Ocean2k - Synthesis

The first step of the synthesis towards an Ocean2k review paper is to analyze the publicly available data at the regional scale. The analysis is based on the compilation of the Ocean2k-relevant metadata that have been assembled in the metadatabase by the volunteer group members.

Here is how we would like to organize and move forward.  Please share results figures and notes as soon as you have them assembled.

1. Deadlines and milestones

1.1. March 16 2012 (deadline): organization and regional group leadership choice (email Mike (This email address is being protected from spambots. You need JavaScript enabled to view it. ) with your selection).

1.2. March 30 2012 (milestone): Data compilation and feedback to Ocean2k leads (questions/problems/answers)

1.3. April 13 2012 (milestone): progress update and/or completed first-pass synthesis figures/tables/discussion sent to Ocean2k leads

1.4. April 20 2012 (deadline): submission of first-pass synthesis figures/tables/discussion to Ocean2k leads

1.5. April 27 2012: teleconference with Ocean2k and regional leads to discuss results.

2. Regions

The synthesis is organized by unique region (see map). The regional approach will help us tackle independent parts of the database and will enable us to later compile regional and global syntheses.



2.1. Arctic: north of 70N and land boundaries

2.2. Atlantic: 70N - 35S and land boundaries

2.3. Mediterranean: by land boundaries but including the Northern Red Sea north of 20N

2.4. Pacific: 65N - 40S, land boundaries of the Americas and of the Malay Peninsula, the Indonesian island arc, and Australia.

2.5. Indian: 35S - northern land boundary, 20E and the Malay Peninsula, Indonesian island arc, and Australia

2.6. Southern: south of 35S

3. Synthesis teams

A listing by keywords of suggested group members is found here [download Excel file]. How you then organize within groups is up to you.  Note: we grouped the Atlantic, Arctic and Mediterranean people together in this listing, but please subdivide yourselves into these groups.  As a first order of business, please choose a group leader (1.1).  For those of you who didn't submit keywords, please stand by until the leaders are chosen and then write to the leadership of the group you wish to join.  Please feel free to move to a different group if you want but recognize that we will need to leverage regional expertise to make efficient progress on this first stage of synthesis.

4. SST focus

We suggest synthesis by surface temperature reconstruction as a first pass; but if there is a critical mass of native measurements (e.g.  planktonic foram o18 from many sediment cores), it is possible/useful to composite these in the native variable as well.

5. Documentation

At all stages we need to document choices made in selecting and compositing.  This will enhance transparency and reproducibility and make the process of peer review of our synthesis paper easier. Mandatory documentation information/procedures:

5.1. Document in a way that the documentation is clear and inseparable from the data.  This might be most easily done in a spreadsheet for preliminary organization and synthesis, and will make peer review of an eventual paper straightforward and transparent.

5.2. Names and email addresses of people performing the synthesis.

5.3. Date of metadatabase export (it will undergo continual revision)

5.4. Metadata used for data extraction/synthesis: a list of the proxy data used in the synthesis, using their metadata titles.

5.5. Notes on choices made in combining data (when choices need to be made).

5.6. Signoff: please confirm with your group leader when you've completed an assignment, i.e. "Synthesis data sent to group leader on [date] by [name]."

6. Time notation

We suggest to work in the CE/BCE timescale; converting all BP ages to CE/BCE as appropriate given stated radiometric reference years (e.g. 1950, 2000).  This means that dates for Ocean2k are roughly in the range 0 - 2000 CE.

7. Age models

If an age model is not provided for a given dataset, but may be available directly from the contributing scientist, then we should write and request it.  The age modeled data should be archived at WDCA or PANGAEA so we can include it in our synthesis.  If this cannot be done, we must exclude it from our synthesis per transparency plans. See (13) below for suggestions on contacting potential contributors.

8. Adding datasets

If groups are aware of data that are clearly relevant but are not in the metadatabase, these should be added to the metadatabase. (This may be because they were not archived at WDCA or PANGAEA, or the search algorithm missed datasets.)  Let's try to find them in the public archives, request authors add them to the public archives, so that we can use them. See (13) below for a draft letter to authors.  At the same time, as we will be working on unique subsets of the paleodata metadatabase, let's correct obvious errors and remove non-ocean2k relevant data as we go.  This should be an incidental effort rather than a distraction from the synthesis work.  Let other groups know if error correction, e.g. correction of longitude, puts a new dataset into another group's domain.

9. Reference data

As a point of reference, we suggest using Levitus data where/when appropriate: let's all use the Levitus database accessible via Ingrid (http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NODC/.WOA98/); please ask Mike or Helen if you need help extracting local or region averages, or seasons, etc.

10. Reference time interval

We should use the reference time interval as appropriate to the regional synthesis.  If we use standard choices, then it may be easier to combine regional anomalies later.  Standard choices are: Levitus climatology, 1800-1900CE, 1000-1500CE.

11. Synthesis products

Here is what we would like from each group:
11.1. Standardized anomalies: averaged over the region and forming a time series relative to each of the reference time intervals in (10).
11.2. Same as 9.1, except for anomalies not normalized by variance (i.e. having original units).
11.3. Distributions of data within well-observed time windows that emerge from the synthesis.
11.4. Any other synthesis product that is supported by the particular region or its data collection.

12. Example

This example [download Excel file] form the Southern Ocean shows how one of us (HVM) worked through these steps given realities of the paleometadatabase and making choices as deemed appropriate to arrive at a synthesis product.

Thanks again for your support of this project -- we are looking forward to seeing your results.



13. Possible template for letter to data originators:

I am writing on behalf of the PAGES Ocean 2k group (http://www.pages-igbp.org/workinggroups/ocean2k). We are attempting to reconstruct and analyse global and regional ocean conditions and marine climate for the past two millennia (more details in the web link above).

We feel the dataset from your paper

Bertler N.A.N., P.A. Mayewski, L. Carter. Cold conditions in Antarctica during the Little Ice Age - Implications for abrupt climate change mechanisms. Earth and Planetary Science Letters 308 (2011) 41-51

could be most useful for our synthesis. However one of our key criteria for including datasets is that they are publically available - we are attempting to make the synthesis process as transparent and reproducible as possible. In the first instance we have been searching for records on the WDC Paleoclimatology and Pangaea archives and we have not come across your record. If your data is available in a public archive may you please send us a link so we can include your data? If not, may we please encourage you to archive the data in a public repository (and let us know that you do)? Archiving data in WDC Paleoclimatology is easy (http://www.ncdc.noaa.gov/paleo/contrib.html ) and the folks at WDC Paleo are very helpful.

We hope we are able to include your data in our synthesis as we feel it will be an important contribution.