Skip to content. | Skip to navigation

Personal tools
You are here: GES DISC Home Precipitation Documentation guide GES DISC DAAC Data Guide: GEOS-1 Multiyear Assimilation Data Set Guide

GES DISC DAAC Data Guide: GEOS-1 Multiyear Assimilation Data Set Guide

warn_icon.gif THIS DOCUMENT IS AVAILABLE ON THE GES DISC WEB SITE FOR HISTORICAL INFORMATION PURPOSES ONLY. Information provided in this document may not be accurate. We recommend checking other sources related to these data or sensors to acquire reliable and updated information.

Explanation: The Dataset or Sensor Guide Document you are accessing is no longer actively maintained. The Dataset Guide Documents were created for earlier versions of the NASA EOSDIS system. The content of these documents, particularly with regard to characteristics of the data or technical descriptions of a sensor, is likely still accurate. However, information such as contact names, phone numbers, mailing addresses, email addresses, software programs, system requirements, and data access procedures may no longer be accurate. We therefore recommend searching for updated information from other sites to insure that reliable and current information is obtained.


Data Assimilation Office GEOS-1 Multiyear Assimilation



Data Assimilation is the process of ingesting observations into a model of the Earth system. The Data Assimilation System uses meteorological observations and an atmospheric model. The result is a comprehensive and dynamically consistent data set that represents the best estimate of the state of the atmosphere at that time. The assimilation process fills data voids with model predictions and provides a suite of data-constrained estimates of unobserved quanitities such as vertical motion, radiative fluxes, and precipitation.

The Data Assimilation Office (DAO) at GSFC produced Multiyear Assimilated dataset. It is global, gridded atmospheric data for use in climate research, produced by assimilating rawinsonde reports, satellite retrievals of geopotential thickness, cloud-motion winds, aircraft, ship, and buoy reports with model forecasts employing version 1 of the Goddard Earth Observing System (GEOS-1) atmospheric general circulation model (GCM).

The output data have been packaged in various datasets of different resolutions and collections of parameters. The full resolution output is refered to as the GEOS- Multiyear Assimilation Timeseries Data is archived at the Goddard Distributed Active Archive Center (DAAC). Other derived datasets are also available from the Goddard DAAC, many via anonymous ftp.

Table of Contents:


1. Data Set Overview:

Data Set Identification:

DAO GEOS-1 Multiyear Assimilated Datasets

Data Set Introduction:

The data described in this guide document are from a multi-year assimilated dataset produced at the Data Assimilation Office of the Goddard Laboratory for Atmospheres with a non-varying assimilation system. The model is called the Goddard Earth Observing System (GEOS-1), and the assimilation run that produced the bulk of the data that the DAO distributes is call the GEOS-1 Multiyear Assimilation.

The output of the GEOS-1 Multiyear Assimilation is very volumunous. The large array of available prognostic and diagnostic fields represent a comprehensive description of the Earth's climate. The output has been packaged in various ways, in various resolutions. This document describes the data assimilation process and the parameters of the GEOS-1 Multiyear assimilation output. Seperate documentation describes the different datasets within which the output are packaged.

Data are currently available for the time period March 1985 through November 1993.


The analysis of historical data using a assimilation system is important to enable researchers to study atmospheric variability and potential short-term climate change. The analysis is simplified by using a non-varying system since the researcher need not account for spurious changes resulting from changes to the assimilation system. DAO's primary mission is the development of the tools necessary to produce research-quality assimilated data sets [NRC, 1991]. The mission of the DAO is unique because it is the quality and the utility of the assimilated data, rather than the forecast, that measures the success of the effort. Ultimately, the assimilation system developed in this effort will be used to assimilate the satellite and other air and surface-based measurements of the Earth system which will become available at the turn of the century from the Earth Observing System (EOS) program.

The DAO has produced this benchmark multi-year data set using version 1 of the GEOS-1 assimilation system. The objectives in producing these data are twofold. First, it is believed that the absence of spin-up in the hydrological cycle and the use of a fixed assimilation system will make these data extremely useful for a wide range of climate studies. Secondly, making the data available to the larger scientific community will attract valuable feedback on the quality and limitations of the assimilated data. This feedback will help guide future development.

from Schubert et al.


Summary of Parameters:

In general the parameters from the GEOS-1 Multiyear Assimilation can be divided into three groups. This naming convention is historically rooted; as more parameters have been added it is no longer strictly accurate.
boundary condition fields
These fields contain predetermined data that is fed into the model. Generally these are derived from averaged data from outside sources, but are time interpolated if necessary by the Data Assimilation System. (phis, albd, gwet, lwi and gtmp are all boundary conditions).
prognostic fields
define the state of the atmos according to fluid dynamical principles. These are the parameters that the model calculates future values of when it goes through a forcast cycle. They are instantaneous values.


diagnostic fields
These fields are derived from the prognostic fields. They are averaged values - usually a three or six hour average centered around the output time. Due to the large number of diagnostic fields in the GEOS-1 Multiyear Assimilation, these fields are further divided into primary and secondary diagnostics.

See the Glossary or Calculated Variables in this document.  The parameters included in the timeseries dataset are grouped under seven data products. Each parameter is described further in the Parameter Table, and in more detail in the document:



    These parameters consist of various surface boundary conditions, prognostic and diagnostic surface quantities and vertically integrated fields. The fields are made available every six hours, though it should be noted the boundary condition fields are interpolated linearly in time from monthly mean fields.


    • surface geopotential height
    • surface albedo
    • ground wetness from off-line bucket model [Schemm et al, 1992]
    • surface pressure minus top of the atmosphere pressure (10mb)
    • surface ground temperature
    • sea level pressure
    • water=1, land=2, permanent ice=3, sea ice=4 flags
    • vertically integrated (barotropic) zonal wind
    • vertically integrated (barotropic) meridional wind



    These parameters are "instantaneous snapshots" reported at 6 hour intervals. Each product is reported as a full three-dimensional field at 18 pressure levels (1000, 950, 900, 850, 800, 700, 600, 500, 400, 300, 250, 200, 150, 100, 70, 50, 30, 20mb).


    • zonal wind speed
    • meridional wind speed
    • geopotential height
    • temperature profiles
    • specific humidity profiles
    • turbulent kinetic energy
    • standard deviation of the height error


  3. DIAGNOSTICS: Precip/Evap

    These parameters are 3-hour averages, reported at the endpoint of the 3-hour interval.


    • surface pressure minus top of the atmosphere pressure (10mb)
    • total precipitation
    • convective precipitation
    • surface evaporation
    • vertically integrated zonal wind * specific humidity (U*Q)
    • vertically integrated meridional wind * specific humidity (V*Q)
    • vertically integrated zonal wind * temperature (U*T)
    • vertically integrated meridional wind * temperature (V*T)
    • vertically integrated precipitable water


  4. DIAGNOSTICS: Momentum/Heat Flux

    These parameters are 3-hour averages, reported at the endpoint of the 3-hour interval.


    • surface pressure minus top of the atmosphere pressure (10mb)
    • zonal momentum surface stress
    • meridional momentum surface stress
    • surface flux of sensible heat
    • surface drag coefficient for temperature and specific humidity
    • surface drag coefficient for winds
    • surface wind speed
    • friction velocity
    • surface roughness
    • planetary boundary layer depth


  5. DIAGNOSTICS: Radiation

    These parameters are 3-hour averages, reported at the endpoint of the 3-hour interval.


    • surface pressure minus top of the atmosphere pressure (10mb)
    • net upward LW radiation at the surface
    • net downward SW radiation at the surface
    • outgoing longwave radiation
    • outgoing longwave radiation clear sky
    • surface longwave radiation clear sky
    • incident SW radiation at the top of the atmosphere
    • outgoing shortwave radiation
    • outgoing shortwave radiation clear sky
    • surface SW radiation clear sky
    • 2-dimensional total cloud fraction


  6. DIAGNOSTICS: Near Surface

    These parameters are 3-hour averages, reported at the endpoint of the 3-hour interval.


    • surface pressure minus top of the atmosphere pressure (10mb)
    • ground temperature
    • surface air temperature
    • surface saturation specific humidity
    • surface pressure tendency
    • zonal wind at 2m
    • meridional wind at 2m
    • temperature at 2m
    • specific humidity at 2m
    • zonal wind at 10m
    • meridional wind at 10m
    • temperature at 10m
    • specific humidity at 10m



    These parameters are 6-hour averages, reported at the midpoint of the 6-hour interval. Each product is reported as a full three-dimensional field at 18 pressure levels (1000, 950, 900, 850, 800, 700, 600, 500, 400, 300, 250, 200, 150, 100, 70, 50, 30, 20mb).


    • zonal momentum changes due to turbulence
    • meridional momentum changes due to turbulence
    • temperature changes due to turbulence
    • moisture changes due to turbulence
    • temperature changes due to moist processes
    • moisture changes due to moist processes
    • temperature changes due to LW radiation
    • temperature changes due to SW radiation
    • vertical velocity

Each parameter is described further in the Parameter Table, and in more detail in the document: We highly recommend that users refer to the data_problems file for a chronology of known problems with the assimilated data. This file will be updated periodically. We also request that if a user discovers additional problems that the user will alert the DAO by e-mail at:

Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, July 12, 1999.


Not Applicable

Related Data Sets:

  • ECMWF [Bengtsson and Shukla, 1988]
  • NMC [Kalnay and Jenne, 1991]

2. Investigators:

Investigator(s) Name and Title:

Name: Siegfried Schubert
DATA ASSIMILATION OFFICE code 910.3 NASA/Goddard Space Flight Center Greenbelt MD 20771
Telephone Numbers:301-286-3441
Electronic Mail Address:

Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, Feb 18, 2000.

Name: Richard Rood

DATA ASSIMILATION OFFICE code 910.3 NASA/Goddard Space Flight Center Greenbelt MD 20771
Telephone Numbers:301-286-8203
Electronic Mail Address:

Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, Feb 18, 2000.


Title of Investigation:

DAO 4D GEOS-1 Multiyear Assimilation

Contact(s) (for Data Production Information):

Name:Chung-Kyu Park
DATA ASSIMILATION OFFICE code 910.3 NASA/Goddard Space Flight Center Greenbelt MD 20771
Telephone Numbers:301-286-8695
Electronic Mail Address:

Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, Feb 18, 2000.

3. Theory of Measurements:

The assimilated data are a synthesis of measurements and short-term model forecasts. In the GEOS-1 system this is accomplished using Optimal Interpolation (OI), in which observations of the following variables are used:
  • geopotential thickness
  • zonal (u) and meridional (v) winds
  • water vapor mixing ratio
  • sea level pressure
  • surface winds over the ocean

The Optimal Interpolation (OI) Scheme

The tropospheric version of the OI analysis scheme being used for the control assimilation has been carried out at a horizontal resolution of 2 degree latitude by 2.5 degree longitude at 14 upper-air pressure levels (20, 30, 50, 70, 100, 150, 200, 250, 300, 400, 500, 700, 850, 1000 mb) and at sea level. The analysis increments are computed every 6 hours using observations from a +/- 3 hour data window centered on the analysis times (00, 06, 12, and 18 UTC). The innovation vector (observation minus background forecast) used as input to the OI is computed using a single forecast valid at the analysis time.


The OI scheme is multivariate in geopotential height and winds and employs a damped cosine function for the horizontal correlation of model prediction error. The height-wind cross-correlation model is geostrophic and scaled to zero at the equator. The multivariate surface analysis scheme over the oceans adopts an Ekman balance for the pressure-wind analysis. The moisture analysis for mixing ratio employs only rawinsonde moisture data. All grid point analyses are done using up to 75 nearby observations from within a radius of 1600 km.

The assimilation system does not include an initialization scheme and relies on the damping properties of a Matsuno time differencing scheme to control initial imbalances generated by the insertion of observations. However, the initial imbalances and spinup have been greatly reduced over earlier versions by the introduction of an incremental analysis update (IAU) procedure [Bloom et al., 1991]. As shown in the figure of the IAU procedure,


standard OI analysis increments are computed at the analysis times (0, 6, 12, 18Z). The increments are then inserted gradually into the AGCM by rerunning the forecast and adding a fraction of the increment at each model time step. Over the 6 hour period centered at the analysis time the full effect of the increment is realized. The assimilation thus effectively consists of a continuous AGCM forecast with additional heat, momentum, moisture and mass source terms updated every 6 hours from observations. It is the output of this IAU procedure that makes up the dataset provided here. An important difference between the IAU scheme and the usual Newtonian nudging procedure is that the IAU forcing terms are held constant over the insertion period, while in Newtonian nudging they are proportional to the difference between a target analysis and the instantaneous current model state.


The implication of the IAU procedure for performing budget (e.g., moisture or heat) calculations with the assimilated data is that, in order to balance the budget, one must include the analysis increments as an additional forcing term. If the assimilating model had no bias the mean analysis increments would, of course, be zero and the increments would have no contribution to the mean budget equations. The current model (GEOS-1), does have a bias, and while the initial timeseries dataset provided by the DAAC does not include the analysis increments, they (the increments) can be obtained as the residual of the other terms in the budget equations.

4. Equipment:

Sensor/Instrument Description:

Collection Environment:

Data were collected from globally deployed in situ and remote observation platforms throughout the assimilation period.


  • NOAA polar orbiters (TOVS)
  • ships
  • buoys
  • rawinsondes, dropwindsondes
  • aircraft winds

Source/Platform Mission Objectives:

The assimilation system synthesizes observations and model first guesses with the intention of producing a consistent and accurate estimate of the climate.

Key Variables:

The following observed variables were ingested into the data assimilation system when and where available:


  • Geopotential thickness
  • Winds
  • Water vapor mixing ratio
  • Sea level pressure
  • Surface winds over the ocean

Note that all diagnostic parameters are calculated from AGCM's physical parameterizations in a manner consistent with the prognostic fields (See Optimal Interpolation).

Principles of Operation:

Not Applicable

Sensor/Instrument Measurement Geometry:

Not Applicable

Manufacturer of Sensor/Instrument:

Not Applicable



Not Applicable


Not Applicable

Frequency of Calibration:

Not Applicable

Other Calibration Information:

Not Applicable

5. Data Acquisition Methods:

The input observational database is one that has been accumulated over the years at GLA. Data for times prior to July 1987 were mostly acquired from the National Center for Atmospheric Research (NCAR). Since that time, data have been obtained directly from NMC, and do not include data which came in after the cut-off time for the operational NMC system. In addition to these two sources, some TOVS temperature soundings have come directly from the National Oceanic and Atmospheric Administration (NOAA) National Environmental Satellite, Data and Information Service (NESDIS), and gaps have been filled with data from the National Climatic Data Center (NCDC) in Ashville, NC. For the global sea level pressure and near surface wind analysis over the oceans, data from surface land synoptic reports (sea level pressure only), ships and buoys are used. The upper-air analyses of height, wind and moisture incorporate the data from rawinsondes, dropwindsondes, aircraft winds, cloud tracked winds, and thicknesses from the historical TOVS soundings produced by NOAA NESDIS. The satellite heights are computed using a reference level which depends on the analyzed sea level pressure. The only situation where data are used as a proxy for un-collected observations are 1000 mb height observations which are generated above pressure reports from ships. These serve to further couple the surface and upper-air analyses.

For a description of how these data are used in the model, see the description of optimal interpoltion.

6. Observations:

Data Notes:

Additional processesing notes are available from the anonymous ftp site Beneath the directory "pub/assimilation/e0054A" with the following structure:
This is a plain text description of the dataset.
This is a postscript description of the dataset.
This is a postscript description of the dataset without the embedded fonts.
Contains a summary of problems discovered with the data. This will be updated periodically.
suspic.tyymm (e.g. suspic.t8503)
These are monthly summaries of outlier counts (and locations) based on the variance at each grid point for the assimilated prognostic fields. These are not necessarily bad data since extreme (many standard deviations) values do occur in nature, especially for specific humidity at upper levels. This is meant to be another source of information for judging the quality of the assimilated fields.
Details of the diagnostics (how they were computed, etc.) and further information about the GEOS-1 GCM used in the assimilation may be found in this postscript file.
This is a place for you to document any problems you have discovered with the data. We are very interested in your feedback. For a quicker response you can also send e-mail to:

Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, Feb 18, 2000.


Contains postscript files (bar graphs of data counts) and other text files summarizing the data going into the assimilation.
Gives a history of the processing (e.g. hardware, software glitches)

Field Notes:

Not Applicable

7. Data Description:

Spatial Characteristics:

Spatial Coverage:


Spatial Coverage Map:

GLOBAL, latitude-longitude grid

Spatial Resolution:

The spatial resolution is 2 deg latitude by 2.5 deg longitude. There are 18 pressure levels (1000, 950, 900, 850, 800, 700, 600, 500, 400, 300, 250, 200, 150, 100, 70, 50, 30, 20mb) interpolated from the following 20 sigma levels.

Sigma Levels


Not Applicable

Grid Description:

There are 144 grid points in the longitude direction with the first grid point at the dateline and with a grid spacing of 2.5 degrees. There are 91 grid points in the latitude direction with the first grid point at the south pole and with a grid spacing of 2.0 degrees. There are 18 pressure levels with the bottom level (highest pressure) first.

Temporal Characteristics:

Temporal Coverage:

The temporal coverage of this data set is currently 5 complete years (from March 1985 to February 1990).

Temporal Coverage Map:

There are no gaps, either temporal or spatial.

Temporal Resolution:

  • PROGNOSTIC fields are sampled 4 times daily at 00Z, 06Z, 12Z, 18Z.
  • DIAGNOSTIC surface or other single level fields are given 8 times daily as time averages over the three hours preceding the time stamp (i.e. at 00Z for 21Z(previous day) to 00Z, at 03Z for 00Z to 03Z, at 06Z for 03Z to 06Z....)
  • DIAGNOSTIC upper air fields are given 4 times daily as time averages over 6 hours centered on the times 00Z, 06Z, 12Z, 18Z.

Data Characteristics:



Surface Prognostic

PHIS(m/s)^2surface geopotential height (z * gravity)
ALBD0-1proportion of solar radiation reflected by the surface (0-1)
GWET0-1ground wetness from off-line bucket model [Schemm et al, 1992]
PS-PTOPmbsurface pressure minus top of the atmosphere pressure (10mb)
GTMPKsurface ground temperature
SLPmbsea level pressure
LWIflagwater=1, land=2, permanent ice=3, sea ice=4 flags
UBARm/svertically integrated (barotropic) zonal wind
VBARm/svertically integrated (barotropic) meridional wind

Upper Air Prognostic

UWNDm/szonal wind speed
VWNDm/smeridional wind speed
HGHTmgeopotential height
TMPUKtemperature profiles
SPHUg/kgspecific humidity profiles
QQ(m/s)^2turbulent kinetic energy
HGHTEmstd dev of the height error


Surface Diag 1 Precip/Evap

PS-PTOPmbsurface pressure minus top of the atmosphere pressure (10mb)
PREACCmm/dytotal precipitation
PRECONmm/dyconvective precipitation
EVAPmm/dysurface evaporation
VINTUQm/s g/kgvertically integrated (mass weighted) zonal wind * specific humidity (U*Q)
VINTVQm/s g/kgvertically integrated (mass weighted) meridional wind * specific humidity (V*Q)
VINTUTm/s Kvertically integrated (mass weighted) zonal wind * temperature (U*T)
VINTVTm/s Kvertically integrated (mass weighted) meridional wind * temperature (V*T)
QINTg/cm^2vertically integrated precipitable water

Surface Diag 2 Momentum/Heat Flux

PS-PTOPmbsurface pressure minus top of the atmosphere pressure (10mb)
UFLUXN/m^2zonal momentum surface stress
VFLUXN/m^2meridional momentum surface stress
HFLUXW/m^2surface flux of sensible heat
CTunitlesssurface drag coefficient for temperature and specific humidity
CUunitlesssurface drag coefficient for winds
WINDSm/ssurface wind speed
USTARm/sfriction velocity
Z0msurface roughness
PBLmbplanetary boundary layer depth

Surface Diag 3 Radiation

PS-PTOPmbsurface pressure minus top of the atmosphere pressure (10mb)
RADLWGW/m^2net upward LW radiation at the surface
RADSWGW/m^2net downward SW radiation at the surface
OLRW/m^2outgoing longwave radiation
OLRCLRW/m^2outgoing longwave radiation clear sky
LWGCLRW/m^2surface longwave radiation clear sky
RADSWTW/m^2incident SW radiation at top of the atmosphere
OSRW/m^2outgoing shortwave radiation
OSRCLRW/m^2outgoing shortwave radiation clear sky
SWGCLRW/m^2surface SW radiation clear sky.
CLDFRC0-12-dimensional total cloud fraction (0-1)

Surface Diag 4 Near Surface

PS-PTOPmbsurface pressure minus top of the atmosphere pressure (10mb)
TGKground temperature
TSKsurface air temperature
QSg/kgsaturation specific humidity at the surface
DPDTmb/dysurface pressure tendency
U2Mm/szonal wind at 2m
V2Mm/smeridional wind at 2m
T2MKtemperature at 2m
Q2Mkg/kgspecific humidity at 2m
U10Mm/szonal wind at 10m
V10Mm/smeridional wind at 10m
T10MKtemperature at 10m
Q10Mkg/kgspecific humidity at 10m

Upper Air Diagnostic

TURBUm/s/dyzonal momentum changes due to turbulence
TURBVm/s/dymeridional momentum changes due to turbulence
TURBTK/dytemperature changes due to turbulence
TURBQg/kg/dymoisture changes due to turbulence
MOISTTK/dytemperature changes due to moist processes
MOISTQg/kg/dymoisture changes due to moist processes
RADLWK/dytemperature changes due to LW radiation
RADSWK/dytemperature changes due to SW radiation
OMEGAmb/dyvertical velocity

For more detailed information see

Sample Data Record:

The first Sea Level Pressure record (1 Mar 85, 00Z) is available for examination: DAO_SLP.85030100. It was created by the fortran read program called: dao_read_1.f. Sea Level Pressure Map

8. Data Organization:

Data Granularity:

A general description of data granularity as it applies to the IMS appears in the EOSDIS Glossary.

Each timeseries file contains one month of data, and is considered a DATA GRANULE. Each file consists of a time sequence of either a single three-dimensional upper air parameter (4 times daily), or a collection of several single level or vertically integrated parameters (8 times daily). The data files are named "e0054A.prs.NAME.bYYMMDD.eYYMMDD" where NAME is one of the filenames listed in the following table. Below is a table summarizing the data granularity (for a more detailed description of the parameters, see Parameter Table:



Data Format:

The data representation is ieee 32 bit floating point, written sequentially by FORTRAN 77. There are no header or trailer records.

9. Data Manipulations:


Derivation Techniques and Algorithms:

A detailed guide which describes how the variables were created is available as a postscript document. It can be accessed via this document or via anonymous ftp from under the directory "pub/gcm". The file is called

Data Processing Sequence:

Processing Steps:


The current tropospheric version of the model (GEOS-1) uses the potential enstrophy and energy-conserving horizontal differencing scheme on a C-grid developed by Sadourney [1975], and further described by Burridge and Haseler [1977]. An explicit leapfrog scheme is used for the time differencing, applying an Assilin [1972] time filter to damp out the computational mode. An 8th order Shapiro filter is applied to the wind, potential temperature and specific humidity to avoid non-linear computational instability. The filter is applied at every time step in such a way that the amplitude of the two-grid interval wave would be reduced by half in two hours. Applying the filter weakly at each time step eliminates the shock that occurred in earlier assimilations by intermittent application of filter. The model also uses a polar Fourier filter to avoid linear instability due to violation of the CFL condition for the Lamb wave and internal gravity waves. This polar filter, however, is applied only to the tendencies of the winds, potential temperature, specific humidity and surface pressure. The model's vertical finite differencing scheme is that of Arakawa and Suarez [1983]. The above dynamics routines are organized into a plug-compatible module called the ARIES/GEOS "dynamical core" developed by M. Suarez and L. Takacs in the Goddard Laboratory for Atmospheres.



The infrared and solar radiation parameterizations follow closely those described by Harshvardhan et al. [1987]. In the longwave water vapor absorption is parameterized as in Chou [1984], the 15 micron band of CO2 as in Chou et al. [1983], and ozone absorption as in Rodgers [1968] with the modifications suggested by Rosenfield et al. [1987]. The shortwave follows Davies [1982], as described in Harshvardhan et al. [1987]. Shortwave absorption by water vapor uses a k-distribution approach as in Lacis and Hansen [1974]. Cloud albedo and transmissivity for the model layers are obtained from specified single-scattering albedo and cloud optical thickness using the delta-Eddington approximation [Joseph et al., 1976; King and Harshvardhan, 1986].



The penetrative convection originating in the boundary layer is parameterized using the Relaxed Arakawa-Schubert (RAS) scheme [Moorthi and Suarez, 1992], which is a simple and efficient implementation of the Arakawa-Schubert [1974] scheme. Unlike the Arakawa-Schubert scheme, which solves an adjustment problem by considering simultaneous interaction among all possible cloud types, RAS considers only one cloud at a time, and rather than adjusting fully every hour or two, it does a series of partial adjustments that tend to relax the state toward equilibrium. The AGCM also includes a parameterization that models the evaporation of falling convective rain as described in Sud and Molod [1988]. Negative values of specific humidity produced by the finite-differenced advection are filled by borrowing from below.



The planetary boundary layer (PBL) is explicitly resolved in a 2 to 4 layer region. Wind, temperature and humidity profiles in an "extended" surface layer (which can be up to 150m thick), and the turbulent fluxes of heat, moisture, and momentum at the surface are obtained from Monin-Obukov similarity theory by selecting similarity functions that approach the convective limit for unstable profiles and that agree with observations for very stable profiles. Surface roughness lengths are taken as functions of vegetation type over land and as a function of surface stress over water. Turbulent fluxes above the "extended" surfaced layer are computed using the second order closure model of Helfand and Labraga [1988]. In this scheme, the turbulent kinetic energy is a prognostic variable, and the remaining second order moments are diagnosed from it and the atmospheric sounding.



The topography used in GEOS-1 was prepared from the 10 minute topography map of the Navy Fleet Numerical Oceanography Center in Monterey. The 2 degree latitude by 2.5 degree longitude elevation values were obtained by averaging the high resolution values (areas with more than 60% water were considered water points), and then applying a Lanczos [1966] filter. The Lanczos filter was designed to remove small scale structure (it completely removes 2 DX waves) while minimizing the Gibbs phenomena.

This version of the AGCM is run without a land surface model. For the assimilation described here, soil moisture is computed off-line based on a simple bucket model using monthly mean observed surface air temperature and precipitation [Schemm et al., 1992]. The snow line and surface albedo are prescribed and vary with the season. The sea surface temperature is updated according to the observed monthly mean values provided by the Climate Analysis Center at NMC and the Center for Ocean, Land and Atmosphere (COLA) at the University of Maryland. Long-term plans call for the incorporation of both land-surface and ocean models.


Processing Changes:


There was a minor change in the output format, where the sea-ice flag was added after 06Z Nov 16, 1985 (See data_problems). The information prior to this date is available upon request from:

Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, Feb 18, 2000.


Special Corrections/Adjustments:

Not Applicable

Calculated Variables:

Since this dataset is the output of a model, technically all variables are calculated. However, the DAO makes a distinction between PROGNOSTIC and DIAGNOSTIC variables (See Glossary). The PROGNOSTIC tend to be those which the model predicts and hence can be adjusted to match observations, when available. The DIAGNOSTIC variables should be considered calculated, in the sense that they are not assimilated from direct observation. They are estimates of the physical processes operating in nature which are generated by the GCM's physical parameterizations in a manner consistent with the observations. For a list of those observed variables which are assimilated, see the KEY VARIABLE list.

Graphs and Plots:

In the DAO anonymous FTP directory are some postscript files which graphically summarize the input data. (See OBSERVATIONS)

10. Errors:

Sources of Error:

The two primary sources of error are observational errors and errors in the first guess (forecast errors).

Quality Assessment:

Data Validation by Source:

As part of the assimilation, a comprehensive statistical analysis is carried out which monitors the statistics of the analysis increments, and other quality control information. Further validation efforts are part of an extensive set of "post-processing" activities both within and outside of the DAO. These activities include intercomparisons with independent satellite and in situ observations, and other analysis products, as well as, the production of monthly summaries of outlier counts and locations (see /pub/assimilation/e0054A/documentation/suspic.tyymm on hera) based on the variance at each grid point of the assimilated prognostic fields. An example of some of the intercomparison work is shown in Schubert et al. 1993. Further comparisons between model simulations and assimilations are being done to help determine the potential impact of model bias on the assimilation.

The file /pub/assimilation/e0054A/documentation/data_problems on contains a summary of problems discovered with the a ssimilated data. This file will be updated periodically.

Confidence Level/Accuracy Judgement:

An important quantity is the OI estimate of the analysis error variance. While this is a rather crude estimate of the true analysis error, it is valuable for assessing the impact of poor data coverage. Maps of these error estimates provide a detailed picture of the regions which have a recent history of poor data coverage.

Measurement Error for Parameters:

The estimated analysis error, mentioned in the previous paragraph, is provided in the timeseries data as the HEIGHT ERROR FIELD. This parameter is the root mean square of the OI's estimate of the analyzed height error variance.

Additional Quality Assessments:

Not Applicable

Data Verification by Data Center:


11. Notes:

Limitations of the Data:

Significant gaps in the input data were recorded in the files "gaps.yy". The following "summary.yy" files present a more detailed list of missing input data during the year yy.

Known Problems with the Data:

See DATA_PROBLEMS, which describes known problems with the assimilated (output) data.

Usage Guidance:

Two PostScript documents are available, written by the DAO which give further information concerning this dataset and the model itself.

Any Other Relevant Information about the Study:

Instances where model output appears suspicious are recorded in the DAO anonymous FTP site.

12. Application of the Data Set:

This dataset serves as a comprehensive description of the climate through an extended period of time. Its intended applications are for the study of climate and its variability, and the evaluation of the the GEOS-1 analysis system.

13. Future Modifications and Plans:

It is anticipated that the assimilated dataset will eventually span the time period January 1979 through December 1993.
The DAO is pursuing long-range data assimilation goals using estimation theory based on Kalman filtering (KF). Estimation theory provides a framework in which to develop algorithms that are suboptimal compared to the KF, yet are improvements over optimal interpolation (OI) and are economically affordable. Such an approach provides an environment in which to incrementally develop systems with a maximum amount of internal physical and chemical consistency. One of the strengths of the DAO effort, and retrospective analyses in general, is that there is no operational constraint to only being able to use observations available up to the analysis time. We can wait for data that are reported late (often from important remote locations), as well as use data collected long after the analysis time. In addition, observations that define the boundary conditions can be obtained. Using data collected after the analysis time more uniformly brackets the analysis time with observations and in theory should improve the quality of the analysis. Estimation theory provides the best framework for developing these retrospective analysis techniques. We also intend to exploit advanced numerical techniques in both model and analysis development to reduce the numerical artifacts of the system.


To guide the development effort we intend to produce data sets and apply them to generalized Earth-science problems. To gain the needed diversity, we intend to distribute the data sets to a wide range of users. An important component of this effort is DAO's involvement in field campaigns to study particular Earth-system processes. Therefore, in addition to the baseline assimilation described above, we plan to carry out shorter assimilations supporting several special observing periods and testing the sensitivity of the results to the analysis system formulation and observational data. These include experiments employing the latest HIRS2 physical retrievals produced in-house [Susskind et. al., 1984], and assimilations in support of the Tropical Ocean Global Atmosphere (TOGA) subproject Coupled Ocean-Atmosphere Response Experiment (COARE), and subprojects of the Global Energy and Water Cycle Experiment (GEWEX).


An important activity in the DAO is stratospheric assimilation. There are currently two configurations to the GEOS-1 data assimilation system. One is the tropospheric system described above. The other is a stratospheric system which employs a 46 level (top pressure at 0.1 mb) version of the GEOS model. The analysis is carried out at 19 pressure levels extending to 0.4 mb. Due to the computational constraints, this system is typically run at a horizontal resolution of 4 degree latitude by 5 degree longitude. This system, which has been published under the name STRATAN, has been used in numerous stratospheric chemistry and meteorological studies [e.g. Rood et al., 1992]. Recently forecasts and analyses from STRATAN have been supplied operationally to the Stratospheric Photochemistry, Aerosols, and Dynamics Expedition (SPADE). The DAO products were used for scientific flight planning and will be used in a wide range of interpretive studies. Previously an older version of STRATAN was used in the Airborne Arctic Stratosphere Expedition in 1989. The participation in these missions has pushed the quality of the stratospheric analysis forward tremendously, and it is this experience that motivates DAO participation in a wider range of applications. The DAO also plans to assimilate data from the Upper Atmosphere Research Satellite (UARS) which has temperature, wind, and chemical constituent measurements. In many ways UARS provides a prototype for the EOS effort, making UARS a very important mission to study prior to the launch of EOS satellites.


Various improvements are planned for both the assimilating AGCM and the analysis scheme. In the short term the model improvements include a more accurate moisture advection scheme, further improvements to the PBL, convection and radiation parameterizations, including the introduction of a land surface model, and a cloud liquid water scheme. The OI scheme is currently being modified to use a variational approach to solve the OI analysis equations; the method solves the system globally, thus eliminating the need to perform data selection. Longer term developments include the introduction of semi-lagrangian dynamics [Bates et al., 1993], a coupled ocean model, and a simplified Kalman filtering scheme.

14. Software:

Software Description:

A sample READ program for data sets is:
       PARAMETER (IM=144,JNP=91)
  • IM is the number of longitude gridpoints
  • JNP is the number of latitude gridpoints
  • NTIMES is the number of time values data was written
  • NXX is either the number of levels in the file for upper air quantities OR it is the number of variables in the file for single level quantities. For upper air files, the first level is the lowest altitude (1000mb).

For each DATA FILE there is an associated TABLE FILE which provides which values to use for these parameters. The TABLE FILE has the same name as the DATA FILE except the part of the name ".prs." is replaced by ".tabl.".



  • NTIMES is provided by TDEF (number of time steps)
  • NXX is provided by ZDEF for upper air files (vertical levels)
  • NXX is provided by VARS for single level files (parameters)
  • IM is provided by XDEF (longitude)
  • JNP is provided by YDEF (latitude)


Software Access:

A program written by the DAAC to read the upper air data products, based on this template is available: UPPER AIR SAMPLE READ PROGRAM. A program to read the single level data products, based on this template is also available: SINGLE LEVEL SAMPLE READ PROGRAM. Both programs must be modified to generate the desired output.

Assign statements (for Cray)

  1. cray data (iau and restart files)
    assign -a $data1 fort.21
  2. ieee data (all other files)
    assign -a $data1  -N ieee -F f77 fort.21

A sample IDL procedure, which produces postscript plots of the output from Ziskin's READ program is called: A sample plot of the Sea Level Pressure is provided in SAMPLE DATA RECORD

15. Data Access:

Contacts for Archive/Data Access Information:

Contact Information:

GSFC DAAC User Services Office
Goddard DAAC Code 610.2 NASA/Goddard Space Flight Center GREENBELT MD 20771
Electronic Mail Address:

Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, Feb 18, 2000.

Data Center Identification:

Goddard Space Flight Center (GSFC) Distributed Active Archive Center (DAAC)

Procedures for Obtaining Data:

For information on how to obtain data via the Goddard DAAC

Data Center Status/Plans:

The timeseries currently consists of five complete years (1985-1989), but will soon be expanded to include three more years (1990-1992). A subset of the timeseries has been placed on anonymous FTP.

16. Output Products and Availability:

Access the subset of the timeseries on Anonymous FTP

17. References:

PostScript Documents:If your environment supports a PostScript viewer (e.g. xmosaic/ghostview) you may link to these documents directly. Otherwise, these files should be retrieved via anonymous FTP.

Arakawa, A., and W. Schubert, 1974: Interaction of a cumulus ensemble with the large-scale environment, Part I, J.. Atmos. Sci., 31, 674-701.

Arakawa, A., and M.J. Suarez, 1983: Vertical differencing of the primitive equations in sigma coordinates. Mon. Wea. Rev., 111, 34-45.

Bates, J. R., S. Moorthi and R. W. Higgins, 1993: A global multilevel atmospheric model using a vector semi-Lagrangian finite-difference scheme. Part I: Adiabatic formulation. Mon. Wea. Rev., 121, 244-263.

Bengtsson, L. and J. Shukla, 1988: Integration of space and in situ observations to study global climate change. Bull. Amer. Meteor. Soc., 69, 1130-1143.

Bloom, S. C., L. L. Takacs and E. Brin, 1991: A scheme to incorporate analysis increments gradually in the GLA assimilation system. Ninth Conference on Numerical Weather Prediction, Denver, CO.

Burridge, D. M. and J. Haseler, 1977: A model for medium range weather forecasting-adiabatic f ormulation. Tech. Rep. No. 4, European Centre for Medium Range Weather Forecasts, Bracknell, Berks, U. K., 46pp.

Chou, M.D., 1984: Broadband water vapor transmission functions for atmospheric IR flux computations. J. Atmos. Sci. , 41, 1775-1778.

Chou, M.D.,and L. Peng, 1983: A parameterization of the absorption in the 15m CO2 spectral region with application to climate sensitivity studies. J. Atmos. Sci. , 40, 2183-2192.

Davies, R., 1982: Documentation of the solar radiation parameterization in the GLAS climate model. NASA Tech. Memo. 83961, 57pp., Goddard Space Flight Center, Greenbelt, MD 20771.

Harshvardhan, R. Davies, D.A. Randall, and T.G. Corsetti, 1987: A fast radiation parameterization for atmospheric circulation models. J. Geophys. Res., 92, 1009-1016.

Helfand, H.M. and J.C. Labraga, 1988: Design of a non-singular level 2.5 second-order closure model For the prediction of atmospheric turbulence. J. Atmos. Sci., 45, 113-132.

Helfand, H. M. and S. D. Schubert, 1993: Contribution of the Great Plains low-level jet to the simulated continental moisture budget of the United States. J. Climate, (submitted).

Joseph, J.H., W.J. Wiscombe and J.E. Weinman, 1976: The delta-Eddington approximation for radiative flux transfer. J. Atmos Sci., 33,2452-2459.

Kalnay, E. and R. Jenne, 1991: Summary of the NMC/NCAR reanalysis workshop of April 1991. Bull. Amer. Meteor. Soc., 72, 1897-1904.

King, M.D. and Harshvardhan, 1986: Comparitive accuracy of selected multiple scattering approximations. J. Atmos. Sci., 43, 784-801.

Lacis, A.A. and J.E. Hansen, 1974: A parameterization for the absorption of solar radiation in the Earth's atmosphere. J. Atmos. Sci. , 31, 118-133.

Lanczos, C., 1966: Discourse on Fourier Series. Hafner Publishing.

Moorthi, S. and M. J. Suarez, 1992: Relaxed Arakawa-Schubert: A parameterization of moist convection for general circulation models. Mon. Wea. Rev., 120,978-1002.

National Research Council, 1991: Four-dimensional model assimilation of data: a strategy for earth system sciencs. Report from the Panel on Model-Assimilated Data Sets for Atmospheric and Oceanic Research, National Academy Press, Washington D.C., 1991.

Rodgers, C.D., 1968: Some extensions and applications of the new random model for molecular band transmission. Quart. J. R. Met. Soc., 94, 99-102.

Rood, R. B., J. E. Nielsen, R. S. Stolarski, A. R. Douglass, J. A. Kaye and D. J. Allen, 1992: Episodic total ozone minima and associated effects on heterogeneous chemistry and lower stratospheric transport. J. Geophys. Res., 97, 7979-7996.

Rosenfield, J.E., M.R. Schoeberl, and M.A. Geller, 1987: A computation of the stratospheric diabatic circulation using an accurate radiative transfer model. J. Atmos. Sci. , 44, 859-876.

Sadourney, R., 1975: The dynamics of finite difference models of the shallow water equations. J. Atmos. Sci. , 32, 680-689.

Schemm, J.-K., S. Schubert, J. Terry and S. Bloom, 1992: Estimates of monthly mean soil moisture for 1979-89, NASA Tech. Memo. 104571, pp 252, Oct. 1992.

Sud, Y. and A. Molod, 1988: The roles of dry convection, cloud-radiation feedback processes and the influence of recent improvements in the parameterization of convection in the GLA AGCM. Mon. Wea. Rev., 116., 2366-2387.

Susskind, J., J. Rosenfield, D. Reuter and M. T. Chahine, 1984: Remote sensing of weather and climate parameters from HIRS2/MSU on TIROS-N, J. Geophys. Res., 89, 4677- 4697.


18. Glossary of Terms:

Assimilation is the process of combining observations and model first guess fields. See Optimal Interpolation.
A Diagnostic parameter is an inference of the processes occuring in the climate system. These parameters are generally not measured, but are calculated by the model's physical parameterizations in a manner consistent with the observations.
A Granule is the smallest sub-division of data. In DAO 4D TIMESERIES data set each granule is either a single parameter at all pressure levels or a collection of surface variables. The temporal coverage of each granule is one month.
A Prognostic parameter is an atmospheric state variable which the model forecasts. During the assimilation these are the parameters most directly influenced by the observations. See Optimal Interpolation for a list of assimilated quantities.
The timeseries is a subset of the DAO 5-year analysis project.

19. List of Acronyms:

AGCM     Atmospheric General Circulation Model
COARE    Coupled Ocean-Atmosphere Response Experiment
COLA     Center for Ocean, Land and Atmosphere at the University of Maryland
DAAC     Distributed Active Archive Center
DAO      Data Assimilation Office
ECMWF    European Center for Medium Range Forecasts
EOS      Earth Observing System
GEOS-1   Goddard Earth Observing System version 1 model
GEWEX    Global Energy and Water Cycle Experiment
GLA      Goddard Lab for Atmospheres
GSFC     Goddard Space Flight Center
HIRS2    High-Resolution Infrared Sounder 2
IAU      Incremental Analysis Update
IMS      Information Management System
KF       Kalman Filtering
LW       Longwave (same as IR)
NASA     National Aeronautics and Space Administration
NCAR     National Center for Atmospheric Research
NCDC     National Climatic Data Center
NESDIS   National Environmental Satellite Data and Information Service
NH       Northern Hemisphere
NMC      National Meteorological Center 
NOAA     National Oceanic and Atmospheric Administration
NRC      National Research Council
OI       Optimal Interpolation
PBL      Planetary Boundary Layer
RAS      Relaxed Arakawa-Schubert
RHS      Right Hand Side (e.g. of an equation)
SPADE    Stratospheric Photochemistry, Aerosols, and Dynamics Expedition
STRATAN  Stratospheric version of GEOS-1
SH       Southern Hemisphere
SW       Shortwave (same as VISIBLE)
TIROS    Television Infrared Observing Satellite
TOA      Top Of the Atmosphere
TOGA     Tropical Ocean Global Atmosphere
TOVS     TIROS Operational Vertical Sounder
UARS     Upper Atmosphere Research Satellite 
URL		Uniform Resource Locator

20. Document Information:

Document Revision Date:Fri May 10 11:51:45 EDT 2002
OCTOBER 12, 1995

Document Review Date:


Document ID:

Intentionally left blank


Intentionally left blank

Document Curator:

Document URL:


Change History

Version 2.0
Version baselined on addition to the GES Controlled Documents List, July 6, 1994.
Document Actions
NASA Logo -
NASA Privacy Policy and Important Notices
Last updated: Jul 21, 2010 11:38 AM ET