CGCM1 (cccsl.asc)

Decadal mean sea level rise data is provided.  The steric
component of sea level rise was computed as the difference, between the
transient runs and the control run, of the vertically integrated specific
volume (essentially as in Bryan,K., Climate Dynamics, 12:545-555, 1996) from
the bottom of the ocean to the surface.  No glacial or ice sheet melt is
included.  The file contains 20 decade mean (from
1901 to 2100) 193X96 gaussian longitude/latitutde (ll) grids of sea-level rise 
for a particular 200 year model run.

DATA FORMAT:

The data file is in ASCII format, containing for each month a header
followed by data values. The FORTRAN statements used to write the data
out are

      WRITE(1,6000) (IHEAD(I),I=1,8)
      WRITE(1,6100) ((F(I,J),I=1,IHEAD(5)),J=1,IHEAD(6))
 6000 FORMAT(1X,A4,I10,1X,A4,5I10)
 6100 FORMAT(1P6E12.5)

where

IHEAD(1) = GRID, type of the data,
IHEAD(2) = YYYYMM, time step,
IHEAD(3) = variable name,
IHEAD(4) = level,
IHEAD(5) = 193, number of longitudes,
IHEAD(6) = 96, number of gaussian latitudes,
IHEAD(7) = 3 for ocean data,
IHEAD(8) = packing density of the original binary data,
F(I,J)   = the data.

Latitudes (J=1,96) are numerated from South to North. J=1 corresponds to
-89.064 degree South and J=96 corresponds to 89.064 North.  Longitudes
(I=1,193) are numerated eastwards from the Greenwich meridian. The grid
contains an additional longitude I=193, identical to I=1, the Greenwich
meridian, to make a cyclic boundary. The land data points are masked with
-999.0

I have also prepared these files into another (FORTNER Transform) format:

csl30.tra (2021-2040) - (1981-2000)
csl90.tra (2081-2100) - (1981-2000)
 
The top four lines are headers, with the third line listing the latitudes
and the fourth line listing the longitudes.  Then there are 96 lines (one
line for each latitude) of 192 longitudes, formated (192f10.2).  Note that I 
have switched the longitudes to go from -180 to 178.12 degrees, W to E.

I'd like to thank Francis Zwiers for supplying these data.



HADCM2 (hsl30.asc, hsl90.asc)

The eustatic sea level change from HADCM2 (1% CO2 plus
sulfates) represents thermal expansion and glacial melt, but not ice sheet 
melt (because of the uncertainty of the net effect of Antarctica and 
Greenland).  Sea level rise varies with location because of the local heating
and thermal expansion.  Remember to factor in isostatic rebound
and subsidence when using these values to compute relative sea
level rise at a particular location.

Files:
hsl30.asc (2020-2040) - (1980-2000)
hsl90.asc (2080-2099) - (1980-2000)

Each file is a 2D Fortran array written out with a single write statement
and format (6e13.5). The fields have spatial resolution of 3.75 degrees of
longitude by 2.5 of latitude, dimensioned (96,73). Longitude starts at 0 and
runs E. Latitude starts at 90S and goes N. Missing data (i.e. land) is
indicated by -1e30.  They are the centres of the boxes. The first and last row
are therefore half-missing (because they extend beyond 90N and 90S), but both 
of them are entirely land anyway.  Also, note that there is an island at the 
North Pole, as in other ocean models, for reasons of numerical convenience.

I have also prepared these files into another (FORTNER Transform) format:

hsl30.tra
hsl90.tra
 
The top four lines are headers, with the third line listing the latitudes
and the fourth line listing the longitudes.  Then there are 73 lines (one
line for each latitude) of 96 longitudes, formated (96e13.5).  Note that I have 
switched the longitudes to go from -180 to 180 degrees, W to E.

I'd like to thank Jonathon Gregory for supplying these data.

Benjamin Felzer