U.S. JGOFS Model-Data Policy

Draft U.S. JGOFS Synthesis and Modeling Project Model-Data Policy

July, 1998

Send comments by August 15, 1998 to:

Robert DeConto (deconto@ncar.ucar.edu)
Scott Doney (doney@ncar.ucar.edu)

Objectives

The overall objectives of the U.S. JGOFS Synthesis and Modeling Project (SMP) depend on the synthesis of observational and model generated data. From its inception, U.S. JGOFS has acknowledged the importance and net benefit of open and timely data exchange. The PIs of the field programs (time-series, process studies, and global survey) have followed a data policy whereby observations are released to the public within two years of the end of the cruise via the web-based JGOFS data system and are made available to other JGOFS PIs prior to that time. U.S. JGOFS SMP recognizes the necessity for a model-data policy management strategy that addresses both the goals of the program and the needs of individual PIs. In order to facilitate the distribution of synthetic data products, model code, and model output generated by JGOFS SMP sponsored research, a unified SMP model-data policy has been established. The JGOFS SMP model-data policy is designed to 1) encourage the timely distribution and exchange of relevant synthetic and model generated data sets, 2) promote the ethical use of such data, and 3) insure the scientific repeatability of results via thorough documentation describing the data, models, and modeling strategies (i.e., boundary conditions and model protocols).

Synthetic and Model-Data

Within this document, the phrase "model-data" encompasses the broad range of numerical and synthetic scientific data products developed under SMP sponsored research. They include, but are not limited to, historical data not currently archived or accessible through the JGOFS data system, compilations and synthesis of JGOFS and historical data across multiple field programs, derived quantities and interpolated fields, calculated data products from remote sensing and satellite observations, and numerical model output, code, and forcing data.

Oversight

The U. S. JGOFS Steering committee, with the advise and guidance of the SMP co-coordinators, is responsible for the review and implementation of the SMP model-data policy. All U.S. JGOFS SMP PIs are strongly encouraged to follow SMP model-data policies. The JGOFS steering committee and SMP co-coordinators should be notified of any instances where SMP model-data policies are not being followed. The situation will be reviewed by the steering committee and co-coordinators, upon which the steering committee may recommend a course of action.

Policy

Like the U.S. JGOFS data policy, the U.S. JGOFS SMP model-data policy is based on voluntary model-data exchange for the mutual benefit of SMP PIs and the broader marine biogeochemical community. The U.S. JGOFS SMP model-data policy encourages the voluntary two-way exchange of observational, synthetic, and model generated data among investigators by enforcing the following guidelines:

PIs may distribute their observational data, models, or model output to anyone at any time.
PIs involved in observational synthesis, model development and/or model applications should publish their results and interpretations of those results in a timely fashion. Model-data should be made available to other SMP PIs no later than the time of publication. However, earlier, voluntary data sharing is encouraged. Model-data submitted to the SMP model-data archive, described below, will be available only to SMP PIs for two years after the data are generated. After two years, the data will be made available to the general scientific community as outlined by NSF data policies. PIs will be responsible for maintaining accessibility to the data for an additional three years, once the model-data is made available to the general scientific community.
PIs who acquire model-data or source code from the SMP model-data archive before the model-data has been made available to the general public must contact the originating investigator(s) and are expected to give the originating investigators due credit. It is expected that the extensive use of archived model-data will result in collaboration among investigators and, in some instances, co-authorship of resulting publications.
U.S. JGOFS SMP Model-Data Policy should be consistent with that of sponsoring agencies (i.e., NSF, NASA). In the case of differences between U.S. JGOFS SMP model-data policy and policies of a sponsoring agency, the agency policies will be followed.

Implementation and Model-Data Archive

A U.S. JGOFS SMP model-data archive will be developed to facilitate data acquisition, public distribution, and long-term preservation. The model-data archive will be established within the framework of the existing U.S. JGOFS Data Management System (DMS) and will be implemented by the U.S. JGOFS Data Management Office (DMO) at the Woods Hole Oceanographic Institution (WHOI).

SMP PIs are encouraged to archive all relevant oceanographic, biologic, ecologic, and biogeochemical data generated by numerical models developed and/or implemented with the support of JGOFS SMP sponsorship within the time limits explained above.

While recognizing the right of model-data developers to the first use of the data they generate, SMP encourages the timely sharing of model-data and discourages the restriction of model-data availability in all but exceptional cases. Access of model-generated data will be limited to other SMP PIs for a period of two years after submittal. Passwords will be required for model-data access during the two year "PI only" phase of availability. The model-data will become available to the general scientific community after two years, unless prior arrangements have been made with the JGOFS steering committee and the DMO.

The SMP PI will be requested to submit complete and accurate data and metadata. The PI will be responsible for the final quality of the data on the DMS. The PI will have access to the DMS and should correct any errors recognized after the data have been submitted.

PIs with Unix workstations are encouraged to store and maintain their own model-data, by installing the SMP DMS on their local computers/storage systems. The DMO will support the installation of the DMS and will provide training for PIs if necessary. PIs without workstations or with limited data storage resources may submit their data to the DMO, where the data will be loaded and maintained at WHOI. Access to model-data will be platform independent, allowing access from any networked MAC, PC, or workstation (see the DMS documentation for more detail).

Archived computer model-data should include model source code and a complete description of model protocols and boundary conditions. The model-data should include sufficient explanation so that they can be understood by persons having knowledge typical of SMP PI's. Documentation of model-simulation protocols should be thorough enough to allow other investigators to repeat the model simulations themselves.

The U.S. JGOFS SMP model-data policy recognizes that each modeling effort supported by SMP is unique. Model generated data sets, especially those produced by Ocean General Circulation Models (OGCMs) can be very large, taxing data storage facilities. Therefore, it will be the responsibility of the individual PIs to edit their model output where appropriate, before submitting the data for distribution. Such editing should be based on the PI's judgment of what model-generated fields may be of use to observationalists and/or other modelers. PIs should contact the DMO directly regarding any questions related to model-data format.

The U.S. JGOFS SMP model-data policy borrows concepts from the U.S. JGOFS Data Policy, U.S. Globec Data Policy (Globec Report Number 10), and the CoOP Data Policy.