.

The BIS bottom-advected plume was observed by Kirincich and Hebert (2005) during a 2-day experiment in April 2002, when the river discharge was strongest (Fig. 2). It is accompanied by an along-shelf coastal jet, which is almost linearly sheared with reversed velocities at the bottom and is in thermal wind balance with the mean density filed (Yankovsky and Chapman 1997). In light of the strong dependence of alongshore velocities on the plume and its front, Codiga (2005) used the velocity field between December 1999 and August 2002 to identify the location of the front and found that the plume experiences strong seasonal shifts, involving the shallowest attachment depth in winter, deepest in spring, and intermediate in fall, as well as a wider plume width in spring than in winter and fall. Based on a series of numerical simulations, Edwards et al. (2004a,b) studied the front in BIS and concluded the downshelf jet is a combination of tide-induced flow (nearshore) and buoyancy-driven flow (offshore), whose position can be modified by the change of buoyancy forcing.

Fig. 2.
Fig. 2.

Wind and Long Island Sound river input. (a) Daily winds from NARR averaged from 2004 to 2009. (b) The corresponding along-shelf (blue) and across-shelf (green) wind components. For the along-shelf component, positive value represents upwelling-favorable wind, while negative represents downwelling-favorable wind. For the across-shelf component, positive is for onshore wind and negative is for offshore wind. (c) Daily and monthly Long Island Sound river input, estimated from the Connecticut River discharge by dividing the Connecticut River by 0.7 (from USGS).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

The studies by both Kirincich and Hebert (2005) and Codiga (2005) pointed out the importance of the highly variable alongshelf (defined as the direction parallel to the southern coast of Long Island) winds (blue line in Fig. 2b) in the surface offshore extension of the plume due to Ekman dynamics, which is consistent with the available literature (Csanady 1978; Whitney and Garvine 2006; Moffat and Lentz 2012). In addition, modeling studies by Tilburg (2003) suggested the importance of winds in the across-shelf direction (the direction normal to Long Island), pointing out that an offshore wind stress can generate surface offshore transport and an onshore returning flow below it, while an onshore wind can generate onshore surface transport and offshore near-bottom transport. During fall and winter, BIS is subjected to strong offshore and upwelling winds; in summer, the winds become weaker and there are episodic occurrences of weak downwelling winds in September while the across-shelf direction is dominated by weak onshore winds. Based on the analysis of ferry-based current observations and numerical simulations, Whitney and Codiga (2011) found that the along-estuary winds are important for the water transport at the mouth of LIS.

It is very common for a plume to encounter an ambient current from upstream. The BIS estuarine plume is affected by the ambient current from RIS. However, most of the available studies have ignored the influence of ambient currents that have been proved important by Chapman and Lentz (1994). Chapman and Lentz (1994) found that this kind of flow limits the offshore spread of a plume and even pushes the front shoreward. Therefore, we believe that the seasonal cyclonic circulation in RIS (Kincaid et al. 2003; Luo et al. 2013) not only exchanges waters between RIS and BIS but also limits the offshore spread of the BIS estuarine plume. Over Rhode Island coastal waters, this upstream current is closely connected with a bottom thermal gradient, which is associated with surface heating (Luo et al. 2013). Hence, the surface heating may modulate the plume by means of both an upstream current and a bottom thermal gradient.

This study will examine the seasonal variability of the plume to the southwest of BIS and its response to seasonally varying winds, river discharge, and surface heating by means of seasonal upstream current and deep thermohaline gradients. The numerical experiment design and model–observation comparisons are described in sections 2 and 3. Model results and discussion are presented in sections 4 and 5. We conclude with a summary in section 6.

2. Design of numerical experiments

The numerical model used for this study is the Regional Ocean Modeling System (ROMS), a widely recognized regional and basin-scale ocean model using a high-resolution, free-surface, terrain-following coordinate (Shchepetkin and McWilliams 1998, 2003, 2005; Haidvogel et al. 2000). Our configuration, covering the domain of RIS, BIS, LIS, and the adjacent inner shelf area, is one-way nested within a larger domain covering the Gulf of Maine/Georges Bank and the New England shelf (Fig. 3; Luo et al. 2013). It has a horizontal resolution varying from 600 m over the RIS and BIS to 1 km along the boundaries (Fig. 3) with the number of vertical layers increased from 15 in Luo et al. (2013) to 30 in order to better capture both surface and bottom boundary layers.

Fig. 3.
Fig. 3.

Model configurations. The local-scale ROMS grid (plot every eight grid points) varies in horizontal resolution from 600 m over the RIS and BIS to ~1 km along the boundaries, and the regional-scale ROMS grid (plot every four grid points) is uniform with a resolution of 5 km.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

A series of experiments are executed to examine the response of the estuarine plume, sourced from the Connecticut River, to tides, wind directions, and surface heating as well as their seasonal variability under realistic atmospheric forcings (Table 1). The first experiment, named Buoy, isolates buoyant discharge from other forcings. The model starts from a resting ocean with uniform salinity of 35 psu and temperature of 14°C and is driven by a constant Connecticut River point source with an input of 1000 m3 s−1. According to O’Donnell et al. (2014), the long-term mean discharge obtained from the U.S. Geological Survey (USGS; http://waterdata.usgs.gov/nwis) in Thompsonville, Connecticut, is about 550 m3 s−1. To account for the ungauged discharge, we multiply the total gauged flow by the ratio of total area to gauged area. For the Connecticut River, the total gauged area is ~11 224 square miles (at USGS site 01194796 in Old Lyme, Connecticut, a site at the mouth of the Connecticut River), which is 1.16 times of the drainage area (~9660 square miles) at site 01184000 in Thompsonville, Connecticut. This gives the freshwater discharge from the Connecticut River as 638 m3 s−1. Therefore, the total freshwater discharge into LIS is ~911 m3 s−1, close to the value used in the idealized experiments (1000 m3 s−1).

Table 1.

List of Experiments with ROMS.

Table 1.

To look at the influence of tides, the second experiment RivTides is carried out. It is driven by the five major regional tidal components (M2, N2, S2, O1, and K1) as obtained from the Advanced Circulation Model for Oceanic, Coastal and Estuarine Waters (ADCIRC; Luettich et al. 1992) tidal simulation and the constant Connecticut River of 1000 m3 s−1. Similar to Buoy, this experiment also starts from a resting ocean with salinity of 35 psu and temperature of 14°C.

The realistic experiment (Real) integrates all available, real forcings and is used to understand the seasonal variations of the plume. In addition to tides, the following are used: daily averaged, climatological atmospheric forcings from the North American Regional Reanalysis (NARR; Mesinger et al. 2006) from 2004 to 2009; daily averaged, climatological river discharges for the Taunton River, Blackstone River, Pawtuxet River, and Connecticut River as obtained from the USGS; and open boundary conditions from our “regional” ROMS domain (from Luo et al. 2013). The case Real is driven by climatological daily averaged data from 2004 and 2009 for 3 yr with the first 2 yr used for spinup and the third-year results used to analyze the seasonal variability of the plume. For verification purpose, we have carried out another similar experiment driven by original, daily NARR forcing from 2004 to 2009. Since there is minor difference in monthly mean currents and plume between these two experiments, and the dynamics behind it is the same, to save computation time, the experiment driven by climatological mean of forcing is used.

To better represent the salinity input from the Connecticut River, we rescale the discharges from the USGS to account for the ungauged portions of the watershed (Chant et al. 2008; Zhang et al. 2010). Because the Connecticut River takes 70% of the riverine input entering LIS, the total discharge into LIS is estimated by dividing the Connecticut River by 0.7 (Fig. 2c). Please note, in this case, the Connecticut River has salinity of 0 psu and the same temperature with the closest water grid. Moreover, for simplicity, all rivers are idealized as point sources at the end of the long channels away from the water points.

In addition to experiment Real, we perform another experiment to examine the winds’ contributions to the seasonal variations of the BIS estuarine plume, that is, experiment NoWind omitting local winds from experiment Real. To test the response of the plume to the four typical winds in the studied region, that is, onshore wind (wind perpendicular to the south coast of Long Island and pointing onshore), offshore wind (similar to onshore wind but pointing offshore), upwelling-favorable wind (wind parallel to the south coast of Long Island and pointing northeast), and downwelling-favorable wind (similar to upwelling-favorable wind but pointing southwest), a series of idealized experiments driven by buoyant discharge, tides, and the corresponding winds are implemented. Each case is restarted from the sixth month of RivTides and then run for another 2 months (Table 1). These experiments are referred to as WndOn, WndOff, WndUp, and WndDown, respectively, in which we set the winds constant as 0.05 N m−2. The magnitudes represent the typical wind stresses in the upwelling and offshore directions in winter and fall.

An experiment with halved, constant river discharge from the Connecticut River, referred to as RivTides_500, is compared with RivTides to examine the river discharge’s impact on the buoyant plume. For the study of the plume’s response to surface heat flux, we construct another pair of experiments by adding constant surface heating from the sixth month of RivTides and run for another 2 months. An experiment, HeatS, driven by the climatological surface heat flux in August is carried out, and in the other experiment, HeatW, we halve the heat flux to provide weaker surface heating, whereby the bottom thermal front still exists but becomes weaker.

3. Model–observation comparisons

Luo et al. (2013) have shown that our realistically forced model of the region does an excellent job of simulating tidal residual currents over the whole domain as well as temperature and salinity at the outer shelf. A depth-averaged currents comparison between the model and observations (Fig. 4) demonstrates the model’s fidelity in RIS by capturing the cyclonic circulation in RIS. The moored observations are a combination of data from the Rhode Island Ocean Special Area Management Plan project and the project funded by Rhode Island Sea Grant and Rhode Island Endeavor Program (C. Wertman et al. 2016, unpublished manuscript) during 2009 and 2010. We average the observations in each season to compare with the modeling results. The average wind from 2004 to 2009 (Fig. 2) is generally southeastward in winter when it is strongest and northeastward in summer. Its seasonal variation is consistent with the data at the deployments carried out by the Ocean Special Area Management Plan (OSAMP) in September 2009, December 2009, March 2010, and June 2010 (Ullman and Codiga 2010). Therefore, the model is under similar conditions with the observations.

Fig. 4.
Fig. 4.

Comparison of depth-averaged currents in RIS in spring and summer between the model (blue arrows) and observations (red arrows). The field data are from Ullman and Codiga (2010).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

In May, both the model and observations show relatively weak (~2 cm s−1) onshore velocities in RIS, with stronger currents (~8 cm s−1) located to the south of Block Island in a region that connects RIS and BIS. In July, the model captures the westward jet offshore of Narragansett Bay as well as the southwestward current leaving RIS that appears in the observations. The root-mean-square error (RMSE) is used as a measure of precision. In May, the RMSEs for u- and υ-component velocities are 1.7 and 1.1 cm s−1, respectively, and in July, they are 2.1 and 1.1 cm s−1, respectively. Therefore, the model well captures the cyclonic circulation at the center and south of RIS in spring and summer, which is proved important to the buoyant plume in BIS in the following discussions.

In addition, Fig. 5 shows that our Real experiment simulates the circulation in BIS quite well by comparing its simulated depth-averaged currents in four seasons with observations from the Front-Resolving Observational Network with Telemetry (FRONT) project (Codiga 2005; Codiga and Houk 2002). Specifically, in spring, the depth-averaged flow features a robust (~15 cm s−1) southwestward current, which is consistent with Codiga’s (2005) observations with an RMSE for the u-component velocity of 2.6 cm s−1 and for the υ-component velocity of 4.6 cm s−1; in summer, our modeling results also agree well with the limited observations. In winter and fall, the model depth-averaged currents are significantly diminished with more variable directions, which may be due to the highly varying winds. The model results show a consistency with the fields in amplitude but only capture about half of the stations in direction. In light of the highly varying wind directions, the comparisons are acceptable. The success in tidal and realistic simulations as well as our previously published work (Luo et al. 2013) indicate that our model successfully represents the essential physical dynamics of RIS and BIS.

Fig. 5.
Fig. 5.

Comparison of depth-averaged currents in four seasons between the model (blue arrows) and observations (red arrows) in Front-Resolving Observational Network with Telemetry project between 2000 and 2001 (Codiga 2005; Codiga and Houk 2002; red arrows).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

To validate the experiment in simulating the river discharge from LIS, we compare velocities across a transect close to the Race (marked as section 1 in Fig. 1) between the case Real and the ADCP observations collected between November 2002 and January 2005 by the Cross Sound Ferry Services of New London, Connecticut (Codiga and Aurin 2007). Both observations and model results show vigorous water exchange across the channel with a surface-intensified, fresher current, ~15 cm s−1, moving out of the estuary along the south of the channel and bottom-intensified, saltier flow moving into the estuary along the northern boundary (Fig. 6), as a result of a gravitational circulation.

Fig. 6.
Fig. 6.

Comparison of velocities across a transect close to the Race (from New London, Connecticut, to Orient Point, New York) between the (a) observation and (b) model. Color represents velocities across the Race with positive values for current moving out of LIS and negative value into LIS. Model-simulated salinity along the transect is shown by black contours on the right panel.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Based on the available observations, mainly the better-sampled inward velocities, Codiga and Aurin (2007) estimated the annual-mean exchange transport as 2.27 × 104 ± 5000 m3 s−1. In the case Real, the annual-mean inward and outward volume transports are 2.51 × 104 m3 s−1 and 2.76 × 104 m3 s−1, respectively, which is in good agreement with the estimation by Codiga and Aurin (2007). In addition, consistent with the observations, the exchange transport undergoes strong seasonal fluctuation with the bottom, inward transport of 2.9 × 104 m3 s−1 in summer and 2.4 × 104 m3 s−1 in winter. The comparison suggests that it is reasonable to treat the rivers in LIS as emanating from the Connecticut River region alone and shows that the advection by residual flow is important for the total exchange between LIS and coastal waters.

To further verify the model in simulating the BIS estuarine plume, we compare the vertical currents in spring with the observations in the FRONT project (Codiga 2005). As shown in Fig. 7, consistent with Codiga’s (2005) observations, the plume features southwestward currents with robust, surface-intensified, along-shelf transport and with the vertical currents veering clockwise with depth for waters deeper than 10 m. The vertical, clockwise-veering structure is due to the decrease of downshelf and offshore velocities with depth. The model and observations have relatively larger discrepancies for velocities above 10 m, which may be due to the extrapolation of measurements. Near the bottom, both the model result and observations show a weak onshore transport even when winds are not strong enough to affect the bottom waters in spring. In brief, the above comparisons indicate our simulation captures the dynamics of the BIS estuarine plume.

Fig. 7.
Fig. 7.

Comparison of vertical currents in spring between (a) the observations in the FRONT project (Codiga 2005) and (b) model results. Color represents water depth with red representing currents at the top layers and blue at the deep, and arrows represent velocities at the corresponding depth.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

4. Results

a. Plume response to tides

Figure 8 depicts the pattern of the surface buoyant plumes after 100 days’ simulation for the idealized experiments of Buoy and RivTides. The case Buoy, an idealized experiment free of impacts of tides and winds, generates a plume extending broadly to the outer shelf region (Fig. 8a). The plume is steered to the right by the Coriolis force and features a relatively large anticyclonic circulation or “bulge” after it moves out of LIS and BIS, which was discussed by Chant et al. (2008), Kourafalou et al. (1996), and Fong and Geyer (2002). Freshwater coming from the Connecticut River is surface trapped, that is, without deep signatures of either salinity or velocity (Figs. 9a,d). Therefore, the horizontal extension of the buoyancy-influenced region must significantly increase to maintain mass conservation, as illustrated in Fig. 8a.

Fig. 8.
Fig. 8.

Time-averaged surface salinity for the case (a) Buoy and (b) RivTides after 100 days’ evolution.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Fig. 9.
Fig. 9.

Cross-shelf sections of buoyant plume normal to the south shore of Long Island (see Fig. 1) for the cases of Buoy, RivTides, and RivTides_500. Along-shelf velocity for the cases (a) Buoy, (b) RivTides, and (c) RivTides_500; across-shelf velocity for the cases (d) Buoy, (e) RivTides, and (f) RivTides_500. The white lines represent salinity, and the thick gray lines represent the isolines of 0.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

In contrast to the broad surface extension in Buoy, the plume in RivTides is limited to a much smaller area encompassing LIS, BIS, south of BIS, and east of RIS with a sluggish, horizontal propagation speed, according to the salinity field comparisons (Fig. 8). After about 4 months’ simulation, the structure of the buoyant plume reaches a steady state with a confined offshore extension. The plume structure in the eighth month is the same with the structure at the fourth month.

The difference between Buoy and RivTides derives from the deep penetration of freshwater in RivTides and is consistent with the available literatures, for example, Simpson (1997), which emphasized the importance of tidal mixing in the plume spread. At the same time, the case RivTides generates a constant upstream flow of ~3 cm s−1 from RIS imposed on the plume to the southwest of BIS. According to a series of process-oriented experiments and the theoretical analysis of the depth-averaged vorticity equation, we infer that the upstream current in RivTides is due to the topographically induced tidal rectification. This current works to turn the trajectories of the particles into RIS back to BIS. Therefore, waters entering RIS through the gap between Block Island and Point Judith are isolated from RIS waters. According to Chapman and Lentz’s (1994) study, because of the limitation by the upstream current, we infer that the plume in RivTides should be narrower and has a larger horizontal density gradient than the theoretical estimates of Yankovsky and Chapman (1997). In the following discussions the effects of tides will be implicitly included.

Figures 9b and 9e show the cross-shelf sections of the along-shelf and cross-shelf velocities and salinity along the section outside BIS, normal to the southern shore of Long Island (section 4 in Fig. 1) after 8 months’ evolution. We choose this section away from the mouth of BIS to avoid the headland eddy around BIS. The total width of the plume is Wp = Wb + Ws and is composed of the part in direct contact with the bottom Wb and the part away from the bottom (from the bottom offshore edge of the plume to its surface offshore edge Ws). Since Wb < Ws, the plume generated by RivTides is classified as “intermediate,” with isopycnals sloping from the bottom at about 20 km offshore of the coast to the surface at 20 km farther offshore. The trapping depth, also called the attachment depth, at which the bottom-advected plume becomes trapped (Yankovsky and Chapman 1997) with the onshore and offshore velocities reaching a balance, is about 20 m for the transect chosen, much smaller than the estimation by Kirincich and Hebert (2005) and Codiga (2005) at a farther upstream region around the mouth of BIS.

The alongshelf flow is surface intensified with almost linear vertical shear, reversing within several meters of the bottom at the foot of the front. It is geostrophic, by comparison of the full physics model output with a thermal wind calculation (Fig. 10) and by the diagnostic analysis of the across-shelf momentum balance (Fig. 11a). During the thermal wind calculation, we approximately set the bottom, weak velocity less than 1 cm s−1 as 0, and in the diagnostic analysis, we cast the momentum equations into a right-hand coordinate system with x directed upshelf along the shore of Long Island and y directed onshore. Meanwhile, the offshore flow responsible for the offshore spread of the inner-shelf fresher waters penetrates to the depth where the along-shelf velocity reverses. Such deep onshore flow has been observed in the southwest of BIS by the moored profiling current meter records in spring when wind effects are excluded (Codiga 2005; Fig. 7). Diagnostic analysis of the along-shelf momentum balance in Fig. 11a verifies Yankovsky and Chapman’s (1997) hypothesis that the deep, onshore transport is due to bottom Ekman dynamics with a balance reached by the pressure gradient, Coriolis force, and vertical viscosity terms.

Fig. 10.
Fig. 10.

Cross-shelf section of the downshelf velocities based on (color) the direct model output and (isoline) the thermal wind relation with the bottom velocity approximately set as 0. The thick gray line represents the zero isoline of the direct model output.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Fig. 11.
Fig. 11.

Momentum terms in the (a) along-shelf and (b) across-shelf directions at 15 km offshore for the cases of RivTides. The thick, solid lines represent pressure gradient; the thick, dashed lines represent Coriolis term; the thin, solid lines represent along-shelf, u-momentum advection; and the thin, dashed lines represent vertical viscosity.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

b. Seasonal variability of the plume in the case Real

We define the average from January to March as winter, April to June as spring, July to September as summer, and October to December as fall. The comparison between the simulated depth-averaged currents and observations from the FRONT project (Ullman and Codiga 2004) indicates a strong seasonal cycle of circulation. Figure 12 shows the seasonal-mean surface salinity, and Fig. 13 shows the seasonal mean of the vertical structure for the transect perpendicular to Long Island. They demonstrate a strong seasonal variability of the BIS estuarine plume generated by the freshwater emanating from LIS.

Fig. 12.
Fig. 12.

Seasonal variability of surface salinity for the experiment Real in (a) winter, (b) spring, (c) summer, and (d) fall.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Fig. 13.
Fig. 13.

Seasonal variability of the buoyant plume’s vertical structure along section 4 in Fig. 1 for the case Real. White lines represent salinity, and colors represent (left) along-shelf velocity and (right) across-shelf velocity. (a),(e) winter, (b),(f) spring, (c),(g) summer, (d),(h) fall.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

BIS is fresher in winter and spring than in summer and fall, with the widest and strongest buoyant plume in spring when river discharge peaks and the narrowest and weakest plume in summer when river discharge is smallest (Fig. 13). Compared with that in spring, freshwater in winter is limited in its downshelf penetration and has a saltier plume, though their offshore spreading is comparable.

The transect normal to Long Island reveals an intermediate plume in spring with the bottom offshore edge (32.6 psu) around 15 km offshore and is accompanied by a robust, surface-intensified, downshelf transport. The downshelf jet centers around 10 km offshore. In the across-shelf direction, there is an offshore transport at the surface and a returning, onshore transport below it. This structure can be explained by the idealized experiments in the next section. At depth, there are currents opposite the transport above it, which is associated with the bottom Ekman layer. Compared with the other seasons, during spring the waters on the landside of the front are freshest; the front has the sharpest gradient and widest frontal width with a slope angle around 0.001. The slope for the isohalines at the offshore edge of the plume is even smaller.

Though there are not enough observations in summer to delineate the buoyant plume, the fact that our modeling result agrees well with the limited current observation gives us confidence in presenting the summer features. During this season, a robust downshelf jet larger than 16 cm s−1 moves through the salinity front. The summer plume is closer to bottom advected with a much narrower plume width at the surface (~22 km) and steeper frontal slope (~0.002 for the isohaline of 32 psu, doubling the slope in spring) than the other seasons. Also, its bottom upshelf layer almost disappears, and the onshore deep transport is weaker and confined to a narrower region.

In winter and fall, the plume is also intermediate. As described by Codiga (2005), the along-shelf transport is significantly weakened and moved offshore with the shallowest penetration depth of the shallow transport, which is especially obvious for winter’s result. In winter, the freshwater patch of 31.4 psu is transported away from the coast; the center of the downshelf jet moves to a region farther than 30 km offshore, and the deep upshelf jet is much thicker than the other seasons. At the same time, the across-shelf flow shows a strong surface offshore transport, even on the landside of the front, corresponding to a broad surface extension and small slope angles. The bottom onshore transport is obviously strengthened, spanning across the bottom of the entire transect and extends even more offshore. Correspondingly, the detachment depth disappears without a balance between the offshore and the onshore transport at the bottom.

In fall, the downshelf jet is also diverted offshore and a strong offshore transport appears, generating a slope angle smaller than summer but larger than spring with a value of about 0.018 for isohaline of 32 psu. In contrast to winter, in fall, at the bottom of ~8 km offshore, the offshore transport at the landside and the onshore transport at the seaside reach a balance, generating a detachment depth of ~35 m.

c. Response to winds

The experiment NoWind, which omits the effects of local winds, produces a bottom-advected plume in winter and fall with strengthened downshelf transport and weakened offshore transport at the surface layers, as shown in Fig. 14. Meanwhile, the deep upshelf transport and onshore transport become weaker and thinner; the center of the surface downshelf transport is located more toward the inner shelf than in the experiment Real.

Fig. 14.
Fig. 14.

As in Fig. 13, but for the case without winds’ effect (NoWind).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

In spring, the plume width in NoWind is narrower than in Real. Taking the isohaline of 31.6 psu as an example, its surface outcropping position moves from 32 to 18 km offshore, while its contact with the bottom is kept intact. Another important feature in the NoWind experiment is that, water at the inner shelf becomes fresher, creating a larger salinity gradient than in Real, especially in spring. This may be due to the change of freshwater delivery out of BIS caused by omission of the winds.

To further examine the roles of wind, we analyze the four idealized experiments driven by winds in different directions: WndOn, WndOff, WndUp, and WndDown. Figure 15 shows the responses of surface salinity after 3 days of wind forcing and the Lagrangian trajectories of the surface drifters during the initial 5 days, in which both the buoyant plume and the trajectories vary with wind direction. In every case, upon leaving the Connecticut River, most of freshwater is stored in LIS, while a smaller fraction is transported out of LIS and mixed with the outer-shelf saline waters. As shown by surface salinity and freshwater flux (Table 2), an upwelling-favorable wind is most effective in spreading the freshwater offshore (Fig. 15c), and a downwelling-favorable wind is most effective in squeezing the plume against the Long Island coast (Fig. 15d).

Fig. 15.
Fig. 15.

Plane view of surface salinity after 3 days of wind forcing and the Lagrangian trajectories of the drifters released at the locations marked by stars at the first day of wind forcing and simulated with 1-h interval for the cases driven by (a) onshore winds (WndOn), (b) offshore winds (WndOff), (c) upwelling-favorable winds (WndUp), and (d) downwelling-favorable winds (WndDown).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Table 2.

Freshwater flux (m3 s−1) across the sections in Fig. 1 for the idealized experiments with different winds and the buoyancy of the plume, that is, g′h0 (m2 s−2) along section 2.

Table 2.

The winds’ effects on freshwater dispersion are accompanied by the changes in the vertical current structure (Fig. 16). In the along-shelf direction, the onshore wind reduces the downshelf velocities at the surface via Ekman dynamics (Fig. 16a); in the across-shelf direction, comparing WndOn with RivTides, the onshore wind generates a surface onshore current (larger than 3 cm s−1) and a returning flow directly below it (Fig. 16e), which is responsible for the sharp halocline around 10-m depth. Such a structure is in agreement with the study of Tilburg (2003), who found a significant surface transport and a returning flow below it, though the depth-integrated transport is feeble. Similarly, while the offshore wind produces an offshore transport at the surface and a deeper returning, onshore transport below it (Fig. 16b), it triggers stronger alongshelf velocities via strong surface and bottom Ekman dynamics (Fig. 16f).

Fig. 16.
Fig. 16.

Buoyant plume’s vertical structure along section 4 in Fig. 1. Alongshelf velocity for (a) WndOn, (b) WndOff, (c) WndUp, and (d) WndDown; across-shelf velocities for (e) WndOn, (f) WndOff, (g) WndUp, and (h) WndDown.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Except in a thin layer around 10 m and offshore of 10 km, an upwelling-favorable wind drastically alters the alongshelf flow by reversing the downshelf currents over the water column (Fig. 16c). In the cross-shelf direction, an upwelling wind drives the waters shallower than 10 m offshore and activates a stronger bottom returning current, bringing about an intrusion of offshore saline waters (Fig. 16g), which detaches the core of the buoyant plume from the coast and diverts the downshelf transport of freshwater offshore. This is consistent with the study by Codiga (2005). He found that in fall and winter, when the upwelling-favorable winds dominate, the fronts extend farther offshore at the surface, the bottom onshore flow becomes stronger, and the downshelf transport becomes weaker. Conversely, a downwelling-favorable wind creates a strong onshore transport in shallower layers and a deep, offshore flow at the bottom, at the same time accelerating the downshelf transport and squeezing the front into a much narrower band (Figs. 16d,h).

d. Response to river discharge

Besides wind effects, buoyant discharge variations can result in changes in the plume and its front (Chapman and Lentz 1994). To examine this, we configure two idealized experiments driven by constant river and tides only: RivTides with a constant river discharge of 1000 m3 s−1 and RivTides_500 with half of that river discharge (500 m3 s−1). The resulting buoyant plumes are shown in the middle (Figs. 9b,e) and bottom (Figs. 9c,f) panels of Fig. 9, whereby we find the stronger river input broadens the plume’s width by more than 10 km at the surface and increases the strength of the front. The doubled river discharge doubles the salinity difference between the seaside edge and landside edge, which is approximately 0.6 psu in RivTides_500 and larger than 1.2 psu in RivTides. Therefore, in light of the pronounced seasonal variation in LIS river input, we can infer that the increase in river discharge from winter to spring as shown in Fig. 2 contributes significantly to the sharper and broader plume in spring, especially when the river discharge peaks in April, and the abrupt decrease from spring to summer should be responsible for the weaker and narrower plume in summer, when discharge subsides to the bottom.

On the other hand, the significant change in river input only slightly changes the offshore extension of the plume and has a limited influence (within 1 cm s−1) on the cross-shelf and alongshelf velocities. However, in the NoWind experiment, we notice much more obvious changes in the velocities and plume’s offshore extension. This fact, together with the study of Chapman and Lentz (1994), prompt us to investigate the effects of the upstream current from RIS and the deep thermal fronts.

e. Plume response to surface heating

Chapman and Lentz (1994) paid attention to the upstream flow’s impacts on the evolution of the buoyant plume, which is highly seasonally variable in BIS and worthy of note. According to current observations at a mooring site to the south of Block Island (referred to as MD-S), Ullman and Codiga (2010) found strong seasonal changes of the upstream current from RIS. During spring and summer, the monthly mean subtidal current reaches to 25 cm s−1, while in fall and winter it is 5–10 cm s−1.

The seasonal change of upstream current from RIS and the deep thermal fronts are both related to the variation in surface heating. By artificially changing shortwave radiation in idealized experiments, we find that the increase in surface heating forms a stronger deep thermal front and a stronger cyclonic circulation around RIS, and, therefore, a stronger upstream current is imposed on the plume front. In addition, the comparison of surface salinity indicates that the increase in surface heating pushes the surface plume onshore (Fig. 17), which results from the increase of the upstream current from RIS according to the theory by Chapman and Lentz (1994).

Fig. 17.
Fig. 17.

Surface salinity for the cases with (a) weak surface heating (HeatW) and (b) strong surface heating (HeatS) after 2 months of simulation.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

In the vertical, compared with the nonstratified case RivTides, the increase of surface heating in HeatW and HeatS significantly increases the surface-intensified downshelf current with a narrower and weaker upshelf current (Fig. 18). For the cross-shelf structure, the experiments with surface heating show slightly weaker and thinner bottom onshore transport; at the same time, a surface offshore transport occurs, between which there is an offshore transport.

Fig. 18.
Fig. 18.

Buoyant plume’s vertical structure along section 4 in Fig. 1. Along-shelf velocity for the experiment with (a) weak surface heating (HeatW) and (b) strong surface heating (HeatS); across-shelf velocity for the experiment with (c) weak surface heating (HeatW) and (d) strong surface heating (HeatS).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Figures 19 and 20 show the momentum balance in the along-shelf and across-shelf directions for the experiments HeatW and HeatS, respectively. According to Fig. 11, for RivTides, the major balance in the alongshelf direction is geostrophic, while at the bottom, the bottom Ekman dynamics dominates. However, for HeatS, the pressure gradient term is drastically reduced to the scale of advection and vertical viscosity, and for waters shallower than 25 m, the pressure gradient direction changes, generating the surface onshore transport. The further addition of surface heating from HeatW to HeatS generates a stronger, surface downshelf pressure gradient and therefore a stronger surface onshore transport. For all three experiments, at the bottom, the dominant balance is the bottom Ekman dynamics, which generates the bottom onshore transport.

Fig. 19.
Fig. 19.

As in Fig. 11, but for the case with weak surface heating (HeatW).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Fig. 20.
Fig. 20.

As in Fig. 11, but for the case with strong surface heating (HeatS).

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

In fact, the across-shelf velocity structure is highly dependent on the location chosen. For a section closer to the mouth of BIS, the surface onshore transport disappears; compared with RivTides, a stronger offshore transport is formed. The difference may be due to the change of the along-shelf pressure gradient with its relative position to the buoyant bulge structure and with the topography along the coast.

In the across-shelf direction, the dominant balance is geostrophic for RivTides (Fig. 11b), HeatW (Fig. 19b), and HeatS (Fig. 20b). The increase of surface heating from RivTides to HeatW to HeatS has strengthened the pressure gradient.

As discussed above, the changes in pressure gradients [−(1/ρ0)(∂P/∂x), −(1/ρ0)(∂P/∂y)] dominate the response of currents and plume to surface heating. Following Csanady (1978) and Ullman and Codiga (2004), the pressure gradients can be decomposed into the part arising from density gradients (baroclinic component) and the part from sea surface elevation gradients (barotropic component):
e1
e2
where P is the pressure, ζ is the free-surface elevation, and ρ is the density relating to a reference density anomaly ρ0 and a dimensionless density anomaly ε {ρ = ρ0[1 + ε(x, y, z, t)]}. The contributions from the barotropic and baroclinic parts for the along-shelf direction are shown in Fig. 21 as solid and dashed lines, respectively. In the along-shelf direction, the pressure gradient arising from the baroclinic part always offsets the effects through the changes in sea surface height. Still, as discussed before, the change of the along-shelf pressure gradient is highly dependent on the location chosen. For a location closer to the mouth of BIS, the effects from barotropic and baroclinic parts change significantly.
Fig. 21.
Fig. 21.

Pressure gradients contributed by (solid lines) sea level variation and (dashed lines) density gradients in the (a) along-shelf direction and (b) across-shelf direction for the cases of (black) RivTides, (blue) HeatW, and (Red) HeatS.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

In the across-shelf direction, the pressure gradient also finds opposite roles played by the barotropic and baroclinic components; the sea level variation produces a stronger onshore pressure gradient, while the density gradient offsets the effect and contributes to a stronger offshore pressure gradient that is responsible for the stronger downshelf current in the experiments with stronger surface heating. Further examination finds that the density gradient is arising from temperature, more specifically, from the formation of a thermal front. Thus, surface heating controls the BIS plume by way of an upstream current and a bottom thermal front, with the effect by the bottom thermal front dominated in the across-shelf velocity.

The changes in vertical velocity structures from RivTides to HeatS are accompanied by a steeper and narrower plume over the whole depth; the surface edge of the plume (isohaline of 34.4 psu) shoals from 42 km offshore in RivTides to 22 km in HeatW to 18 km in HeatS. This can be explained by the increase in the along-shelf freshwater transport [u(∂S/∂x)] because of the increasing u, which suppresses the buoyancy’s offshore spreading. According to the above analysis, we can conclude that, in addition to the seasonally varying wind and buoyancy discharge, surface heating in spring and summer plays an important role in the buoyant plume by steepening the isohalines and narrowing the plume, in agreement with the study of Ullman and Codiga (2004) who stressed the importance of horizontal density gradients through analysis of historical hydrographic data.

5. Discussion

Freshwater delivery is quantified by a freshwater flux that is defined as
e3
where Sa represents ambient salinity, S is salinity, and u represents velocity across the section. The flux is integrated over the whole depth from the bottom −h to the sea surface displacement ζ over the length of the domain starting with ls and ending with le.

Figure 22 shows the monthly mean river input into LIS (black line), and the mean freshwater delivery out of LIS (blue line) and out of BIS into the estuarine plume (red line) across sections 2 and 3 in Fig. 1 as well as into RIS through the gap between Block Island and Point Judith (green line) for the case Real.

Fig. 22.
Fig. 22.

River discharge from the Connecticut River (black line) and the freshwater delivery out of LIS (blue line), out of BIS (red line), and into RIS (green line) for the case Real. The freshwater flux out of LIS (dashed blue line) and BIS (dashed red line) for the case NoWind is compared with the case Real.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

According to the model results, the seasonal variability of the freshwater delivery into the BIS estuarine plume is highly correlated with the river discharge (r = 0.78). In April, when river discharge peaks, the transport is largest, with most of the freshwater moving across the mouth of LIS to the ocean and a relatively smaller part moving westward into the middle and west parts of LIS or mixing vertically. From May to September, consistent with the trend of river discharge, the transport through LIS decreases significantly. However, the freshwater transport out of LIS is larger than river discharge, and the minimum transport out of LIS is delayed by 1 month compared with the discharge. This is related to the seasonal change of turbulent mixing.

a. Relation between freshwater delivery and turbulent mixing

The turbulent mixing can be represented by the vertical turbulent buoyancy fluxes B, which are represented as B = Kρg(∂ρ/∂z), where Kρ is the vertical salinity diffusivity, g is gravitational acceleration, and ρ is potential density. Figures 23a and 23b show the vertically integrated turbulent mixing in LIS and BIS (blue lines), respectively. To examine its relationship with horizontal freshwater delivery, we remove the seasonal trend of freshwater flux from the upstream (green lines in Figs. 23a,b). The horizontal freshwater propagation is negatively correlated with turbulent mixing.

Fig. 23.
Fig. 23.

(a) Correlation between the vertically integrated turbulent mixing in LIS and the horizontal freshwater propagation out of LIS without the seasonal trend from the upstream. (b) As in (a), but for BIS.

Citation: Journal of Physical Oceanography 46, 5; 10.1175/JPO-D-15-0099.1

Turbulent mixing in LIS shows strong values in winter and fall, while much weaker in spring and summer. The strong mixing in winter and fall entrains more freshwater inside LIS. In spring and summer, the weak values favor the horizontal spread. Therefore, although the river discharge declines significantly from April to August, the weakening of vertical mixing deaccelerates the change of delivery out of LIS. Meanwhile, freshwater coming out of the interior and west of LIS, which is originally from the Connecticut River water that moves westward into LIS before flowing eastward into the open ocean, also contributes to the freshwater transport out of LIS. This can explain why the freshwater out of LIS is larger than the river discharge from May to August.

Most of the freshwater entrained out of LIS flows into the open ocean through BIS, and only a small fraction flows into RIS (Fig. 22). Similar to the change of freshwater delivery out of LIS, the seasonal change through BIS is also negatively correlated with the change of turbulent mixing in BIS (Fig. 23). There is more freshwater entering BIS through LIS in April than June. However, the stronger turbulent mixing in April, about 2 times that in June, restricts the propagation of freshwater.

Even though turbulent mixing is affected by the strength of winds (O’Donnell et al. 2014), the comparison of turbulent mixing (not shown here) for the experiments Real and NoWind indicates the major seasonal change of turbulent mixing is not due to winds. However, winds are important for the freshwater delivery by the change of directions.

b. Relation between freshwater delivery and wind directions

Another dominant contributor to the change of freshwater delivery is wind. For an open ocean, the maximum Ekman transport is the same for onshore, offshore, upwelling-favorable, and downwelling-favorable winds with the same amplitude, that is, Mek = τ/ρ0f (τ is the surface momentum stress and f is the Coriolis parameter), perpendicular to the direction of winds. However, the complex geometry in the studied region restricts the exchange of freshwater between the estuary and the ocean. Table 2 lists freshwater delivery out of LIS and BIS for different wind directions. The presence of upwelling-favorable and offshore winds transports more freshwater out of LIS than the case free of winds. In the studied region, a typical wind speed, about 5 m s−1, corresponds to the Ekman layer depth of 47 m (Luo et al. 2013), which is larger than the water depth inside LIS. So the effect of bottom friction steers the wind-driven flow to the right-hand side of wind with an angle smaller than 45° or even in wind direction when the water depth is much smaller than the Ekman depth. Therefore, considering the geometry, the upwelling-favorable and offshore winds deliver more freshwater out of LIS, while the onshore and downwelling-favorable winds suppress the delivery.

After exiting LIS, the upwelling wind rapidly spreads the buoyant plume eastward to the broader outer-shelf region through BIS at a rate of 785 m3 s−1 and through the gap between Point Judith and Block Island at a rate of 256 m3 s−1. By contrast, the offshore wind constrains most of the freshwater from LIS to the outer-shelf region directly through BIS; the freshwater flux through BIS is 608 m3 s−1 when leaving LIS at the rate of 614 m3 s−1 (Table 2).

The application of onshore and downwelling-favorable winds retains more freshwater in LIS by decreasing the freshwater flux leaving LIS to 530 m3 s−1 in WndOn and 331 m3 s−1 in WndDown. Even though the flux in WndDown is small, the wind is much more effective in delivering freshwater to the BIS estuarine plume, while the onshore wind delivers more than half of the freshwater going eastward into RIS through the gap between Point Judith and Block Island. Note that most of the freshwater going into RIS is entrained back to BIS through south of Block Island, which explains why the plume appears to extend farther offshore under onshore winds than offshore winds.

According to Fong and Geyer (2002), for a given latitude and background salinity, the freshwater transport is a function of only g′h0, the buoyancy of the plume, where g′ {g′ = [(ρ0 − ρ)/ρ0]g} is the reduced gravity based on the depth-averaged density anomaly, ρ0 is ambient water density, and ρ is the depth-averaged density. For each case, we calculate the potential energy at the coast (Table 2). The change of the buoyancy of the plume is consistent with the freshwater delivery, that is, larger for WndUp and WndOff and smaller for WndDown and WndOn.

The drifter trajectories shown in Fig. 15 agree with the freshwater flux in Table 2, from which we find that the upwelling-favorable wind rapidly spreads the drifters released at the mouth of LIS eastward through both RIS and BIS, while the offshore wind drives all the drifters, even those closest to the northern shore of BIS, southward. Moreover, the onshore wind pushes the drifters eastward into RIS, and in the experiment with downwelling-favorable wind, drifters move around the mouth of LIS with little chance of leaving LIS.

The above analysis can explain the freshwater flux change from Real to NoWind (Fig. 22). During fall and winter, the study region experiences prevailing upwelling-favorable and offshore winds, both of which strengthen the freshwater advection into the buoyant plume. Therefore, omitting local winds results in a weaker freshwater flux leaving BIS. In spring, the upwelling-favorable wind and onshore wind exert opposite influences on the freshwater flux; unlike an upwelling wind, an onshore wind suppresses the freshwater delivery, which seems more important to the buoyancy transport in spring by holding more freshwater inside the estuaries. The winds in summer act similarly, though the amplitude is much smaller. Outside BIS, Ekman dynamics becomes important by changing the vertical structure of the buoyant plume.

6. Summary

A seasonally varying buoyant plume has been observed southwest of BIS. This study uses a numerical model to investigate its seasonal variability and its response to tides, winds, and surface heating. Idealized model experiments indicate that tidal mixing is required in the formation of the bottom-advected plume southwest of BIS by isolating the freshwater coming out of LIS to a deeper depth and trapping the buoyant plume in a confined region.

The effects of winds on the buoyant plume are examined by two process-oriented experiments and a series of idealized experiments driven by the typical winds—onshore, offshore, upwelling-favorable, and downwelling-favorable winds. Analysis of the idealized experiments reveals that the upwelling-favorable and offshore winds are effective in spreading freshwater out of LIS and BIS in the offshore direction, while the downwelling-favorable and onshore winds tend to resist the delivery of freshwater out of BIS and traps the plume closer to the Long Island coast. This explains the plume’s seasonal variability shown in the realistic experiment: in winter and spring, when the upwelling-favorable winds dominate, an intermediate plume occurs; in summer, when BIS is subjected to weak onshore and downwelling-favorable winds, the bottom-advected plume with steepest front and narrowest offshore extension is generated.

In addition, the seasonal variability in river discharge, mainly from the Connecticut River, plays an important role in the plume’s width and strength. When river discharge peaks in spring, the plume is widest and has a very small frontal slope of about 0.001. The decrease in river discharge from spring to summer contributes to the narrower and weaker plume. However, the river discharge’s impact on the offshore edge at the bottom and the along-shelf and across-shelf velocities across the salinity front is limited. Meanwhile, the study finds that the seasonal fluctuation of the freshwater delivery out of LIS and BIS is controlled by the seasonal change of turbulent mixing. In winter and fall, when turbulent mixing is strong, more freshwater is retained by the increase of vertical mixing; in spring and summer, the decrease of turbulent mixing favors the horizontal propagation of freshwater.

This study also demonstrates that the seasonal variation of surface heating is a competitive contributor to the seasonal plume. According to the idealized experiments driven by the different extent of surface heating, we find that the increase of surface heating eventually produces a plume with a narrower offshore extension both at the surface and bottom. Meanwhile, downshelf transport across the front becomes stronger, which is mainly due to variations in density gradients. Thus, in spring, the growing surface heating increases the upstream current from RIS and forms a bottom thermal front, which tends to compete with the spring discharge peak by shoaling the bottom offshore edge of the plume and limiting the surface spreading of buoyancy. In summer, the narrowest and shallowest plume occurs because of a combination of several physical processes. First, the smallest river discharge limits the freshwater transport into the plume. Second, the appearance of weak onshore and downwelling-favorable winds tends to push the surface plume onshore. Third, the strong surface heating pushes the bottom offshore edge of the plume to an even shallower depth and narrower band.

Acknowledgments

We thank Dan Codiga, David Ullman, and Jim O’Donnell for graciously sharing their observational data with us. Q. Liu would like to thank Dave Ullman, Chris Kincaid, and Jason Dahl for their insightful comments on her work. Y. Luo acknowledges the support of the National Natural Science Foundation of China (41376009) and the Joint Program of Shandong Province and National Natural Science Foundation of China (U1406401). This work was supported by the URI Regional Earth Systems Center through DOE Grant DE-SC0005432 and by the National Science Foundation EPSCOR/STAC program through Grants S000216 and 0004835.

REFERENCES

  • Chant, R. J., S. M. Glenn, E. Hunter, J. Kohut, R. F. Chen, R. W. Houghton, J. Bosch, and O. Schofield, 2008: Bulge formation of a buoyant river outflow. J. Geophys. Res., 113, C01017, doi:10.1029/2007JC004100.

    • Search Google Scholar
    • Export Citation

  • Chapman, D., and S. Lentz, 1994: Trapping of a coastal density front by the bottom boundary layer. J. Phys. Oceanogr., 24, 1464–1479, doi:10.1175/1520-0485(1994)024<1464:TOACDF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Codiga, D. L., 2005: Interplay of wind forcing and buoyant discharge off Montauk Point: Seasonal changes to velocity structure and a coastal front. J. Phys. Oceanogr., 35, 1068–1085, doi:10.1175/JPO2736.1.

    • Search Google Scholar
    • Export Citation

  • Codiga, D. L., and A. E. Houk, 2002: Current profile timeseries from the FRONT moored array. University of Connecticut Department of Marine Sciences Tech. Rep., 19 pp.

  • Codiga, D. L., and D. A. Aurin, 2007: Residual circulation in eastern Long Island Sound: Observed transverse-vertical structure and exchange transport. Cont. Shelf Res., 27, 103–116, doi:10.1016/j.csr.2006.09.001.

    • Search Google Scholar
    • Export Citation

  • Csanady, G. T., 1978: The arrested topographic wave. J. Phys. Oceanogr., 8, 47–62, doi:10.1175/1520-0485(1978)008<0047:TATW>2.0.CO;2.

  • Edwards, C. A., T. A. Fake, and P. S. Bogden, 2004a: Spring-summer frontogenesis at the mouth of Block Island Sound: 1. A numerical investigation into tidal and buoyancy-forced motion. J. Geophys. Res., 109, C12021, doi:10.1029/2003JC002132.

    • Search Google Scholar
    • Export Citation

  • Edwards, C. A., T. A. Fake, and P. S. Bogden, 2004b: Spring-summer frontogenesis at the mouth of Block Island Sound: 2. Combining acoustic Doppler current profiler records with a general circulation model to investigate the impact of subtidal forcing. J. Geophys. Res., 109, C12022, doi:10.1029/2003JC002133.

    • Search Google Scholar
    • Export Citation

  • Fong, D. A., and R. W. Geyer, 2002: The alongshore transport of freshwater in a surface-trapped river plume. J. Phys. Oceanogr., 32, 957–972, doi:10.1175/1520-0485(2002)032<0957:TATOFI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Garvine, R. W., 1974: Physical features of the Connecticut River outflow during high discharge. J. Geophys. Res., 79, 831–846, doi:10.1029/JC079i006p00831.

    • Search Google Scholar
    • Export Citation

  • Haidvogel, D. B., H. G. Arango, K. Hedstrom, A. Beckmann, P. Malanotte-Rizzoli, and A. F. Shchepetkin, 2000: Model evaluation experiments in the North Atlantic basin: Simulations in nonlinear terrain-following coordinates. Dyn. Atmos. Oceans, 32, 239–281, doi:10.1016/S0377-0265(00)00049-X.

    • Search Google Scholar
    • Export Citation

  • Horner-Devine, A. R., R. D. Hetland, and D. G. MacDonald, 2015: Mixing and transport in coastal river plumes. Annu. Rev. Fluid Mech., 47, 569–594, doi:10.1146/annurev-fluid-010313-141408.

    • Search Google Scholar
    • Export Citation

  • Kincaid, C., R. A. Pockalny, and L. M. Huzzey, 2003: Spatial and temporal variability in flow at the mouth of Narragansett Bay. J. Geophys. Res., 108, 3218, doi:10.1029/2002JC001395.

    • Search Google Scholar
    • Export Citation

  • Kirincich, A., and D. Hebert, 2005: The structure of the coastal density front at the outflow of Long Island Sound during spring 2002. Cont. Shelf Res., 25, 1097–1114, doi:10.1016/j.csr.2004.12.014.

    • Search Google Scholar
    • Export Citation

  • Kourafalou, V. H., L.-Y. Oey, J. D. Wang, and T. N. Lee, 1996: The fate of river discharge on the continental shelf: 1. Modeling the river plume and the inner shelf coastal current. J. Geophys. Res., 101, 3415–3434, doi:10.1029/95JC03024.

    • Search Google Scholar
    • Export Citation

  • Latimer, J. S., M. A. Tedesco, R. L. Swanson, C. Yarish, P. E. Stacey, and C. Garza, Eds., 2014: Long Island Sound: Prospects for the Urban Sea. Springer Series on Environmental Management, Springer, 558 pp., doi:10.1007/978-1-4614-6126-5.

  • Lentz, S. J., and K. R. Helfrich, 2002: Buoyant gravity currents along a sloping bottom in a rotating fluid. J. Fluid Mech., 464, 251–278, doi:10.1017/S0022112002008868.

    • Search Google Scholar
    • Export Citation

  • Luettich, R. A., J. J. Westerink, and N. W. Scheffner, 1992: ADCIRC: An advanced three-dimensional circulation model for shelves, coasts, and estuaries. Report 1: Theory and methodology of ADCIRC-2DDI and ADCIRC-3DL. Department of the Army Dredging Research Program Tech. Rep. DRP-92-6, 143 pp.

  • Luo, Y., L. Rothstein, Q. Liu, and S. Zhang, 2013: Climatic variability of the circulation in the Rhode Island Sound: A modeling study. J. Geophys. Res. Oceans, 118, 4072–4091, doi:10.1002/jgrc.20285.

    • Search Google Scholar
    • Export Citation

  • Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343–360, doi:10.1175/BAMS-87-3-343.

    • Search Google Scholar
    • Export Citation

  • Moffat, C., and S. Lentz, 2012: On the response of a buoyant plume to downwelling-favorable wind stress. J. Phys. Oceanogr., 42, 1083–1098, doi:10.1175/JPO-D-11-015.1.

    • Search Google Scholar
    • Export Citation

  • O’Donnell, J., 2010: The dynamics of estuary plumes and fronts. Contemporary Issues in Estuarine Physics, A. Valle-Levinson, Ed., Cambridge University Press, 186–246, doi:10.1017/CBO9780511676567.009.

  • O’Donnell, J., and Coauthors, 2014: The physical oceanography of Long Island Sound. Long Island Sound: Prospects for the Urban Sea, J. S. Latimer et al., Eds., Springer Series on Environmental Management, Springer, 79–158.

  • Shchepetkin, A. F., and J. C. McWilliams, 1998: Quasi-monotone advection schemes based on explicit locally adaptive dissipation. Mon. Wea. Rev., 126, 1541–1580, doi:10.1175/1520-0493(1998)126<1541:QMASBO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Shchepetkin, A. F., and J. C. McWilliams, 2003: A method for computing horizontal pressure-gradient force in an oceanic model with a nonaligned vertical coordinate. J. Geophys. Res., 108, 3090, doi:10.1029/2001JC001047.

    • Search Google Scholar
    • Export Citation

  • Shchepetkin, A. F., and J. C. McWilliams, 2005: The Regional Oceanic Modeling System (ROMS): A split-explicit, free-surface, topography-following-coordinate oceanic model. Ocean Modell., 9, 347–404, doi:10.1016/j.ocemod.2004.08.002.

    • Search Google Scholar
    • Export Citation

  • Simpson, J. H., 1997: Physical processes in the ROFI regime. J. Mar. Syst., 12, 3–15, doi:10.1016/S0924-7963(96)00085-1.

  • Tilburg, C. E., 2003: Across-shelf transport on a continental shelf: Do across-shelf winds matter? J. Phys. Oceanogr., 33, 2675–2688, doi:10.1175/1520-0485(2003)033<2675:ATOACS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Ullman, D. S., and D. L. Codiga, 2004: Seasonal variation of a coastal jet in the Long Island Sound outflow region based on HF radar and Doppler current observations. J. Geophys. Res., 109, C07S06, doi:10.1029/2002JC001660.

    • Search Google Scholar
    • Export Citation

  • Ullman, D. S., and D. L. Codiga, 2010: Characterizing the physical oceanography of coastal waters off Rhode Island, Part 2: New observations of water properties, currents, and waves. Rhode Island Ocean Special Area Management Plan, Tech. Rep., 111 pp.

  • Whitney, M. M., and R. W. Garvine, 2006: Simulating the Delaware Bay buoyant outflow: Comparison with observations. J. Phys. Oceanogr., 36, 3–21, doi:10.1175/JPO2805.1.

    • Search Google Scholar
    • Export Citation

  • Whitney, M. M., and D. L. Codiga, 2011: Response of a large stratified estuary to wind events: Observations, simulations, and theory for Long Island Sound. J. Phys. Oceanogr., 41, 1308–1327, doi:10.1175/2011JPO4552.1.

    • Search Google Scholar
    • Export Citation

  • Yankovsky, A. E., and D. C. Chapman, 1997: A simple theory for the fate of buoyant coastal discharges. J. Phys. Oceanogr., 27, 1386–1401, doi:10.1175/1520-0485(1997)027<1386:ASTFTF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Zhang, W. G., J. L. Wilkin, and H. G. Arango, 2010: Towards an integrated observation and modeling system in the New York Bight using variational methods. Part I: 4DVAR data assimilation. Ocean Modell., 35, 119–133, doi:10.1016/j.ocemod.2010.08.003.

    • Search Google Scholar
    • Export Citation
Save
Cover Journal of Physical Oceanography
Journal of Physical Oceanography

  • Chant, R. J., S. M. Glenn, E. Hunter, J. Kohut, R. F. Chen, R. W. Houghton, J. Bosch, and O. Schofield, 2008: Bulge formation of a buoyant river outflow. J. Geophys. Res., 113, C01017, doi:10.1029/2007JC004100.

    • Search Google Scholar
    • Export Citation

  • Chapman, D., and S. Lentz, 1994: Trapping of a coastal density front by the bottom boundary layer. J. Phys. Oceanogr., 24, 1464–1479, doi:10.1175/1520-0485(1994)024<1464:TOACDF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Codiga, D. L., 2005: Interplay of wind forcing and buoyant discharge off Montauk Point: Seasonal changes to velocity structure and a coastal front. J. Phys. Oceanogr., 35, 1068–1085, doi:10.1175/JPO2736.1.

    • Search Google Scholar
    • Export Citation

  • Codiga, D. L., and A. E. Houk, 2002: Current profile timeseries from the FRONT moored array. University of Connecticut Department of Marine Sciences Tech. Rep., 19 pp.

  • Codiga, D. L., and D. A. Aurin, 2007: Residual circulation in eastern Long Island Sound: Observed transverse-vertical structure and exchange transport. Cont. Shelf Res., 27, 103–116, doi:10.1016/j.csr.2006.09.001.

    • Search Google Scholar
    • Export Citation

  • Csanady, G. T., 1978: The arrested topographic wave. J. Phys. Oceanogr., 8, 47–62, doi:10.1175/1520-0485(1978)008<0047:TATW>2.0.CO;2.

  • Edwards, C. A., T. A. Fake, and P. S. Bogden, 2004a: Spring-summer frontogenesis at the mouth of Block Island Sound: 1. A numerical investigation into tidal and buoyancy-forced motion. J. Geophys. Res., 109, C12021, doi:10.1029/2003JC002132.

    • Search Google Scholar
    • Export Citation

  • Edwards, C. A., T. A. Fake, and P. S. Bogden, 2004b: Spring-summer frontogenesis at the mouth of Block Island Sound: 2. Combining acoustic Doppler current profiler records with a general circulation model to investigate the impact of subtidal forcing. J. Geophys. Res., 109, C12022, doi:10.1029/2003JC002133.

    • Search Google Scholar
    • Export Citation

  • Fong, D. A., and R. W. Geyer, 2002: The alongshore transport of freshwater in a surface-trapped river plume. J. Phys. Oceanogr., 32, 957–972, doi:10.1175/1520-0485(2002)032<0957:TATOFI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Garvine, R. W., 1974: Physical features of the Connecticut River outflow during high discharge. J. Geophys. Res., 79, 831–846, doi:10.1029/JC079i006p00831.

    • Search Google Scholar
    • Export Citation

  • Haidvogel, D. B., H. G. Arango, K. Hedstrom, A. Beckmann, P. Malanotte-Rizzoli, and A. F. Shchepetkin, 2000: Model evaluation experiments in the North Atlantic basin: Simulations in nonlinear terrain-following coordinates. Dyn. Atmos. Oceans, 32, 239–281, doi:10.1016/S0377-0265(00)00049-X.

    • Search Google Scholar
    • Export Citation

  • Horner-Devine, A. R., R. D. Hetland, and D. G. MacDonald, 2015: Mixing and transport in coastal river plumes. Annu. Rev. Fluid Mech., 47, 569–594, doi:10.1146/annurev-fluid-010313-141408.

    • Search Google Scholar
    • Export Citation

  • Kincaid, C., R. A. Pockalny, and L. M. Huzzey, 2003: Spatial and temporal variability in flow at the mouth of Narragansett Bay. J. Geophys. Res., 108, 3218, doi:10.1029/2002JC001395.

    • Search Google Scholar
    • Export Citation

  • Kirincich, A., and D. Hebert, 2005: The structure of the coastal density front at the outflow of Long Island Sound during spring 2002. Cont. Shelf Res., 25, 1097–1114, doi:10.1016/j.csr.2004.12.014.

    • Search Google Scholar
    • Export Citation

  • Kourafalou, V. H., L.-Y. Oey, J. D. Wang, and T. N. Lee, 1996: The fate of river discharge on the continental shelf: 1. Modeling the river plume and the inner shelf coastal current. J. Geophys. Res., 101, 3415–3434, doi:10.1029/95JC03024.

    • Search Google Scholar
    • Export Citation

  • Latimer, J. S., M. A. Tedesco, R. L. Swanson, C. Yarish, P. E. Stacey, and C. Garza, Eds., 2014: Long Island Sound: Prospects for the Urban Sea. Springer Series on Environmental Management, Springer, 558 pp., doi:10.1007/978-1-4614-6126-5.

  • Lentz, S. J., and K. R. Helfrich, 2002: Buoyant gravity currents along a sloping bottom in a rotating fluid. J. Fluid Mech., 464, 251–278, doi:10.1017/S0022112002008868.

    • Search Google Scholar
    • Export Citation

  • Luettich, R. A., J. J. Westerink, and N. W. Scheffner, 1992: ADCIRC: An advanced three-dimensional circulation model for shelves, coasts, and estuaries. Report 1: Theory and methodology of ADCIRC-2DDI and ADCIRC-3DL. Department of the Army Dredging Research Program Tech. Rep. DRP-92-6, 143 pp.

  • Luo, Y., L. Rothstein, Q. Liu, and S. Zhang, 2013: Climatic variability of the circulation in the Rhode Island Sound: A modeling study. J. Geophys. Res. Oceans, 118, 4072–4091, doi:10.1002/jgrc.20285.

    • Search Google Scholar
    • Export Citation

  • Mesinger, F., and Coauthors, 2006: North American Regional Reanalysis. Bull. Amer. Meteor. Soc., 87, 343–360, doi:10.1175/BAMS-87-3-343.

    • Search Google Scholar
    • Export Citation

  • Moffat, C., and S. Lentz, 2012: On the response of a buoyant plume to downwelling-favorable wind stress. J. Phys. Oceanogr., 42, 1083–1098, doi:10.1175/JPO-D-11-015.1.

    • Search Google Scholar
    • Export Citation

  • O’Donnell, J., 2010: The dynamics of estuary plumes and fronts. Contemporary Issues in Estuarine Physics, A. Valle-Levinson, Ed., Cambridge University Press, 186–246, doi:10.1017/CBO9780511676567.009.

  • O’Donnell, J., and Coauthors, 2014: The physical oceanography of Long Island Sound. Long Island Sound: Prospects for the Urban Sea, J. S. Latimer et al., Eds., Springer Series on Environmental Management, Springer, 79–158.

  • Shchepetkin, A. F., and J. C. McWilliams, 1998: Quasi-monotone advection schemes based on explicit locally adaptive dissipation. Mon. Wea. Rev., 126, 1541–1580, doi:10.1175/1520-0493(1998)126<1541:QMASBO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Shchepetkin, A. F., and J. C. McWilliams, 2003: A method for computing horizontal pressure-gradient force in an oceanic model with a nonaligned vertical coordinate. J. Geophys. Res., 108, 3090, doi:10.1029/2001JC001047.

    • Search Google Scholar
    • Export Citation

  • Shchepetkin, A. F., and J. C. McWilliams, 2005: The Regional Oceanic Modeling System (ROMS): A split-explicit, free-surface, topography-following-coordinate oceanic model. Ocean Modell., 9, 347–404, doi:10.1016/j.ocemod.2004.08.002.

    • Search Google Scholar
    • Export Citation

  • Simpson, J. H., 1997: Physical processes in the ROFI regime. J. Mar. Syst., 12, 3–15, doi:10.1016/S0924-7963(96)00085-1.

  • Tilburg, C. E., 2003: Across-shelf transport on a continental shelf: Do across-shelf winds matter? J. Phys. Oceanogr., 33, 2675–2688, doi:10.1175/1520-0485(2003)033<2675:ATOACS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Ullman, D. S., and D. L. Codiga, 2004: Seasonal variation of a coastal jet in the Long Island Sound outflow region based on HF radar and Doppler current observations. J. Geophys. Res., 109, C07S06, doi:10.1029/2002JC001660.

    • Search Google Scholar
    • Export Citation

  • Ullman, D. S., and D. L. Codiga, 2010: Characterizing the physical oceanography of coastal waters off Rhode Island, Part 2: New observations of water properties, currents, and waves. Rhode Island Ocean Special Area Management Plan, Tech. Rep., 111 pp.

  • Whitney, M. M., and R. W. Garvine, 2006: Simulating the Delaware Bay buoyant outflow: Comparison with observations. J. Phys. Oceanogr., 36, 3–21, doi:10.1175/JPO2805.1.

    • Search Google Scholar
    • Export Citation

  • Whitney, M. M., and D. L. Codiga, 2011: Response of a large stratified estuary to wind events: Observations, simulations, and theory for Long Island Sound. J. Phys. Oceanogr., 41, 1308–1327, doi:10.1175/2011JPO4552.1.

    • Search Google Scholar
    • Export Citation

  • Yankovsky, A. E., and D. C. Chapman, 1997: A simple theory for the fate of buoyant coastal discharges. J. Phys. Oceanogr., 27, 1386–1401, doi:10.1175/1520-0485(1997)027<1386:ASTFTF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation

  • Zhang, W. G., J. L. Wilkin, and H. G. Arango, 2010: Towards an integrated observation and modeling system in the New York Bight using variational methods. Part I: 4DVAR data assimilation. Ocean Modell., 35, 119–133, doi:10.1016/j.ocemod.2010.08.003.

    • Search Google Scholar
    • Export Citation
  • <sc>Fig</sc>. 1.
    View in gallery
    Fig. 1.

    Bathymetry and geographic features around the Rhode Island Sound. Place names: Connecticut River (CR), Long Island Sound (LIS), Rhode Island Sound (RIS), Block Island Sound (BIS), Narragansett Bay (NB), Buzzards Bay (BB), and Vineyard Sound (VS). The thick red line is the section used to describe the buoyant plume to the southwest of BIS, and the thick black lines are the sections referenced in Table 2, as discussed further in section 5.

  • <sc>Fig</sc>. 2.
    View in gallery
    Fig. 2.

    Wind and Long Island Sound river input. (a) Daily winds from NARR averaged from 2004 to 2009. (b) The corresponding along-shelf (blue) and across-shelf (green) wind components. For the along-shelf component, positive value represents upwelling-favorable wind, while negative represents downwelling-favorable wind. For the across-shelf component, positive is for onshore wind and negative is for offshore wind. (c) Daily and monthly Long Island Sound river input, estimated from the Connecticut River discharge by dividing the Connecticut River by 0.7 (from USGS).

  • <sc>Fig</sc>. 3.
    View in gallery
    Fig. 3.

    Model configurations. The local-scale ROMS grid (plot every eight grid points) varies in horizontal resolution from 600 m over the RIS and BIS to ~1 km along the boundaries, and the regional-scale ROMS grid (plot every four grid points) is uniform with a resolution of 5 km.

  • <sc>Fig</sc>. 4.
    View in gallery
    Fig. 4.

    Comparison of depth-averaged currents in RIS in spring and summer between the model (blue arrows) and observations (red arrows). The field data are from Ullman and Codiga (2010).

  • <sc>Fig</sc>. 5.
    View in gallery
    Fig. 5.

    Comparison of depth-averaged currents in four seasons between the model (blue arrows) and observations (red arrows) in Front-Resolving Observational Network with Telemetry project between 2000 and 2001 (Codiga 2005; Codiga and Houk 2002; red arrows).

  • <sc>Fig</sc>. 6.
    View in gallery
    Fig. 6.

    Comparison of velocities across a transect close to the Race (from New London, Connecticut, to Orient Point, New York) between the (a) observation and (b) model. Color represents velocities across the Race with positive values for current moving out of LIS and negative value into LIS. Model-simulated salinity along the transect is shown by black contours on the right panel.

  • <sc>Fig</sc>. 7.
    View in gallery
    Fig. 7.

    Comparison of vertical currents in spring between (a) the observations in the FRONT project (Codiga 2005) and (b) model results. Color represents water depth with red representing currents at the top layers and blue at the deep, and arrows represent velocities at the corresponding depth.

  • <sc>Fig</sc>. 8.
    View in gallery
    Fig. 8.

    Time-averaged surface salinity for the case (a) Buoy and (b) RivTides after 100 days’ evolution.

  • <sc>Fig</sc>. 9.
    View in gallery
    Fig. 9.

    Cross-shelf sections of buoyant plume normal to the south shore of Long Island (see Fig. 1) for the cases of Buoy, RivTides, and RivTides_500. Along-shelf velocity for the cases (a) Buoy, (b) RivTides, and (c) RivTides_500; across-shelf velocity for the cases (d) Buoy, (e) RivTides, and (f) RivTides_500. The white lines represent salinity, and the thick gray lines represent the isolines of 0.

  • <sc>Fig</sc>. 10.
    View in gallery
    Fig. 10.

    Cross-shelf section of the downshelf velocities based on (color) the direct model output and (isoline) the thermal wind relation with the bottom velocity approximately set as 0. The thick gray line represents the zero isoline of the direct model output.

  • <sc>Fig</sc>. 11.
    View in gallery
    Fig. 11.

    Momentum terms in the (a) along-shelf and (b) across-shelf directions at 15 km offshore for the cases of RivTides. The thick, solid lines represent pressure gradient; the thick, dashed lines represent Coriolis term; the thin, solid lines represent along-shelf, u-momentum advection; and the thin, dashed lines represent vertical viscosity.

  • <sc>Fig</sc>. 12.
    View in gallery
    Fig. 12.

    Seasonal variability of surface salinity for the experiment Real in (a) winter, (b) spring, (c) summer, and (d) fall.

  • <sc>Fig</sc>. 13.
    View in gallery
    Fig. 13.

    Seasonal variability of the buoyant plume’s vertical structure along section 4 in Fig. 1 for the case Real. White lines represent salinity, and colors represent (left) along-shelf velocity and (right) across-shelf velocity. (a),(e) winter, (b),(f) spring, (c),(g) summer, (d),(h) fall.

  • <sc>Fig</sc>. 14.
    View in gallery
    Fig. 14.

    As in Fig. 13, but for the case without winds’ effect (NoWind).

  • <sc>Fig</sc>. 15.
    View in gallery
    Fig. 15.

    Plane view of surface salinity after 3 days of wind forcing and the Lagrangian trajectories of the drifters released at the locations marked by stars at the first day of wind forcing and simulated with 1-h interval for the cases driven by (a) onshore winds (WndOn), (b) offshore winds (WndOff), (c) upwelling-favorable winds (WndUp), and (d) downwelling-favorable winds (WndDown).

  • <sc>Fig</sc>. 16.
    View in gallery
    Fig. 16.

    Buoyant plume’s vertical structure along section 4 in Fig. 1. Alongshelf velocity for (a) WndOn, (b) WndOff, (c) WndUp, and (d) WndDown; across-shelf velocities for (e) WndOn, (f) WndOff, (g) WndUp, and (h) WndDown.

  • <sc>Fig</sc>. 17.
    View in gallery
    Fig. 17.

    Surface salinity for the cases with (a) weak surface heating (HeatW) and (b) strong surface heating (HeatS) after 2 months of simulation.

  • <sc>Fig</sc>. 18.
    View in gallery
    Fig. 18.

    Buoyant plume’s vertical structure along section 4 in Fig. 1. Along-shelf velocity for the experiment with (a) weak surface heating (HeatW) and (b) strong surface heating (HeatS); across-shelf velocity for the experiment with (c) weak surface heating (HeatW) and (d) strong surface heating (HeatS).

  • <sc>Fig</sc>. 19.
    View in gallery
    Fig. 19.

    As in Fig. 11, but for the case with weak surface heating (HeatW).

  • <sc>Fig</sc>. 20.
    View in gallery
    Fig. 20.

    As in Fig. 11, but for the case with strong surface heating (HeatS).

  • <sc>Fig</sc>. 21.
    View in gallery
    Fig. 21.

    Pressure gradients contributed by (solid lines) sea level variation and (dashed lines) density gradients in the (a) along-shelf direction and (b) across-shelf direction for the cases of (black) RivTides, (blue) HeatW, and (Red) HeatS.

  • <sc>Fig</sc>. 22.
    View in gallery
    Fig. 22.

    River discharge from the Connecticut River (black line) and the freshwater delivery out of LIS (blue line), out of BIS (red line), and into RIS (green line) for the case Real. The freshwater flux out of LIS (dashed blue line) and BIS (dashed red line) for the case NoWind is compared with the case Real.

  • <sc>Fig</sc>. 23.
    View in gallery
    Fig. 23.

    (a) Correlation between the vertically integrated turbulent mixing in LIS and the horizontal freshwater propagation out of LIS without the seasonal trend from the upstream. (b) As in (a), but for BIS.

Fig. 1.

Bathymetry and geographic features around the Rhode Island Sound. Place names: Connecticut River (CR), Long Island Sound (LIS), Rhode Island Sound (RIS), Block Island Sound (BIS), Narragansett Bay (NB), Buzzards Bay (BB), and Vineyard Sound (VS). The thick red line is the section used to describe the buoyant plume to the southwest of BIS, and the thick black lines are the sections referenced in Table 2, as discussed further in section 5.
Fig. 2.

Wind and Long Island Sound river input. (a) Daily winds from NARR averaged from 2004 to 2009. (b) The corresponding along-shelf (blue) and across-shelf (green) wind components. For the along-shelf component, positive value represents upwelling-favorable wind, while negative represents downwelling-favorable wind. For the across-shelf component, positive is for onshore wind and negative is for offshore wind. (c) Daily and monthly Long Island Sound river input, estimated from the Connecticut River discharge by dividing the Connecticut River by 0.7 (from USGS).
Fig. 3.

Model configurations. The local-scale ROMS grid (plot every eight grid points) varies in horizontal resolution from 600 m over the RIS and BIS to ~1 km along the boundaries, and the regional-scale ROMS grid (plot every four grid points) is uniform with a resolution of 5 km.
Fig. 4.

Comparison of depth-averaged currents in RIS in spring and summer between the model (blue arrows) and observations (red arrows). The field data are from Ullman and Codiga (2010).
Fig. 5.

Comparison of depth-averaged currents in four seasons between the model (blue arrows) and observations (red arrows) in Front-Resolving Observational Network with Telemetry project between 2000 and 2001 (Codiga 2005; Codiga and Houk 2002; red arrows).
Fig. 6.

Comparison of velocities across a transect close to the Race (from New London, Connecticut, to Orient Point, New York) between the (a) observation and (b) model. Color represents velocities across the Race with positive values for current moving out of LIS and negative value into LIS. Model-simulated salinity along the transect is shown by black contours on the right panel.
Fig. 7.

Comparison of vertical currents in spring between (a) the observations in the FRONT project (Codiga 2005) and (b) model results. Color represents water depth with red representing currents at the top layers and blue at the deep, and arrows represent velocities at the corresponding depth.
Fig. 8.

Time-averaged surface salinity for the case (a) Buoy and (b) RivTides after 100 days’ evolution.
Fig. 9.

Cross-shelf sections of buoyant plume normal to the south shore of Long Island (see Fig. 1) for the cases of Buoy, RivTides, and RivTides_500. Along-shelf velocity for the cases (a) Buoy, (b) RivTides, and (c) RivTides_500; across-shelf velocity for the cases (d) Buoy, (e) RivTides, and (f) RivTides_500. The white lines represent salinity, and the thick gray lines represent the isolines of 0.
Fig. 10.

Cross-shelf section of the downshelf velocities based on (color) the direct model output and (isoline) the thermal wind relation with the bottom velocity approximately set as 0. The thick gray line represents the zero isoline of the direct model output.
Fig. 11.

Momentum terms in the (a) along-shelf and (b) across-shelf directions at 15 km offshore for the cases of RivTides. The thick, solid lines represent pressure gradient; the thick, dashed lines represent Coriolis term; the thin, solid lines represent along-shelf, u-momentum advection; and the thin, dashed lines represent vertical viscosity.
Fig. 12.

Seasonal variability of surface salinity for the experiment Real in (a) winter, (b) spring, (c) summer, and (d) fall.
Fig. 13.

Seasonal variability of the buoyant plume’s vertical structure along section 4 in Fig. 1 for the case Real. White lines represent salinity, and colors represent (left) along-shelf velocity and (right) across-shelf velocity. (a),(e) winter, (b),(f) spring, (c),(g) summer, (d),(h) fall.
Fig. 14.

As in Fig. 13, but for the case without winds’ effect (NoWind).
Fig. 15.

Plane view of surface salinity after 3 days of wind forcing and the Lagrangian trajectories of the drifters released at the locations marked by stars at the first day of wind forcing and simulated with 1-h interval for the cases driven by (a) onshore winds (WndOn), (b) offshore winds (WndOff), (c) upwelling-favorable winds (WndUp), and (d) downwelling-favorable winds (WndDown).
Fig. 16.

Buoyant plume’s vertical structure along section 4 in Fig. 1. Alongshelf velocity for (a) WndOn, (b) WndOff, (c) WndUp, and (d) WndDown; across-shelf velocities for (e) WndOn, (f) WndOff, (g) WndUp, and (h) WndDown.
Fig. 17.

Surface salinity for the cases with (a) weak surface heating (HeatW) and (b) strong surface heating (HeatS) after 2 months of simulation.
Fig. 18.

Buoyant plume’s vertical structure along section 4 in Fig. 1. Along-shelf velocity for the experiment with (a) weak surface heating (HeatW) and (b) strong surface heating (HeatS); across-shelf velocity for the experiment with (c) weak surface heating (HeatW) and (d) strong surface heating (HeatS).
Fig. 19.

As in Fig. 11, but for the case with weak surface heating (HeatW).
Fig. 20.

As in Fig. 11, but for the case with strong surface heating (HeatS).
Fig. 21.

Pressure gradients contributed by (solid lines) sea level variation and (dashed lines) density gradients in the (a) along-shelf direction and (b) across-shelf direction for the cases of (black) RivTides, (blue) HeatW, and (Red) HeatS.
Fig. 22.

River discharge from the Connecticut River (black line) and the freshwater delivery out of LIS (blue line), out of BIS (red line), and into RIS (green line) for the case Real. The freshwater flux out of LIS (dashed blue line) and BIS (dashed red line) for the case NoWind is compared with the case Real.
Fig. 23.

(a) Correlation between the vertically integrated turbulent mixing in LIS and the horizontal freshwater propagation out of LIS without the seasonal trend from the upstream. (b) As in (a), but for BIS.
All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 331 146 17
PDF Downloads 148 56 11