Contiguous Low Oxygen Waters Between the Continental Shelf Hypoxia Zone and Nearshore Coastal Waters of Louisiana, USA: Interpreting 30 Years of Profiling Data and Three-Dimensional Ecosystem Modeling

Abstract

The multi-decadal expansion of northern Gulf of Mexico continental shelf hypoxia is a striking example of the adverse effects of anthropogenic nutrient enrichment to coastal oceans. Increased nutrient inputs and widespread shelf hypoxia have resulted in numerous dissolved oxygen (DO) water quality problems in nearshore coastal waters of Louisiana. A large hydrographic dataset compiled from research programs spanning 30 years and the three-dimensional hydrodynamic-biogeochemical model CGEM (Coastal Generalized Ecosystem Model) were integrated to explore the interconnections of low DO waters across the continental shelf to nearshore coastal waters of Louisiana. Cross-shelf vertical profiles showed contiguous low DO bottom waters extending from shelf to coastal waters nearly every year in the 30+ year time series, which were concurrent with strong cross-shelf pycnoclines. A threshold Brunt-Väisälä frequency of 40 cycles h⁻¹ was critical to maintaining cross-shelf sub-pycnocline layers and facilitating formation of a contiguous low DO water mass. Field observations and model simulations identified periods of wind-driven bottom water upwelling lasting between several days to several weeks, resulting in both physical advection of oxygen depleted offshore waters to the nearshore and enhanced nearshore stratification. Both upwelling of low DO bottom waters and in-situ respiration were of sufficient temporal and spatial extent to drive DO below Louisiana’s DO water quality criteria. Basin-wide nutrient management strategies aimed at reducing nutrient inputs and shelf hypoxia remain essential to improving nearshore coastal water quality across the northern Gulf of Mexico.

Graphical Abstract

1. Introduction

Oxygen depletion of coastal and continental shelf ecosystems has been exacerbated by anthropogenic nutrient enrichment. Over the last half century, increased incidences of bottom water hypoxia (dissolved oxygen (DO) <2 mg L⁻¹) have been observed globally in coastal and continental shelf regions, primarily associated with widespread alterations in watershed landscapes and demographics that have accelerated nutrient exports to the coast. ^{1, 2} The consequences of low DO can be dramatic and include short- and/or long-term changes in biological communities, trophic interactions, biogeochemical processes and habitats. ^2–4 Bottom water hypoxia on the Louisiana continental shelf (LCS) in the northern Gulf of Mexico has been a long-term recurring seasonal event. The hypoxic area has ranged from 40 to 22,720 km² and averaged approximately 14,000 km² from 1985 to 2019. ⁵ The average estimated area has increased from ~8,300 km² during 1985–1992 to greater than 15,500 km² from 2007–2019, making it the second largest area of marine hypoxia resulting from eutrophication. ² Interannual variations of LCS hypoxic area co-vary with spring discharge and nutrient concentrations from the Mississippi-Atchafalaya River Basin (MARB), where anthropogenic nutrient inputs have increased 2- to 3-fold during the last 50 years. ^6–8

Although decadal scale changes in the area of hypoxia have been linked primarily to elevated nitrogen loads from the MARB,^8–10 the relationship between localized nearshore coastal hypoxia and onset of widespread shelf hypoxia is difficult to assess over the time frame during which hypoxia occurs. The area and volume of the hypoxic water mass is strongly governed by stratification, winds and circulations patterns which vary greatly through time and space on the LCS. ^11–13 Statistical relationships between wind forcing and hypoxia suggest that wind-driven upwelling events in the northern Gulf of Mexico explain between 16–32% of the interannual variability in hypoxic area. ^{12, 14} Complex simulation models have allowed for further examination of the effects of upwelling favorable winds on hypoxic area due to changes in distribution of salinity gradients and organic rich freshwater plumes. ¹⁵ Upwelling currents may advect low oxygen bottom water from the continental shelf to nearshore coastal waters and barrier island shores,¹⁶ similar to observations of deepwater upwelling events at mid-depths of the shelf. ¹⁷ However, the magnitude and temporal/spatial scales of physical advection of low DO waters as influenced by upwelling-favorable winds common during the summer and localized production/respiration in nearshore waters have not been investigated.

There are significant efforts underway to develop and implement effective strategies to control nutrient loading and reduce Gulf of Mexico hypoxia. ^{18, 19} In the US, the Clean Water Act (CWA) provides Federal authority for protection of water quality within US territorial waters, extending to 12 nautical miles (12 NM, or 22 km) from shore (Fig 1). The CWA also provides for state leadership for water quality protection within their nearshore coastal waters, which extend 3 NM (or 5.6 km) from shore in the state of Louisiana. Thus, understanding the interconnections between hypoxia on the LCS, which extends nearly 100 km offshore, and low DO occurrences within nearshore coastal waters has practical implications in regional watershed nutrient management.

Figure 1.

Station locations corresponding to dissolved oxygen and other hydrographic observations. The eastern shelf study area (solid black outline) and cross shelf transects (A’, A, B, C, and D) used to assess contiguous zones of low dissolved oxygen and the spatial extent of hypoxia are highlighted. LUMCON C transect stations C1 and C6 are highlighted for model analysis. The bottom left inset depicts the CGEM model domain.

For protection of biological integrity of coastal waters against effects of low dissolved oxygen conditions, two DO concentration thresholds are often used: 2 mg L⁻¹ and 5 mg L^-1. ^{20, 21} The former is widely recognized as the operational definition of hypoxia and has been used in annual assessments of hypoxic area on the LCS.²² The latter is established for protection against unacceptable chronic growth effects related to physiology and population of specific biological species, especially for their larval and juvenile life stages. ^{20, 21} Louisiana adopted the 5 mg L⁻¹ DO standard for nearshore coastal waters as a management target for water quality protection. However, the frequency and distribution of waters with DO <5 mg L⁻¹ in coastal Louisiana are rarely reported.

In this work, we combined an analysis of 30 years of hydrographic data from multiple research programs with simulations from a hydrodynamic-biogeochemical model to quantify the frequency, temporal and spatial extent of low DO waters within nearshore Louisiana using thresholds of 2 mg L⁻¹ and 5 mg L^-1. We further assessed whether hypoxia in nearshore waters was contiguous with the LCS hypoxic water mass and the factors controlling it. The large hydrographic dataset compiled herein provided a unique opportunity to investigate the relationship between water column stratification and DO, while simulation results allowed for evaluation of the physical and biological drivers of contiguous hypoxia under varying physical forcing conditions.

2. MATERIALS AND METHODS

Field Observations

The study area was bounded by the Louisiana-Texas border on the west and the Mississippi River delta on the east (Figure 1). Field observations were compiled from LCS surveys from 1985 to 2015 that included vertical hydrographic profiles of DO, temperature and salinity within state waters (<5.6 km) and territorial waters (<22 km). Data sources are described in detail in the supplemental information (Text S1; Table S1 and S2), and include the Louisiana Universities Marine Consortium (LUMCON), National Marine Fisheries Service Southeast Area Monitoring and Assessment Program (SEAMAP), U.S. EPA Gulf Ecology Division’s Gulf Hypoxia Research and Modeling Project and Louisiana Inshore Surveys (EPA-GED), and Louisiana Stimulus for Excellence in Research (LaSER). Data were compiled into a single relational database containing 8,798 vertical profiles from 1,388 unique stations and >191,000 observations of DO, water temperature and salinity.

Data Analysis

Within state and territorial water limits, occurrences of DO <2 mg L⁻¹ and between 2 to 5 mg L⁻¹ were determined from vertical hydrographic profiles. In addition to DO concentration, we calculated Apparent Oxygen Utilization (AOU, mg L⁻¹) to remove temperature and seasonality effects on DO concentrations as

$$AOU=\left[O_2^{\ast }\right]-\left[O_2\right]$$

Where $O_2^{\ast }$ is the oxygen solubility concentration calculated using measured temperature and salinity according to Weiss ²³ and O₂ is the measured oxygen concentration. Water masses were assessed based on concurrent temperature and salinity data, calculated density (sigma-t), and Brunt-Väisälä frequency (N). N (cycles h⁻¹) was calculated as

$$N=\sqrt{\left(\frac{g}{\rho }\right)\left(\frac{\partial \rho }{\partial z}\right)}\text{,}$$

where g is gravitational acceleration (m s⁻²), ρ is potential density (kg m⁻³), and z is depth (m).²⁴ N is a measure of vertical stability of the water column, with greater stability (i.e., water column stratification) a critical pre-condition for development of bottom water hypoxia.²⁵

To assess the contiguity of nearshore coastal hypoxia and the larger area of offshore hypoxia, cross-shelf (i.e., perpendicular to the coast) profiles of DO concentrations were analyzed for the occurrence of hypoxic water masses extending continuously across the LCS. Cross-shelf profiles were developed from EPA-GED, LUMCON and NECOP datasets, whose sampling designs were based on a series of synoptic cross-shelf transects. The LaSER surveys focused on plume waters off the Southwest Pass and utilized a random sampling design, which did not permit cross-shelf profile assessment. Cross-shelf profile analyses were restricted to the eastern portion of the LCS because sampling events in shallow nearshore coastal waters west of Atchafalaya Bay were infrequent. Thus, analyses focused on five transects; A’, A, B, C, and D (Figure 1), which were occupied synoptically during the month of July. The distance between stations occupied along these transects averaged 8.3 km. Analysis of bottom water DO utilized the deepest measurement recorded in the hydrographic profile from all datasets. In instances where LUMCON and NECOP observations contained both CTD and Hydrolab measurements for a given transect and sampling date, data from the deepest hydrographic profile was utilized.

Cross-shelf profiles were generated in Matlab® (MathWorks, Inc., Natick, MA). First, transect bathymetry profiles were created by linearly interpolating (0.1 NM resolution) the maximum reported depth at each transect station. Bathymetric relief for all transects were verified using NOAA digital elevation model bathymetry (http://www.ngdc.noaa.gov/mgg/coastal/crm.html). Then, 1 m, bin-averaged observations and calculated sigma-t and N at each transect station were linearly interpolated to 0.1 m resolution vertical bins spanning from the surface to the bottom. The 0.1 m vertical bins were then linearly interpolated horizontally at 0.2 km intervals across each transect. The interpolated cross-shelf profiles were analyzed to identify instances in which low DO appeared contiguous between nearshore coastal waters and offshore LCS waters. Contiguous was defined here as the 2 mg L⁻¹ or 5 mg L⁻¹ DO isopleth extending uninterrupted from shoreward of the 5.6 km state waters boundary to offshore shelf waters. Given the distances between transect stations and possibility of spatial heterogeneity in bottom water DO, cross-shelf interpolations at 0.2 km could potentially overestimate contiguity of low DO between nearshore coastal and LCS waters. However, adjusting the interpolation scale from 0.2 to 0.9 and 1.9 km had minor effects on the results (data not shown).

Wind forcing affecting current patterns on the shelf were analyzed using timeseries of wind velocities taken from the North American Regional Reanalysis (NARR). ²⁶ Alongshore (west-east) wind velocities were averaged across all NARR model grid cells between approximate latitudes of LUMCON stations C1 (29.056°N) and C6 (28.856°N) at 3-hour intervals. Positive (westerly) winds favor bottom water upwelling towards the northern Gulf coast and negative (easterly) winds favor downwelling away from shore. We calculated 4-day average wind speeds (AWS) to compare physical forcing conditions with observed and modeled data at the state water limit of transect C, approximately co-located with LUMCON station C1. Multi-day average wind forcing was calculated using the best fit correlation between maximum bottom water salinity at station C1 with variations of three to six day AWS (Figure S1).

Physical-Biogeochemical Model

We applied the Coastal Generalized Ecosystem Model (CGEM) to describe the physical and biological controls of contiguous low DO across the study area. CGEM is a 3-dimensional biogeochemical model linked to an implementation of the Navy Coastal Ocean Model (NCOM ²⁷) of the LCS. Open ocean boundary conditions for the model including temperature, salinity, and currents are provided by a regional implement of NCOM named the Intra-Americas Sea Nowcast Forecast System (IASNFS) that covers the Gulf of Mexico and Caribbean Sea. ²⁸ Wind stress in NCOM was applied using the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS). ²⁹ CGEM has a horizontal resolution of 2×2 km and 20 equally spaced sigma layers, and includes state variables of carbon, nutrients, oxygen, phytoplankton and zooplankton. Details of the CGEM model and its formulations are described in Lehrter et al.³⁰ Parameter values used in this implementation of CGEM are the same as described in Jarvis et al.³¹ CGEM model output between 2003–2007 were evaluated to supplement field observations with an emphasis on understanding the mechanisms controlling contiguous low DO formation. Model analysis focused on the timing and spatial extent of stratification strength and persistence, nearshore-offshore advection of low DO waters, and sub-pycnocline oxygen metabolism driving bottom water hypoxia. Additional details on modeled sub-pycnocline respiration rates and calculation of the timeframe for hypoxia formation are available in the supporting information (Text S2 and Figure S3). Simulation results were further used to generate an oxygen budget for LUMCON station C1 to compare timeseries of sub-pycnocline oxygen metabolism with advective oxygen exchange. A description of calculations used in the oxygen budget are also available in Text S3 of the supporting information.

3. RESULTS AND DISCUSSION

Low DO in Nearshore Coastal Waters

Bottom waters on the nearshore shelf frequently exhibited DO <5 mg L⁻¹, with ≥55% of all observations falling below the threshold including 28% with DO <2 mg L⁻¹ and 27% between 2 and 5 mg L⁻¹ (Figure S4). Prior review and analysis of Louisiana’s CWA 303(d) listings of impaired waters, which utilized a much smaller dataset, concluded that DO in coastal waters around the Mississippi River delta and those off the Barataria and Terrebonne coastal bays were often below the state standard.^{32, 33} The analysis presented herein utilized 63% more observations than the previous CWA 303(d) assessment, however, only 7% of all stations were within nearshore coastal water boundaries (Table S1). The disparity between available data within nearshore coastal waters and the broader LCS is due in part to the emphasis on shelf-scale hypoxia and basin-scale nutrient loadings in recent decades. From 1985 to 2015, 684 DO profiles were collected at 93 stations within nearshore coastal waters. Most sampling occurred during summer months (June to August, 52% of total; July,32% of total, Figure S4). Consistent with patterns for shelf-scale hypoxia, low DO in nearshore coastal waters was most frequently observed during the summer, with 67% of DO observations <2 mg L⁻¹ occurring in July and 76% of DO observations <5 mg L⁻¹ observed between June to August.

Bottom water hypoxia generally develops beneath the first significant change in density stratification encountered above the bottom ²⁵ due to suppressed oxygen exchange with the upper water column. Previous observations have established that stratification strength, N, of >40 cycles h⁻¹ inhibits oxygen exchange and facilitates oxygen decline with sufficient sub-pycnocline respiration. ³⁴ Comparison of minimum DO concentrations with the maximum Brunt-Väisälä frequency (N_max) from EPA-GED, LaSER, LUMCON/NECOP, and SEAMAP vertical profiles illustrates the dependence of bottom water hypoxia formation on stratification strength (Figure 2A). Dissolved oxygen ≤2 mg L⁻¹ was absent in nearshore coastal waters when the water column was weakly stratified (N_max <37 cycles h⁻¹; log₁₀ N_max=1.57) and thus sufficiently well ventilated with surface oxygen throughout the water column. AOU in coastal waters remained greater than 5 mg L⁻¹ at N_max >40 cycles h⁻¹ as a result of sub-pycnocline respiration (Figure 2B). Across the shelf, 98.5% of DO observations ≤2 mg L⁻¹ and 90% of observations ≤5 mg L⁻¹ occurred at N_max >40 cycles h⁻¹ (Figure 2C). CGEM produced similar patterns of oxygen dependence on stratification strength across all seasons for both nearshore coastal waters and the broader LCS (Figures 3C and 3D). The dependence of hypoxia development on a critical stratification threshold underscores the importance of model physics for accurate simulation of hypoxia observed in previous studies. ¹¹

Figure 2.

A: Minimum DO versus log₁₀ of the maximum Brunt-Väisälä frequency (N_max) for EPA-GED, LaSER, LUMCON/NECOP, and SEAMAP vertical profiles during 1985–2015. Horizontal dashed lines highlight DO thresholds at 2 and 5 mg L⁻¹. Vertical dashed line indicates N_max = 40 cycles h⁻¹ (log10 N_max = 1.6 cycles h⁻¹). Filled circles represent observations within state waters and open circles represent observations beyond state water limits; B: Daily minimum DO versus log₁₀ of the maximum Brunt-Väisälä frequency (N_max) simulated by CGEM at the same stations as presented in A between 2003–2007; C: Apparent oxygen utilization versus log₁₀ of N_max from vertical profiles for field observations presented in A. Vertical dashed line indicates N_max =40 cycles h⁻¹; D: Apparent oxygen utilization versus log₁₀ of N_max from vertical profiles simulated by CGEM at the same stations presented in C.

Figure 3.

A: Percent of cross shelf transects with contiguous zones of DO <2 mg L⁻¹ (black bar) and <5 mg L⁻¹ (gray bar) from nearshore coastal waters (5.6 KM) to the 22 KM territorial sea boundary. The number above each bar represents the annual sample size; B: Representative cross shelf profiles along transect C showing a contiguous zone of low DO (top panel), a large zone of low DO offshore of the coastal water boundary (middle panel), and a non-contiguous, patchy zone of low DO (bottom panel). Black triangles represent station locations. White contours outline isopycnals N > 40 cycles h⁻¹.

Incidence of Contiguous Low DO

Persistent and contiguous low DO bottom waters from nearshore to the continental shelf were observed nearly every year in the 30-year time series, predominately but not exclusively during summer. Cross-shelf profiles indicated that low DO observed within nearshore coastal waters was almost always contiguous, extending beyond territorial waters of the LCS (Figure 3A). Specifically, 141 of 167 (84%) cross-shelf transects exhibited contiguous zones of DO <5 mg L⁻¹, while 36 of 88 transects (41%) contained contiguous low DO <2 mg L^-1. Of the remaining transects, low DO was either restricted to LCS waters outside state boundaries or were non-contiguous across the transect (Figure 3B). During July, 43% of all cross-shelf transects had contiguous zones of DO <2 mg L⁻¹, whereas 98% of the cross-shelf transects had contiguous zones of DO <5 mg L^-1. Observations of contiguous low DO were predominately associated with a strong cross shelf pycnocline along the entire transect with N_max >40 cycles h⁻¹ largely parallel to the bottom (Figure 3B).

Time series of monthly cross-shelf profiles at transect C (located at approximately 90.4°W) from 2000 to 2008 illustrate the temporal and spatial dynamics of bottom water DO and water column stratification (Figure 4A). In this region of the LCS, hypoxic bottom waters were observed more than 75% of the time during midsummer shelf-wide surveys.^{5, 35} Extensive cross-shelf low DO was observed every year from 2000 to 2008 and was contiguous across coastal waters in all years except 2007, in which stormy pre-cruise conditions may have mixed shallow nearshore waters. A closer examination of the 2003–2005 time series shows that water column stratification across the LCS-to-coastal water continuum sometimes persisted over winter, such as in the winter of 2004/2005 (Figure 4B). In other years, strongly stratified waters developed in late winter/early spring on the outer shelf, preceding bottom water DO depletion below 5 mg L^-1. Observations of DO <5 mg L⁻¹ in coastal waters were nearly always preceded by the onset of stratification and subsequent oxygen depletion in the outer LCS, typically occurring around late winter to early spring. Expansion of the offshore pycnocline and low DO bottom waters into the nearshore coastal zone was most frequently observed in late spring to early summer. These data indicate that reduced ventilation caused by strongly stratified waters established the physical conditions conducive to DO depletion in continental shelf and nearshore coastal waters.

Figure 4.

Top: Time series of monthly cross-shelf profiles of bottom water DO along transect C from 2000 to 2008; Bottom: Time series of monthly cross-shelf profiles of bottom water DO and contours of isopycnals N > 40 cycles h⁻¹ along transect C from 2003 to 2005. Triangles correspond to dates of transect sampling.

DO Dynamics During Upwelling Favorable Conditions

The seasonal dynamics of freshwater discharge, surface heating, wind stress and currents interact to establish broad regions of high vertical water column stability in this area of the northern Gulf.^{36, 37} Onset of the strong cross-shelf pycnocline observed throughout the LCS is in part driven by a seasonal shift to westerly winds during late spring/summer,^{15, 38, 39} promoting upwelling favorable conditions and shoaling of isopycnals which transport bottom waters from the shelf into the shallow nearshore coastal waters of Louisiana. Evidence of summer upwelling conditions in the nearshore coastal zone consistent with observations of contiguous low DO across the LCS were revealed in both the hydrographic data and CGEM model simulations. Comparison of monthly observations from inshore stations on transect C indicated that salinity in nearshore coastal waters exhibiting contiguous DO <2 mg L⁻¹ (31.43 ± 2.56 psu, mean + standard deviation) was approximately 12% higher than salinity observed under well-oxygenated (>5 mg L⁻¹) conditions (27.73 ± 3.28 psu), a pattern reflected in the critical relationship between hypoxia and a minimum N_max threshold (Figure 2). Cross-shelf profiles during these periods revealed a uniform layer of high salinity, low temperature bottom water extending from nearshore coastal waters to the LCS, consistent with formation of a strong cross shelf pycnocline and contiguous low DO across the shelf.

CGEM simulations at station C1, located near the state water limit (Figure 1), further identified sustained periods of low DO and hypoxia coinciding with higher bottom layer salinity and N_max ≥40 cycles h⁻¹, similar to mid-summer field observations (Figure 5). Influx of offshore bottom water and onset of low DO resulted from sustained periods of upwelling favorable winds lasting ~3.5 to ≥7 weeks. Northward bottom currents at the nearshore state water limit provides further evidence that upwelling of offshore bottom waters occurs within as little as one day following onset of westerly winds (Figure 5). Wind-driven advection is responsible for much of the cross-shelf displacement of the mid-summer hypoxic plume during both upwelling and downwelling favorable conditions. ⁴⁰ For example, the 2020 shelf-wide hypoxia cruise observed a reduced hypoxic area due to both storm driven mixing and persistent downwelling favorable winds which forced the bottom hypoxic layer deeper offshore. ⁵ Conversely, upwelling favorable conditions due to persistent westerly winds can advect the hypoxic bottom layer nearshore, ^{16, 41} as was observed in 1998 and 2009, resulting in a reduced hypoxic area but similar hypoxic volume. ⁵ Simulated timeseries at station C1 exhibited rapid responses in both bottom salinity and DO to physical forcing conditions, changing as much as 8 psu and 4 mg L⁻¹, respectively, within hours to days of reversals in wind direction and currents (Figure 5).

Figure 5.

Time series of physical drivers and hydrographic conditions for station C1 during summer 2003 and 2004. Top panel: 3-hourly NARR west-east wind velocities (thin black line) presented with a 40-hr low-pass filter (thick black line). Periods of consistent upwelling favorable winds (westerly wind, positive velocity) are highlighted in gray. Middle panel: Modeled bottom layer DO (red lines) are plotted with LUMCON measured bottom DO (red circles). Black lines represent northerly (positive, upwelling) and southerly (negative, downwelling) bottom layer current velocity simulated by the model. Bottom panel: Modeled bottom layer salinity (orange lines) plotted with LUMCON measured bottom salinity (orange circles). Blue lines indicate the maximum water column stratification strength (N_max).

Snapshots from CGEM simulations of cross shelf profiles at transect C prior to, during, and after onset of upwelling favorable winds in 2003 highlight the dynamic nature of oxygen gradients between nearshore state waters and offshore LCS (Figure 6). Strong easterly winds beginning in June and lasting through July 15^th (Figure 5) produced persistent downwelling currents, isolating low oxygen and hypoxia offshore of station C1 (Figure 6). Approximately two weeks later, following transition to consistent westerly winds, upwelling conditions forced advection of high salinity, low oxygen bottom waters from offshore to within state waters. Upwelling conditions persisted until winds again reversed, downwelling bottom waters, dissipating nearshore stratification, and isolating low oxygen offshore. Further evidence of upwelling to the nearshore during this period is apparent in simulated timeseries of temperature and salinity at station C1, where high salinity, lower temperature bottom waters coincide with reduced oxygen concentrations for a period of approximately two weeks (Figure S5). Following reversal of bottom currents offshore, higher temperature, lower salinity bottom waters return, increasing oxygen concentrations above 5 mg L^-1. These simulation results are consistent with continuous measurements of tidally advected hypoxic water masses recorded in the nearshore. ⁴¹ Although nearshore oxygen concentrations in our simulations appear coupled to offshore oxygen dynamics via advection during upwelling favorable conditions, distribution of low oxygen waters were also sensitive to the depth of the pycnocline and its relative height above the bottom. Wind reversals resulting in downwelling of bottom waters from the nearshore typically resulted in a deepening of the pycnocline, such that stratification was eliminated within state waters as the pycnocline intersected the bottom further offshore (Figure 6, bottom panel). These simulation results are consistent with observations from LUMCON transects, where low DO bottom waters restricted to the offshore LCS generally displayed a deeper pycnocline that intersected the bottom at depths ≥10 m, i.e. the approximate maximum depth of the state water boundary (Figure 3).

Figure 6.

Snapshots of modeled cross shelf profiles of transect C in 2003. White vertical line represents the state water limit. Triangles represent the location of LUMCON stations C1 and C6. White contours outline isopycnals of N > 40 cycles h⁻¹. Black lines indicate low oxygen waters less than 5 mg L⁻¹ (dashed line) and 2 mg L⁻¹ (solid black line). Current velocities are presented in green as northerly (positive and onshore) and tan as southerly (negative and offshore).

Field observations and model simulations suggested that sustained upwelling favorable winds lasting as little as four days (4-day AWS) had a significant effect on nearshore bottom water salinity, DO, and stratification gradients. Positive 4-day AWS (westerly wind) resulted in higher bottom water salinity, N_max ≥ 40 cycles h⁻¹, and lower DO and hypoxia in both measured data from station C1 between 1985–2015 as well as the entire 5-year model run (Figure 7). Nearshore stratification gradients were almost always greater than 40 cycles h⁻¹ following positive 4-day AWS, and were most commonly associated with hypoxia observed throughout the summer. Conversely, transition to easterly winds resulted in higher DO and lower salinity, frequently eliminating stratification in nearshore waters by physical mixing or advection of bottom waters offshore. While previous studies have documented the effects of upwelling favorable winds on the hypoxic area across annual and seasonal cycles ^{13, 15, 18}, the combined 30-year dataset and multi-year simulation applied here further demonstrate that periods of sustained westerly winds can alter the distribution of hypoxia between the nearshore and offshore LCS on daily to weekly timescales.

Figure 7.

4-day AWS versus bottom salinity, bottom DO, and N_max. LUMCON observations (black triangles) include all data between 1985–2015 at station C1. CGEM data (grey circles) include daily output at station C1 for summer between 2003–2007. Horizontal dashed lines highlight the DO threshold at 2 mg L⁻¹ (middle panel; B) and N_max = 40 cycles h⁻¹ (bottom panel; C).

While low DO in the nearshore coastal zone appears in part coupled to summer upwelling of oxygen depleted bottom waters formed on the outer LCS, the establishment of a relatively homogeneous sub-pycnocline water mass observed during periods of contiguous low DO facilitates uniform oxygen metabolism across large areas of the shelf. The formation of a widespread stratification envelope, which is clearly visible in the time series (Figure 4B) and cross shelf profiles (Figure 3B), limits ventilation of sub-pycnocline waters from the nearshore to outer LCS.^{25, 37} Given a uniform layer of bottom water within this stratified area, rapid deoxygenation due to sub-pycnocline respiration also facilitates persistent and widespread hypoxia. CGEM simulated nearshore respiration rates capable of reducing oxygen concentrations from saturation to hypoxia in as little as 4–8 days under non-mixed conditions (Figure S3), coherent with the time scales of changes in local wind stress creating upwelling or downwelling conditions. Simulated sub-pycnocline metabolism at station C1 was net heterotrophic throughout the summer, contributing 37–46% of oxygen loss during the upwelling periods identified in 2003 and 2004 (Figure 8). Oxygen loss due to advective oxygen exchange varied with changing winds, resulting in short term (one to four day) fluctuations in advected oxygen loss between 2–5 times greater than losses due to in-situ oxygen metabolism. In comparison, sustained periods of upwelling winds facilitated advective loss of oxygen in the first few days after wind reversal, followed by reduced oxygen exchange as the contiguous hypoxic plume was established. Cross shelf (north-south) advection of the hypoxic plume between state water limits is evident with changes in wind direction, as seen in the period between July 24^th and August 11^th, 2004 (Figure 8). Simulation results further identified periods of enhanced oxygen loss due to along-shelf (east-west) advective oxygen exchange, suggesting spatial heterogeneity and lateral advection of hypoxia within state waters of comparable magnitude to cross-shelf advection. Observations in this region of the shelf similarly suggest both along-shelf and cross-shelf currents may affect bottom DO conditions during upwelling and downwelling cycles. ^{35, 42} These simulation results highlight the complex nature of in-situ metabolism and advective exchange in the nearshore environment, both of which are capable of inducing hypoxic conditions on time scales of a few days to a week provided adequate stratification.

Figure 8.

Daily timeseries of the simulated sub-pycnocline change in DO (ΔDO_sp) from net metabolism and horizontal advection. Net metabolism (yellow bars) includes sub-pycnocline primary production, respiration, and sediment DO consumption. Horizontally advected sub-pycnocline DO exchange are separated into east-west (blue lines) and north-south (red lines) components. Grey highlighted regions indicate periods of sustained upwelling presented in Figure 5. Arrows indicate wind reversals favoring upwelling conditions outside of the sustained upwelling period. Mean net metabolism, east-west ΔDO_sp advection, and north-south ΔDO_sp advection are annotated for the periods of pre-upwelling, upwelling, and post-upwelling favorable wind conditions.

Implications for Nutrient Management

The Mississippi River/Gulf of Mexico Watershed Nutrient Task Force has extended its management goal to reduce hypoxic zone area to 5,000 km² (five year running average) by 2035, an approximate 64% decrease from the average area of 13,950 km² from 1985–2019. Progress towards implementing comprehensive nutrient reduction strategies to improve water quality and reduce the size of the Gulf hypoxic zone continues to be a primary focus of the interagency Hypoxia Task Force.⁴³ Ongoing efforts to elucidate nutrient and organic matter sources ^{44, 45} and model the physical and biogeochemical processes regulating oxygen, carbon and nutrient dynamics on the LCS, ^{15, 31, 46} remain critical to establishing defensible water quality targets. Although a decrease in the area of shelf-wide hypoxia in response to sustained nutrient reductions may not be apparent for several years after reduction targets are met, ^{8, 12} the interrelationships between hypoxia in nearshore coastal waters and the LCS suggest that reductions in shelf hypoxic area may improve DO water quality conditions in the nearshore. Although respiration rates in state waters are sufficient to deplete and maintain DO below the state standard of 5 mg L⁻¹, upwelling of low DO, high salinity bottom waters may further augment net oxygen loss and strengthen stratification gradients necessary to perpetuate nearshore hypoxia. The effects of upwelling driven stratification on hypoxia are projected to intensify in the future due to increases in surface heat flux and enhanced river discharge, resulting in stronger stratification gradients and lower oxygen concentrations across the shelf. ⁴⁷ Wind forcing and upwelling/downwelling conditions may also affect nearshore-offshore oxygen dynamics via changes in the spatial distribution of organic matter delivered via the offshore freshwater plume, potentially increasing nearshore oxygen loss via enhanced respiration. ¹⁵ Considering the frequency and extent of contiguous low DO observed between nearshore and offshore shelf waters, nutrient management strategies focusing exclusively on the nearshore coast are likely insufficient to improve oxygen conditions. Alternatively, a more comprehensive basin-wide nutrient management approach focused on reducing shelf-wide hypoxia may be essential to improving water quality across northern Gulf coastal waters.