Potential impact of changes in marine ecosystems from model simulations and observations

1
Potential impact of changes in marine ecosystems
from model simulations and observations

• Scott Doney and Philip Boyd
• IOC CO2 meeting, Paris,
• April 2007

2
Interactions between climate, ecosystems and
ocean biogeochemistry
The role of the oceans in Earth systems
ecology and the effects of climate variability on
the ocean and its ecosystems, can be understood
only by observing, describing, and ultimately
predicting the state of the ocean as a physically
forced ecological and biogeochemical system.
The Paradigm program Rothstein et al. 2006
ECOSYSTEMS Cycling of organic matter
BIODIVERSITY BIOGEOGRAPHY PHYSIOLOGY CIRCULATION
CHANGING CLIMATE Altered upper ocean properties
BIOGEOCHEMISTRY Elemental fluxes
Modified from Legendre and Rivkin (2005)
3
Bottom-up and Top-down controls together define
ecosystem structure
Anthropogenic Effects - Fishing
Foodwebs and grazing exert top-down control
zooplankton
fish
phytoplankton
Some phytoplankton groups can escape grazing and
Can form blooms
nutrients, light, trace elements
- Environmental drivers
4
Different bloom-formers impact a range of oceanic
biogeochemical cycles
(C, Cd, S)
Coccolithophores
(C, Fe, N)
Diatoms
(S, C, P)
(Fe, N ,P)
N-fixers Trichodesmium
Phaeocystis
5
Environmental drivers will influence which
bloom-forming phytoplankton groups become dominant
The role of iron supply on diatom growth rates is
now well-established But what about other
functional groups? N-fixers Fe? and/or P? (Mills
et al., 2005) CO2? (Levitan et al. 2007
Hutchins et al. 2007) N-deficient waters (Deutsch
et al. 2006)
6
THE BIOLOGICAL PUMP a prominent example of how
ecosystem-scale interactions influence ocean
biogeochemistry and global climate
Based on changes in bulk fluxes (i.e., NPP,
surface and deep export fluxes) the surface and
mesopelagic ocean layers are similar in their
attenuation of the surface productivity
signal. Boyd and Trull (2006)
US-JGOFS, (2004)
7
Ecosystem structure a wide range of possible
interactions

• The greater complexity that characterizes
• foodwebs results in more permutations
• regarding the response of foodweb
• components to climate change.
• Differential tolerance to change
• alteration of stocks of each component
• and hence of grazing dynamics
• Such shifts may trigger trophic cascades
• which can restructure the foodweb
• i.e. the cascading effect that a change
• in the size of one population in the web
• has on the populations below it
• As observed in some lake,
• terrestrial and oceanic ecosystems

From Legendre and Rivkin (2005)
8
An example of the importance of susceptibility to
environmental change between ecosystem components
Benthic algal mats
Differential sensitivity to UVB between benthic
algae and herbivores contributes to
counterintuitive increases in algae in habitats
exposed to UVB. Mesocosm experiments
illustrate that predictions of the response of
entire ecosystems to elevated UVB cannot be made
on single trophic-level assessments.
High UVB More algal production
Low UVB Patchy algal production
(Bothwell et al. 1994 Science)
9
How might climate-mediated alteration of foodwebs
impact ocean biogeochemistry?

• The biological pump and faunistic shifts such as
  salps versus krill different faecal pellets and
  export efficiency of carbon to depth
• Ocean acidification calcareous zooplankton
  shell dissolution and/or a shift to
  non-calcareous zooplankton will alter the rain
  ratio
• The Twilight zone is populated by a range of
  zooplankton functional groups herbivores,
  omnivores, carnivores, detrivores some
  vertically migrate, others occupy discrete depth
  strata. Little is known about how floristic
  shifts will alter carbon export (see simple
  models by Burd Jackson (2001)
• Faunistic shifts and other elements observed
  shifts from nanoflagellates to ciliates results
  in increased DMS production in iron-enriched
  waters (SOIREE)

10
Examples from the ocean regime shifts
See summary table, Boyd Doney, (2002)
1998
2000
Equatorial Pacific
Chavez et al. (1999)
Bering Sea
Merico et al. (2004)
11
Modelling approaches Coupled Ocean Atmosphere
Models (COAMs) provide predictions of how oceanic
properties will change following interactions
with the atmosphere, cryosphere etc.
Doney, (2006)
Such shifts can be related to phytoplankton
processes in both low and high latitudes
Boyd and Doney (2002)
12
The latest COAMs can be used to tease apart the
climate change versus. variability signals
This is an important step forward in
understanding the rate of climate change over the
coming decades, and applying it to the design of
perturbation experiments
Boyd et al. (submitted)
13
Global models Computational costs vs. requisite
detail for understanding climate and ocean
ecosystems
At present the greatest detail that can be
readily incorporated into COAMs are
phytoplankton functional groups. Changes in
their global distributions due to environmental
control can be represented
Le Quere et al. (2005)
Coccolithophore distributions (obs top model
bottom)
Moore et al. (2006)
14
Regional and or simpler models pros and cons
Regional models can supply useful information on
high resolution physical transports, but seldom
provide more biological detail than COAMs (but
see N. Sea model).
Simple models provide useful insights into
foodweb responses, but how can this information
be linked to larger scale models
Rothstein et al. (2006)
Legendre Rivkin (2005)
15
Reconciling observations and model validation
What level of detail is required to capture the
interactions between changes in ecosystems and
subsequent shifts in ocean biogeochemistry?
Dynamic Green Ocean Model
UNKNOWN
In some cases there are no data available for
modellers - and thus the foodweb lacks top-down
control
16
Reconciling observations and model validation
what data are out there?
Time-series of krill and salp distributions
Krill distributions spatial correlation with
chlorophyll, and temporal correlation with winter
sea-ice extent Both correlates can be obtained
from satellite remote-sensing and can also be
modelled
Shifts from krill to salps
Atkinson et al. (2004)
17
Observations based on underway sampling
technology This tells the who and the where but
often not the why
Video Plankton Recorder
Continuous Plankton Recorder
Davis and McGillicuddy (2006)
18
Little is known about the interactions between
large, long-lived organisms and their planktonic
food supply.
Smetacek Nicol (2005)
OR
Smetacek et al. (2004)
19
Linking models and observations tackling the
why
Global Change Newsletter 66 (2006)
20
SUMMARY modelling
COAM predictions are based on interactions with
atmosphere, cryosphere etc. The latest COAMs
can address the climate change vs. variability
issue. The highest ecological detail within
COAMs is algal functional groups. Simulation of
foodweb dynamics therefore requires more
specialised models. Linking such data outputs in
larger scale models remains problematic.
21
SUMMARY observations
The observational database for larger plankton is
improving. This provides insights into
environmental control of zooplankton diversity,
or evidence of faunistic shifts. Better
understanding of how the environment controls
zooplankton distributions (such as sea ice
extent, SST) can potentially be linked to global
models. The mechanisms controlling foodweb
structure, trophic interactions, cascading etc.
are relatively poorly understood. We need
these to the link between climate, foodwebs and
ocean biogeochemistry.
22
Talk Outline

• Climate, ecosystems and ocean biogeochemistry
• How bottom-up and top-down controls define
  ecosystem structure
• Examples from the ocean regime shifts
• Modelling approaches climate change vs. natural
  variability
• Global versus simpler models pros and cons
• Reconciling observations and model validation
• Summary and the future