Dead-Zones and Coastal Eutrophication: Case- Study of Chesapeake Bay W. M. Kemp University of Maryland CES Horn Point Laboratory Cambridge, MD Presentation to COSEE Trends Orientation at UMCES HPL 4 August 2009
Outline Background Concepts of Eutrophication and Dead-Zones Introduce Features of Chesapeake Bay Describe Hypoxia Patterns in Bay Explain Factors Regulating Hypoxia : Physical Flow, Stratification, Mixing Biological Nutrient Loading Ecological Responses to Hypoxia Concluding Comments Epilogue: Ecosystem Feedbacks and Restoration of Eutrophic Coastal Systems
Eutrophication Effects on Coastal Ecosystems Oligotrophic Eutrophic Stratification Dead Zone
Global Scale of Eutrophication Harmful Algal Blooms Seagrass Loss Hypoxia/Anoxia
Background Information Chesapeake Bay
Key Bay Features Large ratio of watershed to estuarine area (= 14:1) Deep, narrow channel is seasonally stratified Broad shallows flank channel (mean z = 6.5m) Most of Bay volume is in the mainstem Most of its surface area in tributaries & embayments Relatively long water residence time (~ 6 mo)
Watershed Changes: Land-Use & Population Exponential growth in watershed population Population x10 6 16 12 8 4 a) Human Population Land-use shift from forest to farm (thru 1850) to developed (1850 2000) % of Total Land 100 80 60 40 20 0 b) Land-Use Forest Agriculture Developed 1650 1700 1750 1800 1850 1900 1950 2000 Year
Susquehanna River Flow is Large (Flood Flow At Conowingo Dam)
Watershed Changes & Variations: Flow & Fertilizer Large variations in river flow (~4X); wet and dry decades but no long-term trends Fertilizer use in basin has been increasing since 1950, tripling since 1960 Mean Flow m 3 s -1 kg N y -1 x10 6 1600 1200 800 400 40 30 20 10 c) Susquehanna Flow Tropical Storm Agnes Wet Drought d) Nitrogen F ertilizer Use Variable wet 0 1950 1960 1970 1980 1990 2000 Year
Hypoxia Patterns in Chesapeake Bay
Summer Hypoxic Dead-Zone : 1950-2003 Clear increasing trend in volume of severely hypoxic (O 2 < 1 mg/l) from 1950-2003 Within long-term trend, hypoxia is greater in high flow years (wet = green dot) compared to low flow years (dry = red dot) Summer Low DO Volume (10 9 m 3 ) 14 12 10 8 6 4 2 0 1950 DO < 1 mg L -1 Wet Year Dry Year 1960 Tropical Storm Agnes 1970 1980 1993 1989 1992 1990 1998 2003 2001 2000 Years ( Hagy et al 2004)
Location of Bay Hypoxic Zone mg/l 7 6 5 4 0 3 Depth (m) 20 40 7/20/1998 60 0 50 100 150 200 250 km from Susquehanna River 2 1 0 (Hagy 2002)
Seasonal Cycle of Algal Productivity, Temperature and Bottom Dissolved O 2 Hypoxia confined to summer (June-September) Hypoxia coincides with peak temperature and productivity
Spatial Distribution of Bay Hypoxia: 1959 vs. 1995 (low flow) Longitudinal sections of summer dissolved oxygen for two years with similar (low flow) freshwater inputs No anoxic conditions in 1959 but large anoxic (dead) zone in summer of 1995 Upper oxic layer was much deeper in 1959 (10-12 m) compared to 1995 (5-10 m)
Factors Regulating Hypoxia Physical Factors Ecological Factors
Stratification Control of Hypoxia: Position & Intensity of Low O 2 Water Depth (m) Susq. R. MD/VA Atlantic 0 5 10 15 20 25 30 35 40 Pycnocline Anoxic DO declines along landward bottom-layer flow
Vertical Exchange between Upper & Lower Layers 250 Distance from Atlantic (km) 200 150 100 50 cm d -1 180 140 100 60 40 20 Vertical exchange is minimal in mid-bay from May-August Corresponds to location and duration of hypoxia. How does it vary inter-annually? 0 1 2 3 4 5 6 7 8 9 101112 Month (Hagy 2002)
River Flow, Stratification, & Mixing (1986-98) 12 Stratification Stratification strongly correlated with river flow. cm d -1 10-3 rad 2 s -2 10 8 28 24 20 16 Vertical Exchange 1996 Vertical exchange relation to flow is weaker (buoyancy vs. mixing). High flow also delivers more nutrients. 12 1995 8 1000 1500 2000 2500 3000 Winter-Spring River Flow (m 3 s -1 ) (Hagy 2002)
Increasing Nitrogen in Susquehanna River: Seasonal and Long-Term Trends Long-term increases in nitrate levels & changes in seasonality seen over five decades Nov Sep Highest nitrate levels (yellow, red) occur in cold months Jul May Nitrate trends are closely related to total Nitrogen N-loads to Bay doubled from 1945 to 1970 Mar Jan 1945 1955 1965 1975 1985 1995 (Hagy et al 2004)
Hypoxia Response to N Loading (1950-2001): Unexpected Shifts in Ecosystem Processes Hypoxic Volume (10 9 m 3 ) (DO<1 mg l -1 ) 12 10 8 6 4 2 0 1950-1969 1970-1979 1980-1989 1990-2001 10 15 20 25 30 35 Winter-Spring NO - 3 Loading (10 6 kg) Hypoxia increases with N loading. Equivalent N load since 1980 generates more hypoxia than in past. Is system less able to assimilate N-load? No clear explanation for regime shift. (Hagy et al. 2004)
Ecological Consequences of Hypoxia & Dead-Zones
Summary of Faunal Responses to Hypoxia in Mississippi River Plume (Rabalais et al. 2001)
Hypoxia Degrades Habitat for Benthic Fauna in Chesapeake Bay Comparing estuaries worldwide (#1-14), benthic animal abundance tends to be proportional to algal food produced in water Upper and lower Bay generally follow this trend, but hypoxic mid Bay has lower animal biomass than expected Loss of bottom habitat causes loss of important fish and invertebrate animals Macrobenthic Biomass (g AFDW m -2 ) 70 60 50 40 30 20 10 12 14 11 13 10 9 8 7 6 2 5 3 Lower Bay Upper Bay Mid Bay 0 0 100 200 300 400 500 600 700 4 Severe hypoxia Annual Phytoplankton Production (gc m -2 y -1 ) (Hagy 2002, Herman et al. 1999) 1
Degraded Bottom Habitats Cause Loss of Benthic Fauna in Hypoxic Regions of Bay With increasing nutrient enrichment and organic production, depth of sediment oxidized zone declines Fauna shift from diverse large deep-burrowing forms to few small surface-dwellers Sediment Profile Photos Conceptual Model Benthic macrofaunal abundance declines markedly Model derived in part from work of by Don Rhoads in LIS Benthic Community Change Disturbance Gradient (Nilsson and Rosenberg 2000)
Degraded Bottom Habitats Alter Fish Community Structure and Harvest Steady decrease in the proportion of fisheries harvest from bottomdwelling animals General degradation of bottom habitats in shallow (loss of SAV) and deep (hypoxia) waters Similar trends are being reported in other systems worldwide Possible loss of trophic efficiency (fish harvest per unit photosynthesis) (Houde in Kemp et al 2005)
Concluding Comments 1 Coastal eutrophication is a global scale phenomenon Many features of Chesapeake Bay make it susceptible to seasonal development and expansion of hypoxic dead zones Seasonal deep-water hypoxia is generally regulated by stratification and enhanced nutrient loading A dramatic upward shift in the size and intensity of Bay hypoxia occurred in the early 1980s Similar hypoxic regime shifts have been reported elsewhere Hypoxic dead zones result in reduced abundance, diversity and production of benthic invertebrates and demersal fish.
Epilogue: Ecosystem Feedbacks and Restoration of Eutrophic Coastal Ecosystems
Bottom-Water Hypoxia Enhances Recycling of Benthic Nutrients Benthic nutrient (PO 4 & NH 4 ) recycling sustains algal production and hypoxia thru summer Hypoxia causes higher rates nutrient recycling rates Thus, hypoxia promotes more algal growth per nutrient input to the Bay For N & P recycling, same effect of low O 2 but different mechanisms Benthic DIP-Recycling (Boynton in Kemp et al. 2005) DIN-Recycling Efficiency (Cornwell in Kemp et al. 2005)
Seagrass (SAV) Decline: Loss of Particle & Nutrient Trapping Dramatic decline in SAV between 1962 & 1982 throughout the Bay Many factors contributed to decline, but increased nutrients was primary factor Plants (kg m -2 ) 0.3 0.3 0.2 0.2 0.1 Response Nutrient to Nutrient ResponseLoads Plant Epi Phyto 1999 0.1 0.0 Control Low High
Excess Nutrients Inhibit SAV Survival But Healthy Beds are Nutrient Sinks Historical Bay SAV beds were capable of removing ~45% of current N Loading Primary pathways of N removal would be trapping particulate N & direct assimilation Calculation only considers mainstem upper (MD) Bay N removal rates would be larger if whole Bay were considered Percent of Total N-Sources 100 80 60 40 20 0 Atmosphere Diffuse Sources Point Sources N-Sources Particulate Nitrogen Trapping Plant Uptake N-Sinks (SAV Beds) Denitrification ( Kemp et al 2005)
Oyster Decline: Loss of Particle Filtration Decline in oyster abundance has caused loss of nutrient filtration capacity Oyster declines due primarily to overfishing and disease Historic oyster populations were able to filter Bay water volume in days WetMeatWeight(kgx10 6 ) 40 30 20 10 Virginia Maryland Time (days) 700 600 500 400 300 200 100 0 Days to Filter Maryland's Portion of Chesapeake Bay 228 700 3.6 1880 1988 2003 (Newell 1988) Current oyster populations filter Bay water in months-years Oyster restoration would help mitigate eutrophication effects 0 1880 1900 1920 1940 1960 1980 2000 Time (years) (Kemp et al 2005)
Oyster Filtration Effects on Bottom Hypoxia Oyster restoration to meet management mandate (10x), and to estimated precolonial conditions (100x) Dramatic declines in phytoplankton with restoration throughout Bay Small improvements in bottom O 2 with oyster restoration (~ effects of reduced nutrient loading) Phytoplankton Chlorophyll-a Time (days) 700 600 500 400 300 200 100 0 Days to Filter Maryland's Portion of Chesapeake Bay Dissolved O 2 228 700 3.6 1880 1988 2003 Restoration improves water clarity (& SAV cover) 10x restoration ~ 50% effect of 100x restoration mid- Bay Large Trib Small Trib Whole Bay mid- Bay Large Trib Small Trib Whole Bay (Cerco and Noel 2007)
Tidal Marshes Expanding with Soil Erosion, But Contracting with Sea-Level Rise Tidal marshes are important features of Bay watershed Marsh area expanded since colonial times due to increased soil erosion from watershed Marshes have served as buffers filtering nutrient inputs from watershed Marsh area is declining due to sea level rise and reduced soil erosion Pre-Colonial (hypothetical) 1903 1949
Tidal Marshes Serve as Nutrient Filters at Watershed-Estuary Margins Tidal marshes have enormous capacity to filter sediments & nutrients Nitrogen removal capacity measured in experimental marsh ecosystems 80% of N-inputs from land and estuary removed in three year-old marshes Similar effects on N-loading for diverse (brackish) and mono-specific (salt) marshes Marsh restoration would help re-establish lost filtration capacity Nitrogen Removal (% Inputs) Time (years) Brackish (High diversity) Saline (Low diversity)
Ecosystem Feedbacks affect Bay Response to Nutrient Management Positive & negative feedbacks control paths of ecosystem change with Bay degradation Among other mechanisms, N & P inputs affect hypoxia & light Hypoxia leads to more nutrients, more algae, & more hypoxia Turbidity leads to less SAV causing more turbidity, less SAV Oysters & marshes tend to reinforce these feedbacks Processes reverse w/ restoration, thus reinforcing trends (Kemp et al. 2005)
Concluding Comments 2 Dead-zones are an expanding problem that is linked to coastal eutrophication at global scales Although some systems are more susceptible to hypoxia due to inherent physics, anthropogenic nutrient loading is a key driving factor Restoration of Bays and estuaries worldwide requires reduction in nutrient loading to coastal systems Diverse ecological feedback processes complicate Bay restoration Hypoxia stimulates more algal growth thru enhanced nutrient recycling Loss of SAV, tidal marshes and oyster beds causes reduced natural filtration of nutrients from coastal waters Ecological positive feedbacks reinforce both coastal ecosystem degradation and restoration Thresholds and delayed responses may be expected with loading,
Fishery Responses to Eutrophication Stages of Fish respond to nutrient enrichment? (1) (2) (3) First: Fishery production increases with nutrients Second: Fishery does not respond to nutrients Third: Fishery production declines with nutrients (Caddy 1993)
Volume of Summer Hypoxic Water is Related to River flow and Nitrate Loading, with Regime Shift in Early 1980s Volumes of summer hypoxic (O 2 < 1 mg/l) and anoxic (O 2 < 0.5 mg/l) clearly related to winter-spring river flow Hypoxia vs. River Flow Abrupt increase in slope of time trend from 1950-1980 (blue line) to 1980-2003 (magenta line). Currently, there is more hypoxia per unit NO 3 Loading Hypoxia vs. NO 3 Loading What factors have contributed to this abrupt regime shift leading to more hypoxia per loading? Positive feedback mechanisms at work? (Hagy et al 2004, Kemp et al 2005)
Dissolved O 2 in Low-Flow Years Susquehanna River 0 10 20 30 40 50 10 20 30 40 50 1959 1995 Atlantic Ocean mg/l 2.0 1.0 0.7 0.5 0.2 0.0 Major differences between conditions in 1950s and present are -- increased intensity -- seaward expansion. (Hagy 2002)
Dissolved O 2 in High-Flow Years Susquehanna River 0 10 20 30 40 50 0 10 20 30 40 50 1952 1998 Atlantic Ocean mg/l 2.0 1.0 0.7 0.5 0.2 0.0 Major difference between earlier years and present years is -- seaward expansion -- deeper hypoxic area (Hagy 2002)
Oxygen Budget for Mid Bay 6.23 6.88 6.46 0.55 0.94 1.88 2.94 mg l -1 Average Flow (1990) 21% 79% 34% 66% 56% 44% Vertical Diffusion Net Horizontal Advection
Chesapeake Bay System: Watershed area = 116,000 km 2 Water surface area = 11,500 km 2 Land-Use in Watershed: Susquehanna R. (55 % of flow) 58 % 28 % Agriculture Barren Developed Forest Water Wetland
Concluding Thoughts Restoration Implications
Salinity and O 2 Seasonal and Vertical Distributions (Kemp et al. 1992)
Concluding Comments Coastal eutrophication is a global scale problem, and Chesapeake Bay is a system that is inherently susceptible to effects of nutrient enrichment Eutrophication effects first evident 200 years ago, with intense hypoxia and dramatic SAV loss first occurring in the 1950s and 1960s A dramatic upward shift in the hypoxic zone size occurred around 1980, with more hypoxia generated per nutrient loading now compared to past Increased turbidity with eutrophication has caused large reductions in benthic primary production (algal & SAV) Changes in abundance and community composition of demersal fish and benthic invertebrates have occurred in response to bottom habitat losses Human-induced changes of oyster and marshes habitats further stimulate Bay ecosystem response to nutrient enrichment and nutrient abatement Ecological positive feedbacks reinforce both Bay degradation response to nutrient enrichment, and Bay restoration response to nutrient reductions Thresholds and delayed responses may be expected with reduced nutrient loading, but habitat restoration may tend stimulate recovery