Transcript Trait-based phytoplankton models

Trait-based models of phytoplankton
Kyle Edwards
2015 CMORE Microbial Oceanography Course
Modeling phytoplankton: why?
• Central players in ocean ecosystem + biogeochemical processes
• We need models to test whether we can explain the present,
and to predict the future
Modeling phytoplankton communities: why?
• Community structure matters for function
Cell size: export, microbial loop vs. higher trophic levels
Variable cellular stoichiometry
Some cyanobacteria fix N
Modeling phytoplankton communities: why?
• Aggregate responses are different from single-species responses
• How does community diversity scale up to aggregate ecosystem
processes?
Scaling of bulk phytoplankton growth
Niches of individual species
How to make complexity tractable?
1000s of species
Genetic diversity
1000s of genes
How to make complexity tractable?
Key traits (parameters)
Constraints define what traits are
possible
Trait constraints + environmental
conditions
=
Emergent community structure
light, nutrients, temperature,
grazers
functional groups
tradeoffs
allometric scaling
Optimal strategies
Global community patterns
How to make complexity tractable?
• Define the upper envelope of temperature responses
• Let the environment select for the optimal strategy (or coexisting strategies)
• We don’t need to measure the temperature response of every phytoplankter
on the planet
Outline
• Case study: Models and traits for nutrient-limited growth
• Ecological theory for traits  community structure
• Trait diversity and potential constraints
Emergent community structure at the global scale: Mick
Monod model
• How is population growth/dynamics affected by nutrient limitation?
• Growth rate depends on the external concentration of the limiting substrate
growth =
mmax S
K +S
µmax
1.0
Growth rate
0.8
0.6
0.4
K
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
Monod model
Rivkin & Swift 1985,
Marine Biology
Monod model
growth =
1.0
mmax S
mmax
+S
a
Affinity = α = µmax/K
Growth rate
0.8
0.6
0.4
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
Monod model
• Very simple model, very phenomenological
• Two very important traits:
1) Growth under chronic nutrient limitation (affinity)
• stratified, well-lit waters
2) Growth under (transiently) high nutrients (µmax)
• upwelling, large mixing events
Limitations of the Monod model
• Measured & works best under relatively steady nutrients (or slow
change)
Uptake rate = (Growth rate)*(Nutrient per cell)
• Assumes constant stoichiometry
• No luxury uptake of transiently elevated nutrients
• Can be difficult to estimate K
Quota model (Droop model)
• Growth should depend more directly on limiting nutrient in the cell
æ Qmin ö
growth = m¥ ç1÷
Q
è
ø
µ∞
1.0
Growth rate
0.8
0.6
0.4
0.2
Qmin
0.0
0.0
0.2
0.4
0.6
0.8
Internal nutrient concentration
(per cell or per C)
1.0
Quota model
• Growth should depend more directly on limiting nutrient in the cell
æ Qmin ö
growth = m¥ ç1÷
Q
è
ø
1.0
Growth rate
0.8
µmax
0.6
0.4
0.2
Qmin
0.0
0.0
Qmax
0.2
0.4
0.6
0.8
Internal nutrient concentration
(per cell or per C)
1.0
Quota model
Caperon and Meyer 1972,
Deep Sea Research
Quota model
Timmermans et al. 2005,
Journal of Sea Research
Quota model
• Can model flexible stoichiometry
• Can decouple uptake from growth: Michaelis-Menten uptake
Vmax
1.0
Uptake rate
0.8
0.6
0.4
Kuptake
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
Quota model
• Uptake affinity
Uptake affinity = Vmax/Kuptake
1.0
Uptake rate
0.8
0.6
0.4
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
Can model growth under variable nutrient concentration, with luxury
uptake
Chlorella sp.
Cells mL-1
P per cell
(10-15 mol)
Time (d)
Grover 1991, J. Phycol
Quota model
• What are the key traits?
Vmax 1
K uptake Qmin
= specific uptake affinity
Uptake rate, under limitation, relative to
demand
Equivalent to affinity in the Monod model
Quota model
• What are the key traits?
µmax
high nutrients for many generations
Vmax and Qmax high nutrients for <1 to several generations
How traits determine community structure – R* theory
How does nutrient limitation determine community structure?
Start simple: a steady-state system (e.g., permanently stratified systems)
1.0
Growth rate
0.8
0.6
0.4
mortality rate
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
How traits determine community structure – R* theory
1.0
Growth rate
0.8
0.6
0.4
0.2
Initial nutrient
concentration
0.0
0
1
2
3
Nutrient concentration
4
5
How traits determine community structure – R* theory
1.0
Growth rate
0.8
0.6
0.4
0.2
Population increase
draws down nutrient
0.0
0
1
2
3
Nutrient concentration
4
5
How traits determine community structure – R* theory
1.0
When growth = mortality, steady-state
Growth rate
0.8
0.6
0.4
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
How traits determine community structure – R* theory
Nutrient
concentration
Population
Size
R*
Time
How traits determine community structure – R* theory
1.0
Steady-state nutrients = R*
Growth rate
0.8
0.6
0.4
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
How traits determine community structure – R* theory
1.0
R1* R2*
Growth rate
0.8
0.6
0.4
0.2
0.0
0
1
2
3
Nutrient concentration
4
5
Under steady-state nutrient supply
the species with the lowest R* competitively excludes all others
because it draws down nutrients below what other species need to persist
Tilman 1982
Tilman 1982
Which phytoplankton are the best competitors?
1000s of species
Ideally, we won’t have to measure R* for every species + nutrient
What constrains R*?
Which phytoplankton are the best competitors?
Insights from the Quota model
For low mortality:
R* ~
1
K
=
Qmin
specific uptake affinity Vmax
Specific uptake affinity ~ competitive ability, under chronic nutrient
limitation
Which phytoplankton are the best competitors?
To be a better competitor
Increase the ratio of uptake affinity : nutrient content
Cell size
Finkel et al. 2010, JPR
Cell size
Specific
nitrate affinity
(L µmol N-1 d-1)
Cell volume (µm3)
•
specific affinity ~ 1/radius2
•
competitive ability for nitrate varies over 4 orders of magnitude!
Edwards et al. 2012, L&O
Cell size
Specific
nitrate affinity
(L µmol N-1 d-1)
Cell volume (µm3)
• specific affinity ~ 1/radius2
• competitive ability for nitrate varies over 4 orders of magnitude!
scaling relationships greatly simplify model complexity
Cellular composition – ways to reduce Qmin
Reduce iron demand by reducing iron-rich photosynthetic machinery
(Strzepek and Harrison 2004)
Reduce phosphorus demand by using non-phosphorus membrane lipids
(Van Mooy et al. 2009)
• What are the costs / tradeoffs?
• Can we quantify the impact of these decisions on growth, competition,
etc?
Cellular allocation
Major physiological components: chloroplasts, ribosomes, nutrient acquisition
Clark et al. 2013, L&O
Cellular allocation
• How does allocation to rapid growth (ribosomes) relate to ecological
outcomes?
Optimize rapid growth
N:P = 8
Optimize R* for N
N:P = 37
Allocation to biosynthesis
Klausmeier et al. 2004, Nature
Cellular allocation
• How does ecological context select for cellular stoichiometry?
Distribution of N:P across species
Klausmeier et al. 2004, Nature
Cellular allocation
• How does ecological context select for cellular stoichiometry?
Distribution of N:P across species
R*
Redfield µmax
Klausmeier et al. 2004, Nature
Theory for variable nutrient supply
Seasonal stratification, shorter-term events, etc.
Simplest version: tradeoff between R* and µmax
‘Opportunist’
‘Gleaner’
Kremer and Klausmeier 2013, JTB
Theory for variable nutrient supply
Seasonal stratification, shorter-term events, etc.
Simplest version: tradeoff between R* and µmax
‘Opportunist’
‘Gleaner’
Kremer and Klausmeier 2013, JTB
Theory for variable nutrient supply
In general, greater resource fluctuation favors rapid growth
strategy
(live fast die young)
Coexistence of strategies: Large resource pulses + periods of
scarcity
Diatoms
Coscinodiscus wailesii
Ditylum brightwellii
Eucampia zodiacus
Nitzschia closterium
Psuedo-nitzschia pungens
Skeletonema costatum
Asterionellopsis glacialis
Cocco
Emiliania huxleyi
Dinos
Gleaners and opportunists: L4 English Channel time series
Prorocentrum micans
Alexandrium tamarense
Gymnodinium catenatum
Gleaners and opportunists: L4 English Channel time series
•
•
•
- nitrate
- mean PAR in the mixed layer
- algal biovolume
Gleaners and opportunists: L4 English Channel time series
Species vary in their response to nitrate
Edwards et al. 2013, Ecology Letters
Gleaners and opportunists: L4 English Channel time series
Specific nitrate affinity
Species with higher affinity increase in relative abundance as nitrate decreases
Edwards et al. 2013, Ecology Letters
Gleaners and opportunists: L4 English Channel time series
µmax
Species with higher µmax increase in relative abundance
when both irradiance and nitrate are high
Edwards et al. 2013, Ecology Letters
Gleaners and opportunists: L4 English Channel time series
Seasonal succession of opportunists vs. gleaners
• Also good light competitors during winter
Edwards et al. 2013, Ecology Letters
Quota model: more complex
Storage capacity (Qmax) and/or rapid luxury uptake (Vmax)
Favored by regular-ish events on the scale of day-weeks
• Meso/sub-meso
Shorter-term ‘opportunists’
Pulsing begins (7 µmol L-1, twice daily)
Cermeño et al. 2011, MEPS
Cermeño et al. 2011, MEPS
Sommer 1984, L&O
Are there constraints to simplify this?
R* vs. storage capacity
rapid growth vs. storage capacity
Residual P affinity
Residual P affinity
R* vs. rapid growth
Excelling at any one function diminishes the others (multidimensional tradeoff)
Empirical tradeoffs can explain the coexistence of strategies
Mechanistic basis?
Predation is difficult
Many kinds of grazers for many kinds of phytoplankton
Trophic interactions less well developed than nutrients, light, temperature
Size provides some important constraints
Theory for size-structured predation
Specific
nitrate affinity
(L µmol N-1 d-1)
Cell volume (µm3)
Why do large phytoplankton exist? Bad at (nearly) everything.
Theory for size-structured predation
Theory for size-structured predation
Fuchs and Franks 2010
Growth
Nutrient
Nutrient
(R*)
After Armstrong 1994, L&O
Nutrient
(R*)
Not enough nutrient for bigger
to persist
After Armstrong 1994, L&O
Predators eat
Nutrient
Nutrient
After Armstrong 1994, L&O
Top-down control of the phytoplankton
(R* for the grazer)
Nutrient
After Armstrong 1994, L&O
Now a larger species can
persist
Which is too big for
to eat
Nutrient
After Armstrong 1994, L&O
Nutrient
After Armstrong 1994, L&O
Nutrient
After Armstrong 1994, L&O
A size-structured food web
Nutrient
After Armstrong 1994, L&O
A size-structured food web
Nutrient
After Armstrong 1994, L&O
A size-structured food web
Important features:
• Size classes are added as nutrient input increases
• Individual populations experience strong grazing
• But the total phytoplankton biomass is controlled by nutrient input
• Ecosystem co-limitation by grazing and nutrient supply
• Explains the coexistence of large phytoplankton
After Armstrong 1994, L&O
A size-structured food web
Important features:
• Both phytoplankton and zooplankton traits can be constrained by size
After Armstrong 1994, L&O
A size-structured food web
Fuchs and Franks 2010 JPR
Summary
The parameters of phytoplankton growth models are key ecological traits
Tradeoffs and other constraints are essential for parsing community
complexity
These constraints determine how community structure and diversity emerges
under different environmental conditions
Future directions
What are the constraints?
Physiological / genetic basis for trait variation and tradeoffs
• Models of allocation
• Synthesize omics approaches with continuous ecological traits
Better validate trait-based community models
• Lab/mesocosm experiments, distributional data
Interactions: how temperature modulates resource competition
How do aggregate patterns emerge from community complexity?