GO-SHIP Easy Ocean

GO-SHIP Gridded Time Series; user friendly WOCE/CLIVAR/GO-SHIP data from CCHDO. Product is available at doi:10.7942/GOSHIP-EasyOcean.

Information for IAPSO Standard Seawater batch-to-batch offset is found in SaltBatchOffset directory.

Reference

K. Katsumata, S. G. Purkey, R. Cowley, B. M. Sloyan, S. C. Diggs, T. S. Moore II, L. D. Talley, J. H. Swift, GO-SHIP Easy Ocean: Gridded ship-based hydrographic section of temperature, salinity, and dissolved oxygen (2022), Scientific Data, doi:10.1038/s41597-022-01212-w, also available in PDF. This publication describes version 1.4.

Output Formats

Uninterpolated data (station data) are called reported data. Horizontally interpolated and vertically smoothed data are called gridded data. These are stored under separate directories. Remarks common to all formats are;

• Temperature is in ITS-90. Use t90tot68.m for conversion to IPTS-68 (e.g. input to gamma surface calculation). The unit recorded in Matlab Station header (e.g. D_pr(1).Station) is the unit of the original CTD data, not the unit in our product.
• Unless otherwise noted (e.g. I01), only good (as defined by flag=2) data are used. This behaviour can be changed by modifying the QC section in read_ctd_exchange.m.
• Missing value is -999 for ASCII and binary outputs and NaN otherwise.
• Vertical coordinate is in pressure. Assuming Depth and/or Corrected depth in the Exchange CTD or SUM is in meters, we convert them to pressure (in reported_data.m). If depth is missing, we assume the bottom of measurement is 10 dbar above seabed. If Uncorrected depth is available but Corrected depth is missing, we use the former.
• Used v3.06 of TEOS-10 to calculate Conservative Temperature and Absolute Salinity.
• Dissolved oxygen concentration is converted to μmol/kg (micro mol per kilogram), often written umol/kg.
• No horizontal interpolation is applied for stations more than 2 degrees apart. This behaviour can be modified by MAX_SEPARATION paramter in configuration_yyyy.m files.

0. Quick start

To visualize a section, e.g. P16 section occupied in 2015, use;

application	reported	gridded
Matlab	reported/P16/p16.mat	gridded/P16/p16.mat
Ocean Data View	reported/P16/p16_2015_ct1.zip	-
Java Ocean Atlas	reported/P16/p16_2015_ct1.zip	-
GrADS	-	gridded/P16/p16.bin.ctl
GMT*	-	gridded/P16/p16_2015.xyz.gz
binary	-	gridded/P16/p16.bin
ASCII	reported/P16/p16_2015_ct1.zip	gridded/P16/p16_2015.xyz.gz
NetCDF	(work in progress)	gridded/P16/p16.nc

* see GMTplotDiff.sh for example.

1. Reported data

This is the clean data with no horizontal interpolation and no vertical interpolation. We support Matlab format and ASCII CSV in WHP Exchange format. Note: we do not perform any additional quality control except for obvious cases listed below. Bad data in the original data set remain in the product.

1.1 All stations

It is possible to include all stations in the original CTD file in the reported data set, e.g., for float calibration purposes. When calling reported_data.m, use a special file name 'all' in the first argument instead of the list file (e.g. P16/p16_1992.list). Note that this output cannot be gridded because of possible duplication and branching of the station tracks.

1.2 Matlab format

D_r is an array holding one Matlab structure for one occupation of the hydrographic section. In typical cases, D_r(1) is by WOCE cruises in the 1980s and 1990s and D_r(2) is by CLIVAR/GO-SHIP cruises.

D_r(1) = struct('Station', {stnW(1), stnW(3), ..} ...
                'lonlist', lon(:), ...
                'latlist', lat(:), ...
                'deplist', depth(:), ...
                'CTDprs', ctdprs(:,:),  ...
                'CTDsal', ctdsal(:,:),  ...
                'CTDtem', ctdtem(:,:),  ...
                'CTDoxy', ctdoxy(:,:)))

Lat/Lon, depths,... etc. can be extracted from stations (see below), but for ease of access, latlist, lonlist, and deplist are provided by reported_data.m. For zonal sections in the Atlantic Ocean, lonlist uses negative longitudes. CTDsal, CTDtem, and CTDoxy are aggregations of station measurements with the pressure given by CTDprs. The entry 'Station' holds a cell list of Station data structure defined as

stations(23) = struct('EXPO', '320620140320', ...
                      'Stnnbr',   '45',        ...
                      'Cast',        1,       ...
                      'Lat',   -45.0002,      ...
                      'Lon',  -149.5998,      ...
                      'Time', datenum(2014,4,17,19,47,0), ...
                      'Depth',     5350,      ...         % in pressure [dbar]
                      'CTDtemUnit', 'ITS-90', ...
                      'CTDsalUnit', 'PSS-78', ...
                      'CTDoxyUnit', 'umol/kg')

• Some Stnnbr has letters (e.g. X12 for cross points) so that Stnnbr is not restricted to numbers but characters are accepted.
• In Lat and Lon, use decimal degree, not degree-minute-second.
• Time follows MATLAB convention with fixed seconds=0.

1.3 ASCII format

One zipped archive corresponds to one occupation of the hydrographic section. In the archive, there are CSV files, one file for one CTD station. Each file has a header showing DATE, LONGITUDE, etc. Note Creation Stamp is dummy.

When unzipped, the data are in ASCII and can be easily edited by your favourite editors. It is in CSV so that spreadsheet program can handle them. In particular, these can be read by Ocean Data View withImport → WOCE Formats → WHP CTD (exchange format) menu. They can also be read by Java Ocean Atlas with File → Open menu.

2. Gridded data

Users interested in data intercomparison should not use gridded data. See Secton 3.1 below.

2.1. Matlab format

Data are stored in 2 dimensional matrices as entries to a structure, one for each occupation. The axes are defined in ll_grid (longitude or latitude) and pr_grid (pressure). The structure has a field NTime which is the time (in Matlab format) of the measurement at the station nearest (in lat/lon) to the grid point.

2.2. Binary format

Binary is in IEEE754, 4-byte float in Big Endian. The first datum is southmost/westmost shallowest temperature datum. The second is the shallowest datum from the next horizontal grid. After XDEF data, data from the second shallowest depth follow. Vertical number of data is YDEF. After temperature, the following data are stored in the order; salinity, oxygen, Conservative Temperature, and Absolute Salinity.

This information and grid lat/lon are found in the form of GrADS control file placed in the output directory (e.g. P16.bin.ctl for P16). One could actually use GrADS to visualize the data, but there is a caveat; GrADS is designed to visualize horizontally collected data (i.e. lat/lon) and not sectional data (i.e. lat/depth). For this reason, the horizontal coordinate (lat/lon) appears as longitude and depth as latitude. Occupation appears in time coordinate. Hopefully this is not a problem for other applications.

2.3 ASCII format

When unzipped, the first line shows the content of the data. The missing value is -999. An example for the use of these ASCII data to plot the difference between occupations with GMT can be found in GMTplotDiff.sh.

2.4 NetCDF format

The NetCDF is CF compliant (CF-1.7 ACDD-1.3). The netCDF files are dimensioned by gridded_section (an integer indicating the gridded section number), longitude, latitude, and depth. time is a time variable associated with each data point, rounded to the day closest to the date of the data used at each grid point.

3. Caveats

3.1 Coordinate for tracklines

When gridding reported data, a choice has to be made as to which horizontal coordinate to use -- longitude, latitude, distance from the first station, staion number, etc. We chose longitude/latitude based on the approximate distance (see sort_stations.m. If the distance between the northernmost and southermost stations are larger than the distance between the easternmost and westernost stations, the the section is meridional. If the other way arond, it is zonal.

Problems arise when the secton has oblique trajectories. For example, P17 section extends from (55°N, 200°E) to (40°N, 225°E) -- southeastward then turns southward until (60°S, 225°E) with a slight turn to avoid islands in the South Pacific. This will be gridded in latitude from 60°S to 55°N. When gridded, the stations north of 40°N appear on the same meridian of 225°E but actually not. The northmost station is more than 1000 km west of the meridian. Those users interested in data intercomparison should only use reported data, where all option might be useful (see Section 1.1.

3.2 Choice of 'occupation'

The definition of occupation or repeat on a section is rather subjective. For example, the cruise 325020080826 on P16 was regarded too short to be an occupation and was not included here, but the choice was subjective. The user is free to make their own choices of cruises to form an occupation.

The cruise 31TUUNES_2 was along 209.5°E, about 0.5° west of later occupations. We included this as part of the 1992 occupation of P16 but it is quite possible to see this section as separate from P16. We did not have any objective standard when deciding which cruise to include. The user is again welcome to make the choice.

3.3 Section names

As far as we know, the section names have been only roughly designated by WOCE or GO-SHIP. For example, the section A01 is often referred to as section AR07. It is also not unusual to cover two or more sections in one cruise, ending up the cuise having mutiple section designators (e.g. 316N145_9 for I04, I05, and I07). The best place to find this information is README.md in SaltBatchOffset/. The table shows a section name along with "also known as" entry. Try this table if you cannot find the section you are looking for.

3.4 Choice of sections

Those data not archived on CCHDO are not included in this product, which does not mean all data available there are included here. Archived data which do not follow the recommended format were included if the format can be converted to Matlab readable format in a minimal effort. We welcome volunteers who could convert those missed cruises into the format. used here.

3.5 Oxygen data

Some cruises did not provide oxygen data. If oxygen field is missing in reported or gridded product, it is most likely because oxygen was not provided in the source data. If you notice oxygen in the source but not in the product, please notify us.

3.6 Bad data

We do not perform any quality control on top of the WOCE flags provided in the source data.

For some obvious cases only, however, did we manually removed suspect data even with flag = 2("acceptable measurement"). They are recorded in README.md's and tabulated here.

section	year	EXPO code	station	cast	description
75N	2006	58GS20060721	266	1	Bad at 1 db
75N	2006	58GS20060721	260	1	Negative oxygen
A05	2011	29AH20110128	6	1	Bad below 568 db
A12	1992	06AQANTX_4			Noisy salinity
A12	2008a	35MF20080207			Pressure missing at p=9 db
A13	1983	316N19831007,316N19840111			Gaps in data, salinity suspect
A22	1997	316N151_4	7	1	Spikes in S below 2500 db
A22	1997	316N151_4	35	1	Spikes in T,S around 2612 db
AR07E	2000	64PE20000926			No quality flag
AR07W	1999--2011				Uncalibrated CTD
I01	1995	316N145_11			FLAG=3
I02	1995		118	1	Wrong P at 2223 db
I06S	2019	325020190403	5-8		Bad oxygen
I08S	2007	33RR20070204	50	2	FLAG=8
IR06E	2000	35MF200009			Noisy DO
IR06E	2000	35MF200009	10		Bad DO
IR06-I10	1995	09FA9503_1			Flag=4 for S
P01	1999	49NZ199901_1			Flag=1 for P
P01	1999	49KA199905_2			Noisy S near surface
P15	2009	09SS20090203	116	1	Extremely low S at P=5784 db
SR01	1997	74JC27_1	24		Wrong pressure P=1113, 1137 db

4. Reprocessing

All input CTD files can be downloaded from CCHDO. List of URLs of the CTD files is getCTD.list, which is generated from README.mds by getCTDlist.pl.

It is possible to reprocess all sections with all_batch.m. Note that PREFIX in batch.m and batch.sh is the output directory. This directory is hard-coded in all_batch.m. Edit all these files as appropriate before running all_batch.m.

5. GLODAP interface

The list files (e.g. P16/p16_1992.list) can be used to extract data from GLODAP (tested on GLODAP version 2) to plot such bottle data as carbons, nutrients, and CFCs. Relevant Matlab scripts are found under GLODAPinterface.

Download the Matlab binary (.mat) of Merged Master File from the GLODAP site. The location of this file is hard-coded in line 36 of fromGLODAP.m. Then

Matlab>> D_reported = fromGLODAP('P16/p16_1992.list', 'cfc12');

will extract the CFC12 data along the stations listed in P16/p16_1992.list from GLODAP data set and put them in the D_reported structure (in the data field) without gridding. For gridding and/or plotting, see an example plotGLODAP.m.

The parameter cfc12 can be one of

aou, c13, c14, ccl4, cfc11, cfc113, cfc11, cfc12, chla, doc, doi, don,
fco2, fco2temp, gamma, h3, he, he3, neon, nitrate, nitrite, o18, oxygen,
pccl, pcfc11, pcfc113, pcfc12, phosphate, phts25p0, phtsinsitutp,
psf6, sf6, sigma0, sigma1, sigma2, sigma3, sigma4, silicate,
talk, tco2, tdn, theta, toc

The algorithm to find stations is described in the source code.

For gridding, use your favourite vertical interpolation scheme. Somewhat classical Reiniger-Ross method (Deep-Sea Res. 1968) is found as GLODAPinterface/vinterp_rr68.m. The example in GLODAPinterface/plotGLODAP.m uses MRST-PCHIP method (Barker and McDougall, J.Atmos.Ocean Tech. 2011) packed in the TEOS-10 GSW Oceanographic Toolbox.

ToDo

• -