Big Picture Perspective
Per NOAA Administrative Order 212-15 and the Data Management Plan Procedural Directive, each new data acquisition project needs a Data Management Plan that identifies what data will be collected, who and where data will be collected, and where and when data will be archived. NCEI can accept data at the immediate end of the data collection process, at the end of the qa/qc process, when a “final version” of data are released to the public, or when the data collecting office no longer wants or is able to keep data locally.
Some key points to keep in mind when managing data:
- Many Program and Project Offices provide good examples and guidance for managing observation data and metadata for marine scientists. While each program, project, and organization may have specific requirements for documenting, organizing, archiving and providing access to data funded by them, some fundamentals are common across the data management spectrum:
- Plan for managing and documenting data collected at the beginning of the project, not the end
- Designate a person to be responsible for managing data and information for the project
- Need assistance with developing a Data Management Plan? Take a look at some of the resources below for good recommendations.
- To get an idea as to how NCEI ingests and archives data in an automated way, see Example Archive Procedures for examples.
Selecting the Data to be Archived
For the intentions of this procedure, we want to focus on region owned/managed observational data assets that are less than 10 TB per year in size. Use the following guidelines to determine which data sets should be included in the archive procedures:
- If the data is already being submitted to a Data Assembly Center (DAC) that has an archive agreement with NCEI, verify your data will be incorporated into that archival process.
- If the data will not be archived through a DAC, proceed with the steps below.
Below is a diagram of the pathway for the various types of IOOS RA data to get to the NCEI archive:
(click for full-size image)
Formatting Your Data
Whenever possible, data should be recorded or translated to scientific units (instead of raw sensor voltages), and be the best, science-quality version of these data available at the time of submission. Ancillary information that are critical for accurate interpretation and reuse of the data, such as calibration information or temperature and pH scale for sensor measurements, should be noted in the data files themselves or in associated documentation files. In order to fully document the data that has been collected, NCEI provides the following recommendations:
- Use a consistent and unique file naming convention for each file. (e.g.
carocoops.cap2.buoy_2014_03_28_05.nc
)- See the Cookbook FAQ on recommended file naming conventions for more details.
- NCEI highly recommends formatting your data following the NCEI NetCDF Templates v2.0. These best practices capture NCEI’s experience in providing long-term preservation, scientific quality control, product development, and multiple data re-use beyond its original intent.
- Data should be in compliance with the IOOS compliance checker, specifically the NCEI plugin for the IOOS compliance checker.
- To see how compliant your files are, use the Online IOOS Compliance Checker.
- For time-series station data, NCEI recommends using the timeSeries Orthogonal template. For other types of data follow the NODC NetCDF decision tree.
- See Available NetCDF Tools for some examples on how to generate, visualize, and use netCDF files.
Providing Data Integrity
The mission of NCEI is to “acquire, process, preserve, and disseminate oceanographic data.” Thus, the acquisition of data from data providers is an important part of the NCEI mission. In order to maintain the integrity and to guarantee availability of the data that NCEI acquires, as well as to ensure the security of both NCEI and remote computer systems, NCEI has a few recommendations for standard practice for data acquisition:
- NCEI uses a ‘manifest’ to list the data files to be archived, and provide checksums of the files in the package to validate the transfer was successful.
- The relative path to the data files should be described in the manifest. If the data files are in a directory structure below the manifest file, we need to know where the data files are in relation to the manifest. If no relative path to the data files is described, NCEI will assume that the data and manifest are in the same directory.
- There are two options when determining how you want to generate your manifest files:
- Generate the manifest files on a per file basis.
- One text file for each netCDF file, which contains the checksum followed by a space and the netCDF file name. See the example at NCEI Acquisition Standards.
- The naming convention for this file follows the same name as the netCDF file followed by
.sha
, indicating it is a checksum file. - NCEI has been using this construct for the current archive automations with GLOS and SECOORA:
- For example, the manifest for the netCDF file
enp.wiwf1.met_2015_06_01_18.nc
would beenp.wiwf1.met_2015_06_01_18.nc.md5
(using an md5 cryptographic hash value). - NCEI’s current best practice is to use the SHA-2 family of crytographic hash functions (sha256 or sha384). But, other hash functions can be used as applicable.
- For example, the manifest for the netCDF file
- Generate one manifest file with all filenames and checksums for the package to be archived.
- This can be a space delimited file with the filenames, including path to the files, and checksums.
- This can be an xml formatted file with the filenames, including path to the files, and checksums. For an xml formatted example see this file.
- Generate the manifest files on a per file basis.
- Review standard practices to guarantee data integrity for more information on manifest files and integrity checks.
Organizing Your Data
Well-organized data promotes the use and re-use of the interested data set throughout the community. Organizing your data into logical structures that the data lend themselves to be housed in is highly recommended (don’t put a square peg in a round hole). Since you, the data provider, have the best handle on how stakeholders and users alike prefer to have the data organized, we default to your recommendations.
NCEI has seen a variety of ways for data managers to format their data. Below are some points to assist in your data management:
- Organize your data in a way that would make the most sense to a future user of these data.
- Organize your data into a consistent logical structure.
- Organize your data how you would like the data represented in the archive.
- NCEI recommends to organize your non-federal buoy assets by station and providing new files for each month, if possible.
- This allows NCEI to generate Archival Information Packages (AIPs) based on station and a higher level of granularity.
-
Some examples of data organization are as follows
- Organizing your data through Directory Structure (One Directory per Station/Buoy):
- Example Directory 1
- Example Directory 2
- Organizing your data through File Naming Convention (One Directory for all Files, NCEI parses based on name):
- File Naming Convention
- Organizing your data through packaging (zip, tar, gzip, etc):
- Packaging
Determine the Web Tool for Documenting and Submitting Your Data
Currently, NCEI has two web tools to collect information about data sets that are intended to be archived within the NCEI archive. The two systems facilitate the submission of data through the development of metadata records. For one-off submissions (data collections that are not episodically or periodically recurring transactions) the Send2NCEI web tool is recommended for the collection of metadata and transfer of data to the NCEI archive. For automated submissions (data transfer will occur on a consistent basis; weekly, monthly, semi-annually, annually, episodic, etc.) the Advanced Tracking and Resource tool for Archive Collections (ATRAC) web tool is recommended for the collection of metadata and details of the transfer mechanisms that will be developed.
Below are some examples to assist in determining which web tool is appropriate for the data that is intended to be archived.
- Send2NCEI:
-
- A Principal Investigator (PI) who has collected data from multiple CTD casts in the North Pacific.
- An aggregation of data collected from one platform over a year.
- Hard copies of historical data found stowed away.
- ATRAC:
-
- Buoy collecting continuous data which can be exported on an episodic/periodic occurrence.
- An automated system that aggregates and exports data on an episodic/periodic basis.
- A Data Assembly Center (DAC) which can aggregate data and submit on an episodic/periodic basis.
Documenting Your Data
Use the appropriate system, Send2NCEI or ATRAC to begin documenting the data intended to be archived. The information collected within the two systems does vary slightly, due to the intention of the request for archiving. However, the following information would be beneficial to have available when populating the appropriate system.
- First and last dates of data collection;
- Sea areas where data were collected (e.g., Gulf of Maine)
- Western and eastern longitudinal boundaries and northern and southern latitudinal boundaries within which the data were collected;
- Instrument types used to collect the data (e.g., CTD or XBT);
- Ocean properties measured (e.g., salinity, temperature, current direction, or dissolved oxygen);
- Organizations that collected the data (e.g., NOAA/NMFS/Northeast Fisheries Science Center and Woods Hole Oceanographic Institution);
- Who was involved in processing, documenting and organizing data;
- Platforms from which the data were collected (e.g., moored buoy and NOAA ship HENRY B. BIGELOW); and
- Projects that funded the collection of these data (e.g., World Ocean Circulation Experiment).
Developing a Record in the Appropriate Web Tool
The Advanced Tracking and Resource tool for Archive Collections (ATRAC) and the Send2NCEI web tool facilitate the collection of metadata related to the packages you have developed according to the procedures above. Once you have collected the pertinent information listed below, it is recommended to have that information available when populating the appropriate system:
- An idea of the size of the data files (both individually and as a collection).
- The frequency you would like to be archiving (weekly, monthly, yearly).
- Accessible examples of your data files (though a WAF, DAP or other service).
- Examples of your ‘Manifest’ files.
For more information on populating the ATRAC record, see ATRAC Guidance. To learn more about ATRAC, see Introduction to ATRAC.
For more information on populating the Send2NCEI record, contact NCEI.
Posting the Data to be Archived
For NCEI to be able to develop the appropriate metadata and archive the data, the provider should provide a location where NCEI can gain access to the data. While developing the location where the files will be posted for transfer to the NCEI archive, we recommend the following best practices:
- Post your consistently formatted data files and directory structure to a location (FTP, HTTP, DAP, THREDDS, etc) where NCEI can pull the data from.
- Allow ample time for NCEI to develop the ingest and archival procedures.
- Verify that the location where the files are posted will not vary throughout the archival process.
- Anonymous FTP sites are okay for NCEI to pull data from, we just need the appropriate information to gain access.
- Provide access to the files for approximately one month after the data has been ingested and archived at NCEI.
Monitoring Archival Procedures
Once the archival process has been developed and implemented, some time should be devoted to verifying the defined process has been implemented correctly. Typically this is accomplished through the development of an example metadata record and example Archival Information Package. Once the example is approved, we recommend the following checks to verify the process is functioning correctly:
- Search the Geoportal for your archived data collection. The metadata records could take up to one-day to be indexed in the geoportal, so allow a little bit of leeway.
- Check at least one of the Archival Information Packages to verify no errors have occurred.
Approximately one month after you have been notified that a package has been archived, you have the flexibility to remove the archived packages from your access site. NCEI now has your data archived!
Access Archived Data
- The NCEI accession number is a ‘unique tracking number’ for each collection of data from a data provider. The NCEI Accession Tracking Data Base (ATDB) is a database used by NCEI to keep track of archival collections of original data and provide discovery metadata about those collections through the NCEI Ocean Archive System (OAS).
- Use the NCEI accession number to search the Ocean Archive System or to reference a specific NCEI accession number with this URL:
http://data.nodc.noaa.gov/accession/#######
(where ####### is the NCEI Accession number; multiple NCEI Accession numbers may be used in one URL, separated by commas, e.g., #######,#######,#######) - To search through the Geoportal use the following Geoportal Tips Page to assist in developing the appropriate query.
- All IOOS Regional Association and Data Assembly Center data will be tagged with the following metadata key: “Integrated Ocean Observing System Data Assembly Centers Data Stewardship Program” (The link directs you to the pre-queried Geoportal search page for the project “Integrated Ocean Observing System Data Assembly Centers Data Stewardship Program”).