Output files from SPEEDY v.42 ensembles described in the paper: "Multi-decadal pacemaker simulations with an intermediate-complexity climate model" by F. Molteni, F. Kucharski and R. Farneti (part 2 of 2)

The monthly-mean output from SPEEDY v.42 ensembles (either driven by prescribed sea-surface temperature (SST) or coupled to the TOM3 model) consists of a series of IEEE little-endian binary files and metadata files in text format.
For each year of integration (indicated by a 4-digit number YYYY) and ensemble member (indicated by a 3-digit number NNN), two binary files are present, named:
•    attmNNN_YYYY.grd, including data on the 120x60 grid-point atmospheric grid;
•    sftmNNN_YYYY.grd, including data on the 360x180 grid-point surface grid.
The metadata for these files are contained in the text files attmEEE.ctl and sftmEEE.ctl respectively, where EEE is a 3-digit ensemble identifier (usually, but not necessarily, equal to one of the ensemble-member number NNN).

This repository contains data from:

(part 1) a 41-year 5-member ensemble (653) run with prescribed SST

(part 2) a 70-year 5-member ensemble (104) run with the coupled SPEEDY-TOM3 model.

Integration years are 1980 to 2020 for ensemble 653 and 1951 to 2020 for ensemble 104.

The structure of the binary data and metadata files follows the conventions for gridded datasets set by the GrADS diagnostic and plotting package (developed by the Center for Ocean-Land-Atmosphere Studies of George Mason University), as described here:
 http://cola.gmu.edu/grads/gadoc/aboutgriddeddata.html

In addition to the COLA-GMU web site, free version of the GrADS package for different platforms can be downloaded from the OpenGrADS web site:
http://opengrads.org/

Specifically, the SPEEDY v.42 output consists of sequential-access files where each record contains a two-dimensional field. Three-dimensional fields are stored as a sequence of consecutive records, one for each of the 8 pressure levels where model-level data are interpolated by the post-processing routines. For each month of the year:

the attmNNN_YYYY.grd files contain a sequence of 9 3-D variables and 26 2-D variables;
the sftmNNN_YYYY.grd files contain a sequence of 21 2-D variables.

Within each record, grid-point data are stored as a NLONxNLAT array with longitude varying from west to east and latitude varying from south to north. The list of variables and levels is specified in the attmEEE.ctl and sftmEEE.ctl files. These files contain descriptors which allow the data of each ensemble to be accessed as a single dataset by the GrADS package.

Although the metadata files are specific to the GrADS package, the binary data can be read by different types of code. As example of fortran90 instructions to read the content of the attmNNN_YYY.grd and sftmNNN_YYY.grd files for one year/ens.member is as follows:

integer, parameter :: nlon=120
integer, parameter :: nlat=60
integer, parameter :: nlev=8
integer, parameter :: nlon0=360
integer, parameter :: nlat0=180

integer :: jmonth, jvar3d, jvar2d, jlev
real :: fld3d(nlon,nlat,nlev)
real :: fld2d(nlon,nlat), fld0(nlon0,nlat0)

open (unit=1, file=”attmNNN_YYY.grd”, form=”formatted”, access=”sequential”)
open (unit=2, file=”sftmNNN_YYY.grd”, form=”formatted”, access=”sequential”)

do jmonth=1,12

   do jvar3d=1,9
        do jlev=1,nlev
             read (1) fld3d(:,:,jlev)
             …………
        enddo
    enddo

   do jvar2d=1,26
        read (1) fld2d(:,:)
         ………
    enddo

   do jvar2d=1,21
        read (2) fld0(:,:)
         ………
   enddo

enddo

close (1)
close (2)