Skip to content

Background

In response to urging from the climate modeling community, atmospheric research centers are now regularly producing consistent, global analysis of atmospheric fields spanning back in time over several decades. This technological feat is achieved by employing a frozen, state-of-the-art global data assimilation system and an atmospheric general circulation model to produce as consistent as possible a set of atmospheric fields. The resulting "reanalysis products" are highly suitable for use in a variety of climate modeling applications and in particular those directed towards understanding interannual and decadal climate variability, such as AOMIP.

NCEP/NCAR Reanalysis

The National Centers for Environmental Protection (NCEP) and the National Center for Atmospheric Research (NCAR) carry out atmospheric reanalysis projects.

Their first notable product, NCEP/NCAR-40, produced a 40 year record (1957-1996) of global analyses of atmospheric fields (Kalnay et al., 1996). Recently, that reanalysis product has been extended, NCEP/NCAR-50, resulting in a record that now spans more than 50 years (1948-Present) (Kistler et al., 2001).

The NCEP/NCAR-50 reanalysis product is available in NetCDF format from a data server at the Climate Diagnostics Center (CDC), an affiliate laboratory of the National Oceanographic and Atmospheric Administration (NOAA). Modeling groups participating in the AOMIP 50-Year Experiment will use only the NetCDF data originating the AOMIP web site that has been copied from the CDC archive. This is to ensure that each AOMIP modeling group is using identical atmospheric forcing. The NCEP/NCAR-50 data is available from other data centers and in other formats, (e.g., the more cumbersome GRIB format from a data archive at NCAR) but, again for consistency of intercomparison in AOMIP, only the NetCDF data set on the AOMIP website will be employed.

ECMWF Reanalysis

The European Center for Medium-Range Weather Forecasting (ECMWF) also carries out atmospheric reanalysis projects.

The first major project was ERA-15 and covered a 15 year period (1979-1993). The reanalysis product from the ERA-15 project is available to UCAR affiliate institutions through a data server at the National Center for Atmospheric Research (NCAR).

Currently a 40 year reanalysis project, ERA-40, covering the period 1957-2001 is occurring. The completed reanalysis product will likely be available in Fall, 2002.

The ECMWF reanalysis products will not be used in the AOMIP 50-Year experiment. As noted above, the AOMIP 50-Year Experiment will utilize the NCEP/NCAR-50 reanalysis product to simulate the Arctic Ocean during the period 1948-Present. However, later stages of the AOMIP project may involve use of the ECMWF data.

OMIP Forcing

As a forcing data set for OMIP, the Max Plank Institute for Meteorology (MPI) has created an annual atmospheric climatology from the ERA-15 reanalysis product. The climatology has typical daily fluctuations imposed upon it using Gaussian filtering. There are 13 forcing fields in the OMIP Forcing data:

Forcing Fields
Zonal Wind Stress (east/west)
Meridional Wind Stress (north/south)
Scalar Wind
Daily Std. Dev. Scalar Wind
Air Temperature (2 m)
Dewpoint Temperature (2 m)
Sea-Surface Temperature
Total Shortwave Radiation
Net Shortwave Radiation
Total Cloud Cover
Precipitation
Evaporation
River Runoff

he data is available in this electronic format:

  • ASCII   (as separate ~ 0.1 GB files for each forcing field) from the OMIP website

Basic details about the ASCII data are found in descriptive text at the OMIP website, and are useful for understanding general aspects of the OMIP Forcing. There are currently no plans for use of the OMIP forcing in the AOMIP project. However, later stages of the AOMIP project may involve use of the OMIP data.

OMIP ASCII data descriptive text

The tar-files contain 12 files each: one file for one month.
Each of these files contains 30 days of data.
The filenames are composed of the parameter name (see the list below)
and the month (01 for January, ..., 12 for December).
There are no missing values in the data.

Reading routine for one file containing data for one month:
===========================================================

C
C     320 longitudes
C     160 latitudes
C     30 days per month
C
parameter(nxatm=320,nyatm=160,ndays=30)
dimension field(nxatm,nyatm),ifield(nxatm)

iunit=20
open(iunit,file=filename,
2       form='formatted',access='sequential')

C
C---> signif is dependent on the format
C
signif=10**4   [ or 10**5 or 10**6 ] 999    format(20I4)   [ or 16I5  or 12I6,  see list below]

do nd=1,ndays
do j=1,nyatm
read(iunit,999) (ifield(i),i=1,nxatm)
do i=1,nxatm
field(i,j) = fac*float(ifield(i))/signif + add
enddo
enddo
enddo

data
====

parameter                             add         fac         format   unit
------------------------------------------------------------------------------
2m dewpoint temperature               190.        120.        20I4     K
2m temperature                        190.        130.        20I4     K
zonal stress (east/west)               -4.3         8.6       12I6     Pa
evaporation                            -2.8E-7      3.2E-7    16I5     m/s
net short wave solar radiation          0.        410.        16I5     W/m^2
meridional stress (north/south)        -3.3         6.0       12I6     Pa
runoff                                  0.          1.7E-5    12I6     m/s
scalar wind                             0.         27.        16I5     m/s
sea surface temperature               180.        165.        20I4     K
daily std.dev. scalar wind       (1)    0.         20.        16I5     m/s
total cloud cover                       0.          1.1       20I4     (0-1)
total precipitation              (2)   -4.1E-7      2.8E-6    12I6     m/s
total short wave solar radiation        0.        910.        16I5     W/m^2

land sea mask                           0.          1.        80I1     1:sea/0:land

(1) Daily standard deviation of scalar wind (two values per day).
(2) Negative precipitation due to inconsistencies in the use of GRIB.

The horizontal grid is described using the syntax of a GrADS control file:
==========================================================================
*
* Longitudes are specified by a linear relationship:
*
*                 starting    increment(=360/320)
*                 longitude
*                 ---------   -------------------
XDEF 320 LINEAR      0.0       1.125000
*
*-------------------------------------------------------------------------
*
* All latitudes are specified starting at the South Pole:
*
YDEF 160 LEVELS
-89.1415 -88.0294 -86.9108 -85.7906 -84.6699 -83.5489 -82.4278 -81.3066
-80.1853 -79.0640 -77.9426 -76.8212 -75.6998 -74.5784 -73.4570 -72.3356
-71.2141 -70.0927 -68.9712 -67.8498 -66.7283 -65.6069 -64.4854 -63.3639
-62.2425 -61.1210 -59.9995 -58.8780 -57.7566 -56.6351 -55.5136 -54.3921
-53.2707 -52.1492 -51.0277 -49.9062 -48.7847 -47.6632 -46.5418 -45.4203
-44.2988 -43.1773 -42.0558 -40.9343 -39.8129 -38.6914 -37.5699 -36.4484
-35.3269 -34.2054 -33.0839 -31.9624 -30.8410 -29.7195 -28.5980 -27.4765
-26.3550 -25.2335 -24.1120 -22.9905 -21.8690 -20.7476 -19.6261 -18.5046
-17.3831 -16.2616 -15.1401 -14.0186 -12.8971 -11.7756 -10.6542  -9.5327
-8.4112  -7.2897  -6.1682  -5.0467  -3.9252  -2.8037  -1.6822  -0.5607
0.5607   1.6822   2.8037   3.9252   5.0467   6.1682   7.2897   8.4112
9.5327  10.6542  11.7756  12.8971  14.0186  15.1401  16.2616  17.3831
18.5046  19.6261  20.7476  21.8690  22.9905  24.1120  25.2335  26.3550
27.4765  28.5980  29.7195  30.8410  31.9624  33.0839  34.2054  35.3269
36.4484  37.5699  38.6914  39.8129  40.9343  42.0558  43.1773  44.2988
45.4203  46.5418  47.6632  48.7847  49.9062  51.0277  52.1492  53.2707
54.3921  55.5136  56.6351  57.7566  58.8780  59.9995  61.1210  62.2425
63.3639  64.4854  65.6069  66.7283  67.8498  68.9712  70.0927  71.2141
72.3356  73.4570  74.5784  75.6998  76.8212  77.9426  79.0640  80.1853
81.3066  82.4278  83.5489  84.6699  85.7906  86.9108  88.0294  89.1415