MARine Ecosystem Models Intercomparison Project (MAREMIP)
Summary of preparation meeting held at the British Antarctic
Survey, Dec 3 2007.
Present: Shubha Sathyendranath, Icarus Allen (on the phone), Erik
Buitenhuis, Laurent Bopp, Meike Vogt, Yasuhiro Yamanaka, and Corinne Le
Quéré (chair).
The goals of MAREMIP are to:
- progress towards the resolution of important scientific questions (see below)
- develop a community around marine ecosystem modellers
- promote the development of targeted observations
- link modellers with observationalists
- document the evolution of marine ecosystem models
Scientific questions to be addressed:
- What are the main physical, chemical and biological processes that determine
the global marine biomass, its regional distribution, its distribution among size classes
and plankton groups, its organisation in biological provinces, its internal and external
fluxes, its variability, and its resilience to change? Are there multiple steady states
in marine ecosystem distribution?
- What are the impacts of global environmental changes on marine ecosystems, including
climate change, changes in nutrient input, and pressure by fisheries?
- What is the feedback from marine ecosystems to climate and to the availability of food resources?
- What observations are we missing in order to further our understanding of marine ecosystems?
Previous MIPs efforts
MAREMIP is building on two parallel efforts: the Ocean Carbon-Cycle Model Intercomparison
Project (OCMIP) and the Carbon Cycle Climate Coupled Model Intercomparison Project
C4MIP. OCMIP ran from about 1996 to 2004, and had four phases:
OCMIP-0 compared the Princeton and Hamburg models. OCMIP-1 compared the Princeton,
Hamburg, IPSL, an Hadley center models. OCMIP-2 had 11 models and a wider range
of observations. OCMIP-3 (NOCES) had 7 models and focused on interannual variability.
C4MIP is ongoing since about 2000. It includes 11 models.
The ocean archive from these models is far
smaller than that of OCMIP. 3 of these models (IPSL, Hamburg, and NCAR)
are being compared as part of the EU Carbo-Ocean project. The current focus
is on primary production, export, and ocean CO2 sink.
MAREMIP will focus on the comparison of global Dynamic Green Ocean Models based
on the representation of multiple Plankton Functional Types (PFTs). We will initially
focus on the 1996-2007 time-period. MAREMIP will start in 2008 and hopefylly
last for about 10 years. Further phases will consider ecosystem changes over
paleo time-scales.
Model evaluation
Models participating in MAREMIP will be compared with a range of observations. The
comparison should ideally consider model efficiency and model bias, following Allen 2007
and Doney et al. (subm. to JMS).
The comparison will include basic properties, such as:
- the annual mean biomass of PFTs, as available in the EUROCEAN's database (during 2008)
- surface chla from satellite data, its regional distribution, and its seasonal and interannual variability
- the frequency of PFT dominance from satellite data, its regional distribution, and its seasonal and interannual variability
- export of carbon at 100 m
- nutrient distribution (for long simulations)
- time series data (including JGOFS and zooplankton time-series)
- pCO2, O2 fluxes
The comparison will also include emergent properties, such as:
- biological provinces
- light saturation parameter
- Chl:C ratio
- zoo/chl vs chl trends
- dynamics properties of flows among PFTs
Simulations
In this first phase, we will focus on the 1996-2007 time period.
Although we do not impose any restrictions, we recommend modelling groups to
initialise their models in 1990 with nutrient observations.
In Phase II, simulations will consider:
- historical: 1996-2007 (with initialisation from observations in 1990)
- historical: 1948-2007 (with initialisation from observations in 1948)
- historical: 1948-2007 (with spin up to equilibrate nutrients)
- climate change scenario (A2 anomalies on top of NCEP)
- climate change scenario (A2 with fully coupled models)
- scenarios of top-down pressure from fisheries (to be determined)
Technical issues:
MIPs can take a lot of peoples' time. We need to be pragmatic in order not
to waste too much time, and focus on things we want to do anyway. With this
in mind, we propose that:
- Everyone stores their own results locally and provide remote access with OpenDAP
- For Phase I, we request results to be interpolated to a regular grid
Funding:
We have currently no common source of funding. There are two possibilities to get the
comparison going in 2008: we applied for support from the EU EUROCEANS project
for 3 months of salary for Meike Vogt to get Phase I going; also members of
the group of Yasuhiro Yamanaka would be willing to volunteer some time already in the
winter. We also have a bit of funding secured to meet in October 2008.
To get the effort going, a small group composed of Meike Vogt, Yasuhiro
Yamanaka and Corinne Le Quéré will meet as soon as the Euroceans's results are out
and decide who does what. If other groups would like to provide human time
in 2008, please let Corinne know so we can co-ordinate the efforts.
For the long term effort, we need to explore future sources of funding.
Corinne will explore the possibility of including EU funding in Greencycles-II project.
Corinne will explore NERC standard or consortium grant.
Travel money from international projects may be available (e.g. GCP, AIMES, IMBER, SOLAS).