[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

U.S. OCMIP phase 3



Dear OCMIP participants,

I would like to provide you with an update of our current plans
to ensure the continuation of the U.S. component of OCMIP (OCMIP-3).
These plans are based on the discussions we had during the recent OCMIP 
meeting in Princeton and some additional discussions during the JGOFS SMP 
meeting in Woods Hole. I think we have made great progress in defining why 
we want to continue this project and what the focus of this next phase 
should be. However, there are still some major outstanding issues that 
we need to discuss and decide upon before we are ready to submit a 
proposal. I therefore would like to encourage you to comment on the plans
outlined below.

The current "straw-person" proposal looks as follows:

Overarching Goal: To develop global-scale, 3-D ocean carbon-cycle models
that are capable of predicting the future oceanic uptake of CO2 from the
atmosphere, and in particular the response of the ocean carbon cycle to
global climate change.

Strategy: To improve the predictive capacity and speed the development of
these models through standardized model evaluation, and model
intercomparison. (taken from OCMIP2)
 
Focus: The focus of the next phase will be on the role of biology for the 
partitioning of CO2 between the ocean and the atmosphere.  Or in other
words: "how important is the biological pump?". We will investigate 
the steady-state situation as well as start to assess its role under 
perturbed conditions.  This phase will thus provide the framework for 
future studies on the response of ocean biology to climate variability and
change.

Proposed Simulations: Two sets of simulations have been proposed:
- nutrient depletion simulations: 
	The goal of these simulations is to evaluate how a change in the 
	strength of the biological pump is changing atmospheric CO2.  It builds 
	on the work of Broecker et al.  (1999) and Archer et al.  (submitted, 
	in press?)  who showed that the atmospheric CO2 drawdown obtained by 
	increasing the biological pump to maximum strength varies dramatically 
	from model to model, with the GCMs generally predicting smaller 
	atmospheric CO2 decreases than simple box models.  The evaluation and 
	understanding of the processes that control this "efficiency" of 
	changes in the biological pump are very important to better understand 
	how models respond to ocean fertilization, global warming and also to 
	glacial-interglacial biological pump scenarios.

	These simulations would be based on the OCMIP-2 nutrient restoring 
	model, and the depletion would be performed by 	
	setting the restoring nutrient concentration to zero. It is
	usually very instructive to perform several simulations, where
	the depletion is only done over a limited region (i.e. Southern 
	Ocean vs. high latitudes of the NH, equatorial regions, subtropical
	regions)

- common prognostic biogeochemistry model: 
	The goal of this simulation is to evaluate the sensitivity of the 
	biologically controlled carbon fluxes to different physical circulation 
	models.  It will be particularly interesting to compare and evaluate 
	the results from these simulations with the nutrient restoring 
	simulations, because of the much looser constraints inherent in a 
	prognostic model.  Furthermore, this simulation will form a baseline 
	against which the group's individual ecosystem models can be compared 
	and it will also provide the baseline for running perturbation 
	experiments in the future (global warming and interannual variability).
	
	The big issue we need to resolve is what kind of common prognostic 
	biogeochemistry model we want to implement and run, particularly since 
	none of the existing models stand out as a clear candidate.  I 
	therefore want to initiate some discussion on how we want to tackle 
	this problem:

    - First and foremost, we probably want to define a set of criteria that 
         this model should fulfill:
      - euphotic zone model:
        - what we would love to have:
          from a geochemical perspective: 
             input: Si, N, Fe, (P)
             output: POC, PON, PIC, DOC, DON, Chlorophyll
          from a biological perspective:
             nutrient co-limitation
             at least two size classes, including the
			 differentiation into (at least) diatoms and non-diatoms
      - aphotic zone model:
          what we would love to have:
          from a geochemical perspective: 
             input: fluxes of POC, PIC; conc of DON, DOC, O2
             output: rates of remineralization, dissolution, denitrification
          from a biological perspective:
             particle aggregation, zooplankton feeding, microbial loop
             etc

      - feasibility:	  
		  To my knowledge, currently no such model exists that has been 
		  successfully used in the context of a large-scale GCM. This 
		  raises a serious challenge, since I am not particularly eager to 
		  use OCMIP as a" guinea pig".  I also advocate strongly
                  against the 
		  idea of OCMIP being in charge of developing its own prognostic 
		  biology model!  Therefore, how are we going to respond to this 
		  challenge?
		  
		  There are (at least) two potential strategies: The cautious 
		  approach is that we decide to use a model that doesn't include 
		  everything in the wishlist above, but one that has been tested 
		  already in a global GCM. One such model is that of Olivier 
		  Aumont.  Other candidates?  A second, more audacious approach 
		  would be to wait with the decision for a year, and work with 
		  JGOFS SMP to speed the development of the community model that 
		  JGOFS SMP is aiming for.  In addition, several among us are very 
		  near a breakthrough in their efforts to develop a model with the 
		  above characteristics, and one year might suffice to make it 
		  happen! What is your impression?
		  
       - simulations: The appealing part of these simulations is that
	      they don't necessarily have to be run to a full ocean steady-state
		  (although that would be preferable, so that issues like 
		  interhemispheric transport, etc. can be investigated as well).
		  I suggest that the simulations should be carried out until the
		  net air-sea flux is less than 0.1 PgC/yr (10 times higher limit
		  than OCMIP-2). When starting from the OCMIP-2 conditions, that
		  probably shouldn't take much more than 1000 years.
		  


Analysis Studies:
  We need to develop a strategy how we are going to analyze the
  results, particularly those that will come out from the common
  prognostic biogeochemistry simulations. There are a number of
  standard fields that we probably want to subject to systematic evaluation
  and comparison
  
    - Satellite Chlorophyll 
    - POC, PIC standing stocks
    - NPP and Export production, f-ratios
    - POC, PIC (opal) fluxes
	- seasonal nutrient and pCO2 fields
	- CO2 and oxygen air-sea fluxes

Participation:
- Modeling groups:
    All four U.S. modeling groups that participated in OCMIP-2
    are interested in continuing their participation. In order
    to overcome the limited range of physical models represented
    by MIT, PU, NCAR and LLNL, Scott Doney suggested that we approach
    and invite the Miami group to participate (they just developed
    a new global version of their isopycnal model). Raleigh Hood who is
    currently working with the Atlantic version of this model had the
    same idea and approached me last week during the JGOFS SMP meeting
    and told me that he and Victoria Coles would be very interested
    in participating with the new global Miami model in the next phase of
    OCMIP. In the meantime, they have already contacted Rainer Bleck and
    others in Miami and have received full support. I am very excited
    by this development, since the similarity of the MIT, PU, NCAR and
    LLNL models (particularly with the decision of NCAR to use POP)
    represents a problem. I therefore would like to officially invite
    the Miami Group, Raleigh Hood and Victoria Coles to participate in
    the next phase.

- Data groups:
    I believe (as many others do) that the entrainment of data people
    into OCMIP-2 was very important and contributed to its success
    so far. However, the shift in focus from geophysical problems more 
    towards biological processes in the next phase requires a 
    slightly different set of observations to evaluate our model results
	with (see above).

    Once we have better defined the analysis projects, we should
    start thinking about the people that we like 
    to entrain and to discuss whether it will be necessary to provide
	some funds for them to produce data products specifically for
	OCMIP (as was done during OCMIP-2). (e.g. I don't think
    Sat. Chl. needs support, but we might need to support the flux people
    a bit in order for them to produce gridded fields, possibly even
    in interaction with model results). 

- Data management and support:
    Steve Hankin made a very convincing case during the OCMIP
	meeting that the combination of DODS (Distributed Ocean Data System (?))
	and LAS (Live Access Server) would provide the data support needed
	for U.S. OCMIP. The recommendation to use DODS/LAS is further strengthened
	by the decision of U.S. JGOFS SMP to use the same system. I therefore
	propose to allocate funds for building a U.S. OCMIP LAS server. It 
        is
	very important that this happens in very close collaboration with the
	international data management activities led by Jim Orr in Paris, building
	on their expertise in setting this up and making it work. I suspect
	that the U.S. OCMIP server will act as a mirror site for the data from
	OCMIP-2, and as the primary data storage site for U.S. OCMIP-3 and 
	derived data products, and mirroring it back to Paris (or vice versa, 
	depending on how things evolve in terms of funding).
	
	The situation with regard to the analysis package (GAP) is less clear.
	Do we want to seek funds for this as well or are we happy with the 
	current status of the GAP package? In any case, we probably need to build
	some sort of a data base of analysis scripts, programs, files etc.
	(e.g. ferret scripts, mask files, etc). A moderated user group with
	a good archive might be all we need (a la ferret --> Steve Hankin).
	
	
General Remarks:

Please note that the reason for U.S. OCMIP pushing the agenda right now is 
that we have an excellent funding opportunity coming up this fall.  
However, I gained the impression from the responses given to me during the 
discussion at the OCMIP meeting that the ideas outlined above are in 
line with our international OCMIP partners.  In fact, Yashu Yamanaka was 
probably the person who got us going in the direction currently proposed.  
However, not everybody was present, and I therefore wanted to invite 
you to comment on this current strawperson proposal.  Nothing is set 
in stone!

I am looking forward to your comments and suggestions! 

-- best regards

   Niki

------------------------------------------------------------------
Nicolas Gruber
AOS Program, Princeton University  
Sayre Hall, Forrestal Campus, P.O. Box CN710
Princeton, NJ 08544-0710
Phone: (609) 258-1314  Fax: (609) 258-2850
------------------------------------------------------------------