JOA Data


All files must be downloaded before use. If your browser opens a file, try again using [ctrl/right]-click.

NODC SD2 files must have their extensions changed from ".TXT" to ".sd2".

Java OceanAtlas Data Files

The data sets provided for the UCSD Java OceanAtlas site were put together by Jim Swift in order to provide students and researchers alike some of the best available 'blue water' ocean vertical profile data.

The emphasis in most of the data collection is on the discreet sample ("bottle") data. For some of the cruises and transects we also supply CTD data.

All data are available here in Java OceanAtlas binary format, and many are also available in ASCII. (Note, however, that one can use JOA to export bottle data in ASCII-spreadsheet "WHP-Exchange" format.)

About the Data on This Website

J. Swift

The data sets provided at this website were put together in order to provide Java OceanAtlas users a collection of the best available 'blue water' ocean vertical profile data for education and research use. These data are mostly in Java OceanAtlas binary format, for direct use in the JOA application, but we also supply ASCII versions of some of the data sets. (Note that the JOA application can output/save any JOA binary data file in ASCII format.) We have added value to many of these data by organizing sections, by combining and deleting stations, and by correcting some errors. We also supply a special collection of data files meant to be used, chapter by chapter, with the "DPO JOA Examples" which accompany the Talley et al. textbook, "Descriptive Physical Oceanography".

The collection, "The Best" Vertical Section Data, contains some of the finest-quality vertical section CTD and bottle data yet obtained from the World Ocean. This ever-growing collection began as an effort to prepare 'cleaned' versions of WOCE Hydrographic Program transects, and as the post-WOCE 'CLIVAR' repeats of key WHP transects began to come available, these have been added to the collection. By "cleaned" we mean that where it took several cruises to cover one very long transect, the data were combined, and overlapping or off-transect data were eliminated. In most 'cleaned' bottle data sets, known bad data were eliminated. (The WOCE formats deliberately retain bad data - properly coded - but these bad values confuse students and are of dubious research utility.) Most transects were re-sorted, where needed, with the south-to-north or west-to-east left-right orientation preferred by Swift and many colleagues. Originally "The Best" Vertical Section Data collection was bottle data only, and in JOA binary format only, but over time we are adding CTD data and also ASCII (WHP-Exchange format) versions of many of the cleaned transects. Note that we occasionally replace a data file with a preferred version or add new transects to the collection.

Syd Levitus and colleagues at NOAA have assembled and occasionally re-issued collections of quality-controlled mean temperature, salinity, dissolved oxygen, and nutrients on standard level surfaces on a 1-degree positional grid for the World Ocean, known as World Ocean Atlas data files, typically abbreviated WOAyy, where yy is the two-digit year of publication, e.g., WOA98, WOA05, WOA09. These are extremely valuable and versatile statistical data compilations, though of course the combination of averaging and gridding make WOAyy data different from those from measured profiles. Here we use the multi-year all-year average WOA gridded data files. These are the 'mean ocean' data often used in ocean models. The WOAyy data were originally released by NOAA on a layer-by-layer basis at standard levels. To assemble our data files from these we picked the values at each grid point from all the levels and assembled them into vertical profiles, with data at each WOAyy standard level for which there are reported values. We provide some of the WOAyy data as areal, gridded data at the original 1-degree grid spacing, and most at a sub-grid spacing which provides adequate basin-scale visualization with reduced computer resource requirements. Various collections of these grids and sub-grids are provided. We also provide vertical sections taken from WOAyy at 5-degree intervals of latitude and 10-degree intervals of longitude. Finally, so that JOA users can compare WOAyy data to the non-mean, non-gridded data provided in our other data collections, we also are adding vertical sections taken from the closest WOAyy grid points along the tracks of some of the transects in the "The Best" Vertical Section Data collection.

Our largest collection in terms of sheer number of sections is ca. 2200 mostly pre-WOCE (pre-1990's) vertical sections assembled from the original NODC data files contributing to SIO Professor Joseph Reid's World Ocean data collection. From the many cruises in NODC Archives, and some newer cruises originally from colleagues, over more than two decades Joe sifted for the best available data from each region of the World Ocean (except for the Arctic). His intended purpose was to make maps of water properties (which he used to great effect in his many publications), not vertical sections, and so he cut off stations from a cruise where a better (better for his purposes) cruise overlapped it. He made corrections to many of the data for known offsets (mostly standard seawater batch offsets). Here we went into the NODC files and retrieved the entire cruise data set for each cruise Joe used, assembled the stations into sensible vertical sections, applied the same corrections Joe used, and then we culled many bad data values. The resulting 2200 sections represent the basic data knowledge of the deep ocean basins prior to the 1990s. The data sets all include temperature and salinity, nearly all have dissolved oxygen data, and most have nutrient (NO3, PO4, and/or SiO3) data. The data are available in JOA binary and also in an NODC text data format called "SD2". NODC no longer delivers data in SD2 format, but the JOA application will import/read any of these data files.

Under "Other Data Files" we have a somewhat dated collection of pre-1990s sections as Atlantic, Indian, and Pacific Pre-WOCE Basin Scale Sections, along with sections from the highly-regarded GEOSECS, TTO, and SAVE expeditions. A collection of Arctic Ocean and Nordic Sea data is also provided.

We also supply a special collection of data files meant to be used, chapter by chapter, with the "DPO JOA Examples" which accompany the Talley et al. textbook, "Descriptive Physical Oceanography". These are normally accessed from within the DPO JOA Examples, but we provide them here as an alternate source of the data sets.

Other Data Sources

Java OceanAtlas can read several data formats in addition to JOA binary, such as the old NODC SD2, EPIC netCDF, WOCE 'WHP-Exchange', and spreadsheets (certain caveats apply for spreadsheets).

The CLIVAR and Carbon Hydrographic Data Office (CCHDO; see is a comprehensive 'one-stop shop' for WOCE and many post-WOCE vertical section data sets from the World Ocean. Data from most cruises are available in several formats. The JOA application directly reads the "WHP-Exchange" format data from this web site (data sets with suffixes _hy1.csv, _ct1.csv, and and also the netCDF format data. The CCHDO web site also includes cruise maps, cruise documentation, and information about data formats, data collection, and data processing.

NODC seems to no longer be supplying data in their SD2 text format, but JOA at some point will be updated to read data from the searchable World Ocean listings at (Conversion from the NODC ".csv" output choice to a JOA-compatible spreadsheet format would take some time and experience, though is possible.)

It is straightforward to prepare one's own data for the JOA application, though nearly all first-time attempts involve some trial and error (and frustration). We suggest that when creating an ASCII spreadsheet data file for loading into JOA, one pay very close attention to the details on this subject in the JOA User Guide (and also covered briefly in the DPO JOA Examples). We very highly recommend starting with a small, simple text data file - a minimum file - and then taking what one learns from that into modification and assembly of the data file desired.