The Physical Modulation of Seasonal Hypoxia in Chesapeake Bay
Malcolm Scully
Outline:
1) Background and Motivation
2) Role of Physical Forcing
3) Simplified Modeling Approach
4) Sensitivity Studies
5) Physical Mechanisms and Oxygen Budget
6) Historical Observations of Hypoxia
7) Conclusions
Center for Coastal Physical OceanographyOld Dominion University
CCPO Seminar:Monday, March 28, 2011
Center for Coastal Physical Oceanography
From Chesapeake Bay Program newsletter: http://ian.umces.edu/pdfs/do_letter.pdf
Map of Mean Dissolved Oxygen -- Summer 2005
• Low DO has significant impact on a wide array of biological and ecological processes.
• Large regions of Chesapeake Bay are impacted by hypoxia/anoxia.
• Over $ 3.5 billion was spent on nutrient controls in Chesapeake Bay between 1985-1996 (Butt & Brown, 2000)
• Assessing success/failure of reductions in nutrient loading requires understanding of the physical processes that contribute to the inter-annual variability.
From Chesapeake Bay Program newsletter: http://ian.umces.edu/pdfs/do_letter.pdf
Conceptual Model for Hypoxia in Chesapeake Bay
Physical forcing is thought to play an important role in extent and severity of hypoxia: 1) River Discharge; 2) Temperature; 3) Wind forcing
Seasonal and Inter-Annual Variability in Hypoxic Volume (from CBP data 1984-2009)
Maximum observed
Minimum observed
Data compiled from Murphy et al. (submitted)
Regional Ocean Modeling System (ROMS)
ChesROMS Model GridModel forcing
• Realistic tidal and sub-tidal elevation at ocean boundary
• Realistic surface fluxes from NCEP (heating and winds)
• Observed river discharge for all tributaries.
• Temperature and salinity at ocean boundary from World Ocean Atlas.
• Very simple oxygen model
Oxygen Model
• Oxygen is introduced as an additional model tracer.
• Oxygen consumption (respiration) is constant in time, with depth-dependent vertical distribution.
• No oxygen consumption outside of estuarine portion of model
• No oxygen production.
• Open boundaries = saturation
• Surface flux using wind speed dependent piston velocity following Marino and Howarth, 1993.
• No negative oxygen concentration and no super-saturation.
Depth-dependent Respiration Formulation
€
Flux = k DOsat −DOsurf( )
Surface Oxygen Flux using Piston Velocity:
€
k = 3 e0.25W10
From Marino and Howarth, Estuaries, 1993
Model assumes biology is constant so that the role of physical processes can be isolated!
Comparison with Bottom DO at Chesapeake Bay Program Stations
July 19-21, 2004 August 9-11, 2004
Comparison with Chesapeake Bay Program DataBottom Dissolved Oxygen Concentration (mg/L)
Simple model captures seasonal cycle of hypoxia as well as a more complicated bio-geo-chemical model.
In addition to seasonal cycle, model captures some of the inter-annual variability
485 km3days
476 km3days
707 km3days
Model predicts roughly 50% more hypoxia in 2004 than in 2005, solely due to physical variability.
Variability of Physical Forcing
What is relative importance of different physical forcings in controlling seasonal cycle of hypoxia?
Sensitivity to River Discharge
Sensitivity to Temperature
Sensitivity to Wind
Differences between 2004 and 2005 are almost entirely due to wind forcing
Sensitivity to Summer Wind Magnitude
Average Monthly Wind Speed from Model at PNAS
Wind speed during May-August was increased/decreased by 15%
Changes in average summer wind speed of %15 result in roughly 2-fold change in hypoxic volume.
Base Summer Winds Positive 90°
Negative 90° 180°
Sensitivity to Summer Wind DirectionModeled summer wind direction
Sensitivity to Summer Wind Direction
Along axis winds result in less total hypoxic volume
The Physical Modulation of Seasonal Hypoxia in Chesapeake Bay
Malcolm Scully
Outline:
1) Background and Motivation
2) Role of Physical Forcing
3) Simplified Modeling Approach
4) Sensitivity Studies
5) Physical Mechanisms and Oxygen Budget
6) Historical Observations of Hypoxia
7) Conclusions
Center for Coastal Physical OceanographyOld Dominion University
Center for Coastal Physical Oceanography
CCPO Seminar:Monday, March 28, 2011
Lateral AdvectionLateral AdvectionLateral Advection
Mechanisms for Oxygen “Ventilation”
Direct Vertical Mixing
Along-Channel Advection
detrainment
€
Qd = uin∂A∫ − uout∂A∫[ ] ×O2
July 2004 average bottom Oxygen
Fixed Volume for Budget calculations
Oxygen Budget Calculations
€
∂O2∂t+ u∂O2∂x+ v∂O2∂y+ w
∂O2∂z=∂
∂zKz∂O2∂z− R
Rate of change
Advection: Turbulent mixing
Respirationhorizontal lateral
Integrate all terms over entire volume
Monthly Averaged Sub-Pycnocline Oxygen Budget
Ekman
wind stress
North Wind
x
Ekman
wind stress
South Wind
Ekman
wind stress
West Wind
xEkman
wind stress
East Wind
Response of Chesapeake Bay to Wind Forcing is Strongly Impacted by Rotation
For most of the deep areas of the Bay, the gradient Richardson almost never drops below 0.25 in pycnocline (year round!).
Richardson Number for CPB Station 4.3
Sensitivity to Summer Wind Direction
Along axis winds result in less total hypoxic volume
The Physical Modulation of Seasonal Hypoxia in Chesapeake Bay
Malcolm Scully
Outline:
1) Background and Motivation
2) Role of Physical Forcing
3) Simplified Modeling Approach
4) Sensitivity Studies
5) Physical Mechanisms and Oxygen Budget
6) Historical Observations of Hypoxia
7) Conclusions
Center for Coastal Physical OceanographyOld Dominion University
Center for Coastal Physical Oceanography
CCPO Seminar:Monday, March 28, 2011
Historical Observations of Hypoxia in Chesapeake Bay
Data from Jim Hagy
1) Multiple regression based on estimated nitrogen loading explains relatively small amount of observed variance.
2) Residuals to fit suggest hypoxic volume is increasing, despite recent reductions in estimated nitrogen loading.
3) One interpretation is that Bay is less able to assimilate nutrient inputs because of ecosystem degradation.
Observed hypoxic volume (< 1 mg/L)Regression Model
€
V = β 0 + β1NL
Residual
Duration of Summer Wind
River Discharge
Nitrogen Loading
N NE E SE S SW W NW
< 2 mg/L 0.00 0.08 0.18 -0.49 -0.37 0.04 0.69 0.32 0.16 0.36
< 1 mg/L -0.02 0.04 0.15 -0.48 -0.34 0.03 0.71 0.36 0.24 0.44
< 0.2 mg/L -0.10 -0.08 0.05 -0.42 -0.17 -0.10 0.55 0.30 0.33 0.62
* values in red denote significance at 95% confidence interval
Correlation of Historic Data (Hagy et al.) with Wind Direction
Wind data from Patuxent Naval Air Station (1950--2007)
Multiple Regression based on Nitrogen Loading and Duration of Westerly Winds
When you account for changes in wind direction, residual slope is no longer significant.
€
V = β 0 + β1NL + β 2W%
Observed hypoxic volume (< 1 mg/L)Regression Model Residual
Have the Winds over Chesapeake Bay Changed in recent Decades?
Conclusions
1) A relatively simple model with no biological variability can reasonably account for the seasonal cycle of hypoxia in Chesapeake Bay.
2) Wind speed and direction are the two most important physical variables controlling hypoxia in the Bay.
3) Model results are largely insensitive to variations in river discharge.4) The model suggests that the dominant balance controlling hypoxia is between
respiration and advective processes not vertical mixing.5) During winter months ventilation is dominated by longitudinal advection.6) During the summer months ventilation is greater by lateral advection.7) Because of the width of Chesapeake Bay, the rotational response to wind forcing is
greater for along-channel winds than for across-channel winds.8) Winds from the north enhance the residual circulation, increasing the longitudinal
flux of oxygen into the hypoxic zone.9) However, winds from then north are not common during the summer months and
subtle shifts between south and west winds may play a significant role in the observed inter-annual variability.