Modelling Marine Ecosystems Mick Follows Dept of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology

http://ocean.mit.edu/~mick/Downloads.html

What is the marine ecosystem? • Food web • Focus on

phytoplankton Bacteria, archaea

Why model marine ecosystems? • To understand and quantify global carbon cycle and relationship to climate

– e.g. how much carbon is sequestered in the ocean due to the formation of sinking organic particles?

• To understand and describe fundamental

ecological controls on marine ecosystem and its evolution – e.g. why do particular species or functional types of phytoplankton occupy particular ocean regions?

Where is phytoplankton biomass in the oceans? • Satellite based observations of

Chlorophyll-a, annual cycle 2005 (MODIS)

Where is phytoplankton biomass in the oceans? • Upwelling regions of tropics and subpolar oceans bring nutrients to surface, sustaining high productivity

Seasonal cycle of “plant pigment” at Georges Bank – in situ observations (Riley, 1946)

Phytoplankton abundance

month

Modelling Marine Phytoplankton Riley (1946)

dP = P  μ−K r −gZ  dt

Modelling Marine Phytoplankton Riley (1946)

dP = P  μ−K r −gZ  dt

P = phytoplankton biomass (mol C m-3)

Modelling Marine Phytoplankton Riley (1946)

dP = P  μ−K r −gZ  dt Growth rate (s-1)

P = phytoplankton biomass (mol C m-3)

Modelling Marine Phytoplankton Riley (1946)

dP = P  μ−K r −gZ  dt Growth rate (s-1)

P = phytoplankton biomass (mol C m-3)

Respiration rate (s-1)

Modelling Marine Phytoplankton Riley (1946)

dP = P  μ−K r −gZ  dt Growth rate (s-1)

P = phytoplankton biomass (mol C m-3)

Respiration rate (s-1)

Grazing g = grazing rate Z = abundance of grazers

Riley’s (1946) model of seasonal cycle at Georges Bank

observed

Phytoplankton abundance,

P

modelled

dP = P  μ−K r −gZ  dt

month

Growth rate, μ , depends on environmental conditions • Information from laboratory cultures

Growth rate (day-1)

Chisholm lab, McCarthy (1981)

Temperature

light intensity

nutrient

Global marine ecosytem model • Embed a model like Riley’s in a description of ocean circulation, temperature, nutrient and light distributions • Capture broads regional and seasonal dynamics • Understand nutrient and light controls

Captures regional and seasonal patterns of biomass and productivity

modelled 0-50m biomass (uM N) Stephanie Dutkiewicz

Model captures observed regional and seasonal patterns

Remote Chl-a observations

modelled biomass

Reveals environmental regulation of primary production

Iron limited

Light limited

Nitrogen limited

Current Question: What regulates phytoplankton “community structure”? coccolithophores (CaCO3 plates)

Prochlorococcus, Synechococcus

4 μm

20 μm

diatoms (Silicate frustule)

0.5 μm Small, buoyant, locally recycled. Inefficient export of organic carbon

Large, blooming, aggregating. Efficient export of organic carbon

Observations of phytoplankton community structure • Pigment observations (Aiken et al, 2000) Diatoms, coccolithophores

Prochlorococcus

Global phytoplankton community structure (January) • Interpretations of remote ocean color observations (Alvain et al, 2006)

January: dominant functional types from SeaWifS (Alvain et al, 2005) red Green yellow blue

- diatoms - Prochlorococcus - Synechococcus-like - includes coccolithophores

What determines phytoplankton community structure? physical and chemical environment

genetics and physiology

competition predation selection

ecosystem structure and function

How can we reflect this in mathematical model? • Seed model with many potentially viable

phytoplankton types (c.f. Riley’s model with only one) • Competition for resources, ability to avoid predation, etc… determine “fitness” of each phytoplankton type • Model “self-organizes” selecting for phytoplankton with “fittest” physiological characteristics Follows,Dutkiewicz, Grant and Chisholm (Science, March 30th 2007)

How can we reflect this in a model? DP j = P j  μ j −k Rj −g j Z  Dt • Represent a large variety of phytoplankton types (index j) • Provide characteristics (μ, KR, g) for each type by random draw from ranges determined in lab • Explicit “natural selection” in model brings to the fore the “fittest” types for model environment Follows, Dutkiewicz, Grant and Chisholm (Science, March 30th 2007)

Modeled ecosystem structure

Characterize dominant types in terms of real-world functional types Prochlorococcus analogs

Synechococcus & small eukaryotes

.

Diatoms

Other large eukaryotes

January: dominant functional types from SeaWifS (Alvain et al, 2005) red green yellow blue

MODEL: January dominant functional types red green blue brown

- diatoms - Prochlorococcus - other “small” - other “large”

- diatoms - Prochlorococcus - Synechococcus-like - haptophytes

What now? • Use model ecosystem as platform to

explore and test ecological theories • Explore relationships between ecosystem, nutrient cycles and climate change

Summary and Outlook • Mathematical and numerical models of marine

ecosystems are developed to help understand marine ecology and biogeochemical cycles • Recent models embrace complexity of ecosystem and help us understand relationship of “community structure” and environment • New, genetically based observations of marine ecosystems provide new opportunities and challenges