This page's content is no longer actively maintained, but the material has been kept on-line for historical purposes.
The page may contain broken links or outdated information, and parts may not function in current web browsers.

Surface Exchange of Carbon Dioxide (Fung et al. 1987, 1990)

This page offers access to a number of datasets which describe the exchange of carbon dioxide between Earth's surface and atmosphere. They were used in simulations of atmospheric CO2 using the GISS three-dimensional tracer transport model.

Individual files according to source may be downloaded via the following links; All are gzipped ASCII-text tables and none exceed 18-kb in length.

Alternatively, you may download this entire collection of carbon dioxide data in one of two ways:

Available Information

The following datasets are available. All are gzipped ASCII-text files.

CO2 release from fossil fuel burning
Derived from Marland (1989). We distribute 1987 data for each country according to the population density in each 1°×1° cell for that country. The resolution is 1°×1°, that of the population data base. The unit of the array is 106 kg C/m2/y. No seasonal variations. Global source strength = 5.3×1012 kg/y
CO2 release from land use modification
Derived from Houghton et al. (1989). We distribute Houghton et al.'s data uniformly across the area of each country. The resolution is 1°×1°, that of the country data base. The unit of the array is 108 kg C/m2/y. No seasonal variations. Global source strength = 0.3×1012 kg/y.
CO2 exchange with the oceans
At 4°×5° resolution, this is the del(pCO2) pattern reported as Figure 1 in Broecker et al. (1986), multiplied by gas exhange rates based on 14C. The pattern was used in curve (d) in Figure 1 of Tans et al. (1990). More recent data have not been released by Takahashi. The unit of the array is 107 kg C/m2/y. No seasonal variations. Global source strength = - 2.6×1012 kg C/y.
Monthly exchange with natural ecosystems (e.g., vegetation and soils):
Twelve files: JAN, FEB, MAR, APR, MAY, JUN, JUL, AUG, SEP, OCT, NOV, DEC
These are the arrays used in Fung et al. (1987) and have resolution of 4°×5°. The unit of the array is 1013 kg C/m2/s. One array per month. The global annual source should add up approximately to zero (given vagaries of number of days per month and ways to interpolate between months).

Reading the Data

A snippet of FORTRAN code like the following can be used to read the ASCII text files.

C   UCOMMENT ONE OF THE FOLLOWING DEPENDING ON RESOLUTION:
C      PARAMETER (IM=360,JM=180)      FOR 1X1
C      PARAMETER (IM=72,JM=46)        FOR 4X5
C
      PARAMETER (IM=???,JM=???)
      CHARACTER*80 INFO1, INFO2, INFO3
      INTEGER   IARRAY(IM,JM)
C

      READ(10,910) INFO1
      READ(10,910) INFO2
      READ(10,910) INFO3
      READ(10,920) IARRAY
C
C   CAUTION:  CHECK FOR MISSING AND UNDEFINED VALUES IN IARRAY FIRST,
C      BEFORE RESCALING AND OPERATING ON THE NUMBERS.   THE VALUES OF
C      THESE PARAMETERS ARE GIVEN IN INFO1 AND INFO2.
C
  910 FORMAT (A80)
  920 FORMAT (10(1X,I7))
C
      STOP
      END

Explanation of Arrays

In the sample code, four variables are used to store the extracted information: INFO1, INFO2, INFO3, and IARRAY. The first three are character strings which describe the dataset; the latter contains the actual data. Following are some samples of INFO1, INFO2, and INFO3.

CO2FOS.MRL

 CO2: FOSSIL FUEL       (KG C/M2/Y) DIMENSION = 360 X 180    SCALE = 0.10E+07
 MISSING =  9999999  UNDEF =  -999999    MIN =        0      MAX =  4483131
 NUMREC = 6483 (NUMBER OF RECORDS IN DATA FILE)

CO2VEG.JAN

 CO2: JAN  VEGETATION   (KG C/M2/S) DIMENSION =  72 X  46    SCALE = 0.10E+14
 MISSING =  9999999  UNDEF =  -999999    MIN =  -121330      MAX =   222301
 NUMREC =  335 (NUMBER OF RECORDS IN DATA FILE)

The surface grid IARRAY(I,J) has two possible arrangements depending on the resolution of the dataset.

For a 1°×1° gridded dataset, the layout is:

  • I = 1 is centered at 179.5°W
  • I increases eastward
  • J = 1 is centered at 89.5°S
  • J increases northward

For example:

       87S - | - - - | - - - | - - - | - -
             | (1,3) | (2,3) | (3,3) |
       88S - | - - - | - - - | - - - | - -
             | (1,2) | (2,2) | (3,2) |
       89S - | - - - | - - - | - - - | - -
             | (1,1) | (2,1) | (3,1) |
       90S - | - - - | - - - | - - - |
            180W    179W    178W    177W

For a 4°×5° gridded dataset, the layout is:

  • I = 1 is centered at 180°W
  • I increases eastward
  • J = 1 is centered at 90°S
  • J increases northward
  • Note that J=1 and J=46 are only 2° of latitude wide.

For example:

       80S - | - - - | - - - | - - - | - -
             | (1,3) | (2,3) | (3,3) |
       84S - | - - - | - - - | - - - | - -
             | (1,2) | (2,2) | (3,2) |
       88S - | - - - | - - - | - - - | - -
             | (1,1) | (2,1) | (3,1) |
       90S - | - - - | - - - | - - - |
          177.5E  177.5W  172.5W  167.5W

References

Papers and reports which describe the datasets include: