Welcome to the web pages containing the public access directory of the Global Undersea Pressure data set (GLOUP). In 1999, an IAPSO resolution recommended that GLOUP be established as a part of the IAPSO Permanent Service for Mean Sea Level (PSMSL). GLOUP is the global data bank for ocean bottom pressure measurements.

A Description of the data set

The data come in two forms: High frequency (typically hourly or every 15 minutes), and daily values. High frequency data contain the total pressure, predicted pressure from a tidal analysis, and residual pressure. The tidal analysis is performed using 63 constituents for records longer than 12 lunar months (354.36 days: constituents listed here), 55 independent and 2 related constituents for records longer than 6 lunar months (177.18 days) but shorter than 12 lunar months (constituents listed here), and 27 independent and 8 related constituents for records longer than 25 days but shorter than 6 lunar months (constituents listed here). Records shorter than 25 days contain no tidal predictions or residuals. The tidal analyses are generally performed on an integer number of lunar months of data, ignoring the first few days, unless gaps in the record make a different subset (or all) of the data preferable.

The daily values are only produced for records longer than 25 days. These contain two values for each day, calculated from residual pressures from the high frequency data. The first value is a simple average of the residual pressures measured on that day. The second value is a value filtered with a Doodson X0 filter to remove any residual signal at tidal frequencies. The daily value is the filtered value closest to 12:00 GMT.

For each pressure record, there is a corresponding documentation file. Some of the records have contributed to the IAPSO pelagic constants data set of tidal constants from deep deployments. Where this is the case, the corresponding IAPSO number is noted in the header of the high frequency data file.

Data can be accessed file by file, or as zipped collections of files sorted in various ways (zipped files should be unzipped either on unix, using unzip -a filename, or on MSDOS using PKUNZIP):

Tables containing links to individual data files and documentation:

Zipped files:

Single concatenated files for documentation and tides:

Format of the data

High frequency data

The high frequency data are in `TASK-2000' format, which is a modified ACCLAIM format, with 10 columns rather than 11. This consists of a 20 line header, followed by 10 columns of data: The data are in Fortran format (I6,I2,I5,I4,F7.3,5F8.2)

An example header, followed by the first 5 records, is:

     GLOBAL UNDERSEA PRESSURE DATA        FORTRAN format(i6,i2,i5,i4,f7.3,5f8.2)
 1: Line count 2: Flag (good data = 0) 3: Year 4: Day (1st Jan = 1) 5: Hour     
 6: Relative pressure (mbar) Digiquartz sensor  DQ43118              
 7: Tidal predictions (mbar)                                                    
 8: Tidal residuals   (mbar)                                                    
 9: Temperature (deg C) QIT6                                                
10: Temperature (deg C) QIT1                                               
     2           Number of pressure sensors on this instrument
     0           File index of best pressure record (0 if it is this file)
     0           File index of 1st deployment at this site (0 if this is 1st)
 -59.7280        North latitude   (degrees, -90 to +90)
 -55.4920        East longitude   (degrees -180 to +180)
3690.0000        Depth            (metres)
1992 318 14.000  Start date       (year, day, hour)
1996 324 15.993  End date         (year, day, hour)
 1.000   35211   Sample interval  (hours); Number of data entries
Data collected by: Proudman Oceanographic Laboratory                            
Note: 3 IAPSO number 4.2.30                                                     
Name: myr9296.a1d.a
 Count f Year Day  Hour      6       7       8       9       10                 
     1 0 1992 318 14.000   16.18  -10.23   26.41    0.15    0.25
     2 0 1992 318 15.000  -19.42  -45.41   25.99    0.15    0.24
     3 0 1992 318 16.000  -47.45  -73.40   25.95    0.14    0.23
     4 0 1992 318 17.000  -63.16  -88.46   25.30    0.14    0.23
     5 0 1992 318 18.000  -62.36  -87.79   25.43    0.13    0.23
The first three numbers after the column descriptions indicate the relationship between this file and others in the data set.

The number of pressure sensors refers to the number of simultaneous pressure measurements at the same site, either on the same rig or on rigs deployed and recovered at about the same time. The number given is the number contained in the GLOUP data set, there may be others which malfunctioned or have not been found.

Where there are several sensors, one has been chosen as `best', on a subjective basis depending on length of record and apparent noise in the residuals. If this file is not the best, then the index of the file containing the best data is given here.

Sometimes a series of deployments has been made at the same place (this need not mean exactly the same position, but must be within a few kilometres to count). In that case, the index of the file containing the first data from that site is given here. If there were several sensors on that first deployment, the index of the best is given.

The file index is sufficient to identify the GLOUP file: All high frequency files have names of the form ghfNNNNNN.dat, where NNNNNN is the file index.

The files also have more individual `Names' reflecting their origin, for example myr9296 refers to the deployment of the MYRTLE instrument which produced a 4 year record from 1992 to 1996. There is not a consistent system for producing these names, but they give the files a more memorable identity than the index.

When the pressure has been used to calculate tidal constituents in the pelagic tidal constants data set, the corresponding IAPSO number is noted here. The first character of the Note is `0', `1', `2', or `3', according to whether there is no tidal analysis, or analysis with the short, medium, or long record set of constituents. Other important information may also appear here, but more complete information will be found in the documentation file.

If the Year/Day form for the date is inconvenient, several Fortran subroutines/functions are available to convert to other forms:

The header can be read using the Fortran subroutine An example of its use to produce a list of notes from all the headers, and to read all the data checking that the start and end dates in the header are correct, is

Daily data

Daily data are in files with names of the form gdaNNNNNN.dat, where NNNNNN is the index number as for the high frequency files. They each have a header similar to the high frequency header, e.g.

G L O B A L   U N D E R S E A   P R E S S U R E   D A T A                       
         DAILY VALUES of PRESSURE in millibars                                  
          FORTRAN FORMAT (i6,i5,i4,2i2,2f8.2)                                   
 1: Line count    2: Year           3:Day of year                               
 4: Flag for mean 5: Flag for filter (0 is good)                                
 6: Daily mean    7: Doodson X0 filtered value                                  
     3           Number of pressure sensors on this instrument
    99           File index of best pressure record (0 if it is this file)
     0           File index of 1st deployment at this site (0 if this is 1st)
 -58.3630        North latitude   (degrees, -90 to +90)
 -56.3550        East longitude   (degrees -180 to +180)
3776.0000        Depth            (metres)
1994 324         Start date       (year, day)
1996 196         End date         (year, day)
 49676     603   MJDN of day before first record; Number of data entries
Data collected by: Proudman Oceanographic Laboratory                            
Note: 3 none                                                                    
Name: dpc9496a.i4d.c                                                            
 Count Year Day 1 2    Mean  DoodX0                                             
     1 1994 324 0 0   -0.75   -0.49
     2 1994 325 0 0    0.56    0.45
     3 1994 326 0 0    0.22    0.31
     4 1994 327 0 0   -0.33   -0.30
     5 1994 328 0 0    0.33    0.18
In place of the sample interval found in the high frequency header, the Modified Julian Day Number, MJDN0, of the day before the first record is given, so the MJDN of each record is MJDN0 + Count.

The columns labelled `1' and `2' contain flags for the daily mean and Doodson X0 filtered values respectively. The mean is still calculated where it is flagged, and will be the mean of however many observations are unflagged for that day. The filtered value is set to zero and flagged when any of the hourly values which go into the 39 hour filter are flagged.

The header can be read using the Fortran subroutine

Tidal analyses

Tidal analyses are in files with names of the form gtiNNNNNN.dat, where NNNNNN is the index number as for the high frequency files. An example of a tidal analysis for a long record is gti000147.dat:

Following initial text relating to the reading and blocking of the data, this gives a list of tidal constituents, their name, speed in degrees per hour, and analysed amplitude in millibars and Greenwich phase lag in degrees.

Contact information

For further information contact Dr C. W. Hughes at

Proudman Oceanographic Laboratory
6 Brownlow Street
Liverpool L3 5DA

Telephone:   (UK international code 44) 151-795-4800
Fax:         (UK international code 44) 151-795-4801
Email:       cwh@pol.ac.uk