Transcript Hurtado_OWU_ScienceSeminar_28mar2013x

Hypoxia in Lake Erie’s Central Basin:
How Annual Variation In Temperature,
Dissolved Oxygen Affect Fishes
Paul J Hurtado (MBI; OSU), Yuan Lou (OSU), Elizabeth Marschall (OSU),
Kevin Pangle (CMU) & Stuart Ludsin (OSU)
Population Density 2005
Gridded Population of the World, v3 (http://beta.sedac.ciesin.columbia.edu/data/collection/gpw-v3)
Based on Hadley Centre HadCM3 climate model
http://www.globalwarmingart.com/wiki/File:Global_Warming_Predictions_Map_jpg
Warming
Hypoxia
Global phenomenon: 400+ marine systems & large inland lakes
Hypoxia
Diaz and Rosenberg (2008) Science
Hypoxia effects on fish?
– Direct mortality
• Higher for sessile benthic organisms than mobile organisms
• Mobile pelagic species can be susceptible
Atlantic Menhaden
(Narragansett Bay)
(www.geo.brown.edu)
An “island” of
dead menhaden
Gulf Menhaden
(N. Gulf of Mexico)
(www.leeric.lsu.edu)
– Sub-lethal effects → more likely for mobile species
• Caused by reduced access to optimal temperature, prey or refugia
 Small
 Shallow
 Warm
 Productive
Volume = 484 km3
Depth = 24 - 60m
> 200 Frost Free Days
All Trophic Levels
Canada
Lake Erie hypoxia
Central
USA
East
West
Warm epilimnion
Cool hypolimnion
(becomes hypoxic)
High
Cool
Temperature
Productivity
Low
Excess Nutrients
Hypoxia
Bottom hypoxia is typically also augmented by eutrophication…
Nutrients
(N, P)
Phytoplankton
Bacteria
Max.
Hypoxic (< 2 mg/l)
Area = 10,000 km2
Central Lake Erie Food Web
Piscivorous
Fish
Walleye
(www.buckeyeangler.com)
Yellow Perch
(©Shedd Aquarium)
Benthivorous
Fish
Planktivorous
Fish
Emerald shiner
(www.cnr.vt.edu)
Rainbow smelt
(nas.er.usgs.gov)
White Perch
(www.cnr.vt.edu)
Benthic
Macroinverts.
Zooplankton
Brandt et al.
Modeling Approach
• Simple model: bring together behavior, ecology,
physiology/bioenergetics, environmental data.
– Physical environment (Temperature, DO, Climate)
– Bioenergetics (growth rate, habitat preference)
– Mortality & sub-lethal effects (body condition)
– General enough to model multiple fish species
Warming + Hypoxia
Hypoxia negatively affects smelt by modifying habitat
Warm
Normoxic
conditions
Cool
Dark
Thermocline
Rainbow
Smelt
Hypoxic
conditions
Hypoxic
Warming + Hypoxia
Hypoxia negatively affects smelt by modifying habitat.
Consequences?
 access to optimal
(cool) temperatures
 access to prey
Warm
Cool
Dark
Normal
 condition (health, weight)
Hypoxic
Model Foundations
• Fish movement, quality & survival affected by:
– Temp, DO
– Food
– Predation risk
Warm
Cool
Dark
Normal
– Population impact?
– Fish body condition?
– Climate variation?
– Disease risk?
Hypoxic
• Goals:
Physical Environment (1987-2005)
Temperature
Rucinski et al, 2010
Physical Environment (1987-2005)
Rucinski et al, 2010
Annual Variation (1987-2005)
Bioenergetics Model
Fish Mass
ZP prey
Oxygen
Growth Rate
(Gi)
Temperature
1.0
0.8
0.6
0.4
0.2
0.0
Field
data
0
2
4
6
8
10
Zooplankton biomass (mg/l)
Consumption scalar
Proportion of
maximum consumption
• Bioenergetics parameters: Arend et al 2011, Lantry & Stewart 1990
1.0
0.8
0.6
0.4
0.2
Bartell (1990)
0.0
0
2
4
6
8
10
DIssolved oxygen (mg/l)
Bioenergetics Model
• Growth rate given by
C = Biomass consumption rate,
F = Egestion, U = Excretion, R = Respiration
• Ex:
Mortality
• Mortality from predation, low DO, high T, other:
• Predation (by Walleye) depends on light, attack
rates, response rates, success rates.
• Hypoxia:
• Temperature:
•
Movement Model
Central Basin of Lake Erie, ignore horizontal space. Water column divided
into 24 “patches” each roughly 1 meter deep. Ni = # fish in patch i
N1
N1
N2
N2
N3
N…
N24
Leave based on
“patch” quality
Redistribute
N3
N…
N24
Movement Model
Central Basin of Lake Erie, ignore horizontal space. Water column divided
into 24 “patches” each roughly 1 meter deep. Ni = # fish in patch i
N1
N1
N2
N2
N3
N…
N24
∑mξiNi
mфi(N)
N3
N…
N24
Improving the Movement Model
• Movement based on Q = G/μ can lead to
tolerance of terribly high mortality rates!
• Solution? Stimulus ≠ Response! Q = g(G)f(1/μ)
f(1/μ)
1/μ
1/μ
1/μ
Improving the Movement Model
• Movement based on Q = G/μ can lead to
tolerance of terribly high mortality rates!
• Solution? Stimulus ≠ Response! Q = g(G)f(1/μ)
• What about g(G)?
– Predator encounters (stochastic) inhibit foraging.
– Solution: use μ to discount ideal growth rate G
g(G, μ) = G exp(-λ(μ) h)
λ(μ) = attack rate, h = displacement duration.
Improving the Movement Model
• Movement based on Q = G/μ can lead to
tolerance of terribly high mortality rates!
• Solution? Stimulus ≠ Response! Q = g(G)f(1/μ)
• Better solution!
– Sensitivity h should be high during hard times,
moderate to low during good times.
– Solution: For time t, define qi = Qi/max(Qi). Then
ξi = exp(-h Qi)  ξi = exp(-h/max(Qi) Qi)
ξi = exp(-hqi)
Model
Central Basin of Lake Erie, ignoring horizontal space. A closed system, with
the water column divided into 24 “patches” each roughly 1 meter deep.
Annual Variation (1987-2005)
Remarks
Water warming and environmental hypoxia …
…can have strong, negative impacts on pelagic fish:
– Survival (direct) and growth (indirect)
– Increased aggregation/density.
This approach provides a unique and promising avenue to better quantify
these effects.
Micro-scale (<1m) environmental factors likely have big effects!
Mortality/survival is a complex and poorly understood piece of the puzzle.
But what about…
• Fish with other natural histories? Trophic interactions?
• Disease?
Model + Infectious Disease
S = # susceptible, I = # infected/infectious, R = # recovered/immune
Infection Risk Index
Infection Risk Index
Low Hypoxia (1999)
High Hypoxia (2001)
Remarks
• Hypoxia induced aggregation can increase
disease risk.
• Stress effects appear to be less significant.
• Aquatic parasites:
– more complex life histories than terrestrial
parasites?
– macro- versus micro-?
– When does “stress + symbiosis = parasitism”?
In Conclusion
• Not a good time to be a cold-loving species…
• Human Induced Rapid Environmental Change
(HIREC) is a complex cornerstone of “our” future.
• Bridging ecology, evolution, physiology, physical
processes, climate change, epidemiology, etc.
…requires REALLY good data!
• Simple (and not-so-simple) models, combined with
data, are increasingly essential research tools.
Questions?