Difference between revisions of "GEOS-FP"

Revision as of 18:38, 28 May 2020

This page contains some basic information about GMAO's GEOS-FP ("forward-processing") data product.

GEOS FP ("forward processing") is the current operational met data product from NASA/GMAO. GEOS-FP has finer native horizontal resolution (0.25° lat x 0.3125° lon) and temporal resolution (hourly data and 3-hourly data) than older GMAO met products such as GEOS-5.2.0, the MERRA reanalysis, and GEOS-4.

For more information on the GEOS-FP meteorology data product as well as other products from GMAO, please see:

1. GEOS-FP (v1.0) file specification document (11 Jun 2013)
2. List of GEOS-FP met fields for GEOS-Chem
3. Information about implementing GEOS-FP into GEOS-Chem
4. Overview of GMAO met data products
5. Version history of GMAO met data products
6. Glossary of variables produced by the GEOS-DAS
7. Evolution of the GEOS-5 system from GEOS-5.2.0/MERRA to GEOS-5.7.2 (aka "Fortuna")

Important notes

Before using GEOS-FP data, please thoroughly understand the following caveats which may impact your model results:

1. GEOS-FP data is generated using a model which changes frequently and data are not retroactively generated following changes. See https://gmao.gsfc.nasa.gov/GMAO_products/NRT_products.php for update boundary dates.
2. Simulations using GEOS-FP which cross an update boundary will have an inconsistent representation of the atmosphere. We therefore recommend that users avoid using GEOS-FP across update boundaries as they can generate spurious trends or results.
3. Result comparisons from simulations on either side of an update boundary will include a spurious signal due to changes in the model.
4. The update applied on January 24th 2017 is particularly significant, as the underlying model resolution was increased, affecting products such as total convective mass flux.
5. Users interested in performing long-term simulations, investigating trends, or who need to run a simulation which crosses one of these boundaries are recommended to consider using MERRA-2 meteorology instead of GEOS-FP. MERRA-2 uses a consistent model for all years (see MERRA-2)

Overview

From the GEOS-FP file specification document, Version 1.0 (11 Jun 2013), p. 1:

The GEOS FP Atmospheric Data Assimilation System (GEOS-5 ADAS) uses an analysis developed jointly with NOAA’s National Centers for Environmental Prediction (NCEP), which allows the Global Modeling and Assimilation Office (GMAO) to take advantage of the developments at NCEP and the Joint Center for Satellite Data Assimilation (JCSDA). The GEOS-5 AGCM uses the finite-volume dynamics (Lin, 2004) integrated with various physics packages (e.g, Bacmeister et al., 2006), under the Earth System Modeling Framework (ESMF) including the Catchment Land Surface Model (CLSM) (e.g., Koster et al., 2000). The GSI analysis is a three-dimensional variational (3DVar) analysis applied in grid-point space to facilitate the implementation of anisotropic, inhomogeneous covariances (e.g., Wu et al., 2002; Derber et al., 2003). The GSI implementation for GEOS-5 FP incorporates a set of recursive filters that produce approximately Gaussian smoothing kernels and isotropic correlation functions.

The GEOS-5 ADAS is documented in Rienecker et al. (2008). More recent updates to the model are presented in Molod et al. (2011). The GEOS-5 system actively assimilates roughly 2 × 10⁶ observations for each analysis, including about 7.5 × 10⁵ AIRS radiance data. The input stream is roughly twice this volume, but because of the large volume, the data are thinned commensurate with the analysis grid to reduce the computational burden. Data are also rejected from the analysis through quality control procedures designed to detect, for example, the presence of cloud.

To minimize the spurious periodic perturbations of the analysis, GEOS-FP uses the Incremental Analysis Update (IAU) technique developed by Bloom et al. (1996).

The assimilation is performed at a horizontal resolution of 0.3125-degree longitude by 0.25-degree latitude and at 72 levels, extending to 0.01 hPa. All products are generated at the native resolution of the horizontal grid. The majority of data products are time-averaged, but four instantaneous products are also available. Hourly data intervals are used for two-dimensional products, while 3-hourly intervals are used for three-dimensional products. These may be on the model’s native 72-layer vertical grid or at 42 pressure surfaces extending to 0.1 hPa.

--Bob Y. 15:00, 16 August 2013 (EDT)

Version history

See: https://gmao.gsfc.nasa.gov/GMAO_products/NRT_products.php.

Acknowledge the source of GEOS-FP data in your publications

If your GEOS-Chem research depends on the GEOS-FP meteorological data producs, please consider adding an acknowledgment to your citations, such as:

The GEOS-FP data used in this study/project have been provided by the Global Modeling and Assimilation Office (GMAO) at NASA Goddard Space Flight Center.

--Bob Yantosca (talk) 16:14, 12 February 2018 (UTC)

Developers

The following table credits the various individuals who have assisted with the development of the GEOS-FP met data product and its incorporation into GEOS-Chem:

--Bob Y. 11:27, 27 February 2014 (EST)

GEOS-FP grid structure

This section describes the horizontal and vertical grids used by the GEOS-FP data products.

Input grids

From the GEOS-FP file specification document, Version 1.0 (11 Jun 2013), p. 7:

Fields are produced on the model’s native horizontal grid, with a resolution of 5/16 degree longitude by 1/4 degree latitude.

The GEOS-FP global horizontal grid consists of IMn=1152 points in the longitudinal direction and JMn=721 points in the latitudinal direction. The horizontal native grid origin, associated with variables indexed (i=1, j=1) represents a grid point located at (180°W, 90°S). Latitude and longitude of grid points as a function of their indices (i, j) can be determined by:

     Lon(I) = -180° + [ ΔLon * ( I - 1 ) ],  I = 1, IMN
     Lat(I) =  -90° + [ ΔLat * ( J - 1 ) ],  J = 1, JMN

Where ΔLon = 5/16° = 0.3125° and ΔLat = 1/4° = 0.25°. For example, (i = 577, j = 361) corresponds to a grid point at ( 0, 0).

and also on p. 7:

Gridded products use four different vertical configurations: horizontal-only (can be vertical averages, single level, or surface values), pressure-level, model-level, or model-edge. Horizontal-only data for a given variable appear as 3-dimensional fields (x, y, time), while pressure-level, model-level, or model-edge data appear as 4-dimensional fields (x, y, z, time). In all cases the time dimension spans multiple files, as each file (granule) contains only one time. Pressure-level data is output on the LMp=42 pressure levels shown in Appendix B. The model layers used for GEOS-5 FP products are on a terrain-following hybrid sigma-p coordinate. Model-level data is output on the LM=72 layers shown in the second table of Appendix B. The model-edge products contain fields with LMe = LM + 1 levels representing the layer edges. The pressure at the model top is a fixed constant, PTOP=0.01 hPa. Pressures at model edges should be computed by summing the DELP(I,J,L) starting at PTOP. A representative pressure for the layer can then be obtained from these. In the GEOS-4 eta files, one could compute the pressure on the edges by using the “ak” and “bk” values and the surface pressure. In GEOS-5, the full 3-dimensional pressure variables are explicitly provided through (DELP(I,J,L)) and PTOP. Even though the model-level fields are on a hybrid sigma-p coordinate and their vertical location could be obtained from the “ak-bk” relationship, this may change in future GMAO systems. We thus recommend that users rely on the reported 3D pressure distribution, and not use ones computed from the “ak” and “bk”.

Note that the indexing for the GEOS-5 FP vertical coordinate system is top to bottom, i.e., layer 1 is the top layer of the atmosphere, while layer LM is adjacent to the earth’s surface. The same is true for edge variables, with level 1 being the top of the model’s atmosphere (PTOP), and level LM+1 being the surface.

The table below lists the combination of the horizontal and vertical grids onto which the GEOS-FP "raw" data products are placed:

Output grids

We download the GEOS-FP data at its native 0.25° x 0.3125° horizontal resolution. For use with GEOS-Chem, we cut and/or regrid the 0.25° x 0.3125° data to the following grids:

• 0.25° x 0.3125° China, Europe, North America nested grids with 72 hybrid layers
• 2° x 2.5° with 72 hybrid layers
• 4° x 5° with 72 hybrid layers

--Bob Yantosca (talk) 21:11, 23 November 2016 (UTC)

GEOS-FP file naming convention

The following description of the naming convention used for GEOS-FP raw data files is paraphrased from Section 5.1 and 5.2 of the GEOS5-FP file specification document, Version 1.0 (11 Jun 2013), pp 8-11:

Standard names

NOTE: For GEOS-Chem we only use the GEOS-FP assimilation product. We have therefore omitted information about the GEOS-FP forecast product.

The standard generic complete name for the assimilated GEOS-5 FP products will appear as follows:

GEOS.config.mode.collection.timestamp.file_ver.nc4

A brief description of the node fields appear below:

GEOS

Identifies output as a data assimilation system product produced by GEOS.

config

GEOS-5 runs in multiple operational configurations, targeted at different user communities. This document is specific to a single configuration, fp.

fp

Operational forward-processing assimilation tailored for mission customers and other realtime users, nominally 12 hours behind real-time using a DAS version that is updated more frequently to include the latest advancements. The current horizontal resolution is 5/16° (lon) x 1/4° (lat).

mode

GEOS-5 runs both assimilation and forecast jobs for operational forward-processing.

asm

Assimilation. Uses a combination of atmospheric data analysis and model forecasting to generate a time-series of global atmospheric quantities.

fcst

Forecast. After atmospheric data assimilation has completed for a given synoptic time, typically at 00z and 12z, a model forecast is used to generate a time-series of forecast products some number of days into the future. Five or ten days is typical.

collection

All GEOS-5 FP data are organized into file collections that contain fields with common characteristics. These collections are used to make the data more accessible for specific purposes. Fields may appear in more than one collection. Collection names are of the form freq_dims_group_HV, where the four attributes are:

freq

time-independent (const), instantaneous (instF), or time-average (tavgF), where F indicates the frequency or averaging interval and can be any of the following:

• 1 = Hourly
• 3 = 3-Hourly

dims

2d for collections with only 2-dimensional fields or 3d for collections with a mix of 2- and 3-dimensional fields.

group

A three-letter mnemonic for the type of fields in the collection. It is a lowercase version of the group designation used in the ESDT name, as listed in the next section.

HV

Horizontal and Vertical grid.

H can be:

• N: Nominal horizontal resolution on lat/lon grid. See config above.

V can be:

• x horizontal-only data (surface, single level, etc.) ; dims must be 2D
• p: pressure-level data (see above for levels) ; dims must be 3D
• v: model layer centers (see above for levels) dims must be 3D
• e: model layer edges (see above for levels) dims must be 3D

timestamp

This node defines the date and time associated with the data in the file. It has the form yyyymmdd_hhmm for assimilation files.

• yyyy - year string (e.g. , 2002)
• mm - month string (e.g.., 09 for September)
• dd - day of the month string
• hh - hour (UTC)
• mm - minute

file_ver

This denotes the file version in the form V##. Under normal conditions ## will be 01. In the event of a processing error that requires a re-processing, this number will be incremented to identify the new version of this file.

nc4

All files are in NetCDF-4 format, thus the suffix .nc4.

EXAMPLE:

GEOS.fp.asm.tavg1_2d_slv_Nx.20131015_0430.V01.nc4 

• fp: forward-processing
• asm: assimilation
• tavg1_2d_slv_Nx: time-averaged 1-hourly data, 2-dimensional, single-level parameters on the native resolution grid.
• 20131015_0430: valid time is 04:30 GMT, which represents the center point of a 1-hour time-averaging period from 04:00 GMT to 05:00 GMT.

--Bob Y. 15:42, 16 August 2013 (EDT)

ESDT names

As required by the EOSDIS system, all GEOS-5 FP products are identified by a relatively short ESDT name. While GEOS-5 FP products are not currently distributed from the GES DISC, we have retained the ESDT designations for assimilation products. This name, also known as the ShortName, is a short handle for users to access and order data products. It takes the form: DFPTFHVGGG, where: where

1. T: Time Description:
   • I = Instantaneous
   • T = Time-averaged
   • C = Time-independent
2. F: Frequency
   • 0 = Time-independent
   • 1 = Hourly
   • 3 = 3-Hourly
3. H: Horizontal Resolution
   • N = Native
4. V: Vertical Location
   • X = Two-dimensional
   • P = Pressure
   • V = model layer center
   • E = model layer edge
5. GGG: Group
   • ASM = assimilated state variables (from the IAU corrector, see Appendix A)
   • TDT = tendencies of temperature
   • UDT = tendencies of eastward and northward wind components
   • QDT = tendencies of specific humidity
   • ODT = tendencies of ozone
   • LND = land surface variables
   • FLX = surface turbulent fluxes and related quantities
   • MST = moist processes
   • CLD = cloud-related quantities
   • RAD = radiation
   • TRB = turbulence
   • SLV = single level
   • INT = vertical integrals
   • CHM = chemistry forcing
   • AER = aerosol diagnostics
   • ADG = aerosol diagnostics (extended)
   • LSF = large-scale flux
   • OCN = ocean
   • LFO = land-surface forcing
   • NAV = navigation

--Bob Y. 15:50, 16 August 2013 (EDT)

GEOS-FP data file collections

The following information is taken from the GEOS-FP file specification document, Version 1.0 (11 Jun 2013), pp. 12-13.

The GEOS-FP product is organized into the collections listed below...All data is on the model’s native horizontal grid, which is a regular latitude-longitude grid with a spacing of 1/4° in latitude and 5/16° in longitude. Horizontal arrays (1152,721) are ordered by longitude first, with the first point at the Dateline and the South Pole, with the inner index increasing eastward. All 3d collections, except the inst3_3d_asm_Np, are on the model’s native, hybrid sigma-p vertical grid.

This table shows only those GEOS-FP file collections that are required for GEOS-Chem. For a complete list of file collections, see the GEOS-FP file specification document.

As described above, the full file name for the GEOS-FP data products takes the form:

GEOS.fp.asm.CCCCCCCCCCCCCCC.YYYYMMDD_hhmm.V01.nc4

where

1. CCCCCCCCCCCCCCC is the collection name (i.e. one of the entries in the first column of the above table)
2. YYYYMMDD_hhmm is the UTC date (year/month/day, hour:minutes)

--Bob Y. 16:20, 16 August 2013 (EDT)

Future removal of the tavg3_3d_rad_Nv collection

In the near future, we may be able to remove relying on the tavg3_3d_rad_Nv collection. The only variable from this collection that we need for GEOS-Chem is CLOUD (the 3-D cloud fraction). This variable may be constructed from other variables in the tavg3_3d_cld_Nv collection as follows:

CLOUD = min( CFAN + CFLS, 1.0 )

This will allow us to save disk space by not having to download 8 files from the tavg3_3d_rad_Nv collection for each day of data that we process.

--Bob Y. 15:27, 19 August 2013 (EDT)

GEOS-FP time archiving

Raw data

The GEOS-FP raw data are archived/averaged at the following times:

1. const: Time-invariant data
2. inst3: 3-hr instantaneous data
   • Times: 00:00, 03:00, 06:00, 09:00, 12:00, 15:00, 18:00, 21:00 GMT
3. tavg1: 1-hour time-averaged data
   • Center times: 00:30, 01:30, 02:30 ... 23:30 GMT
4. tavg3: 3-hour time-averaged data
   • Center times: 01:30, 04:30, 07:30, 10:30, 13:30, 16:30, 19:30, 22:30 GMT

Regridded data

The regridded GEOS-FP data for GEOS-Chem shall keep the same temporal resolution as the GEOS-FP raw data. To prevent file sizes from becoming larger than 2GB (i.e. the maximum netCDF-3 file size), we split the regridded GEOS-FP data among the following file types:

1. CN (time-invariant data)
   • Filenames GEOS572.20110101.CN.{RESOLUTION}.nc
   • Timestamps: For convenience we assign a file timestamp of 2011/01/01 at 00:00 GMT.
2. A1 (1-hr time-averaged data)
   • Filenames GEOS572.YYYYMMDD.A1.{RESOLUTION}.nc
   • Timestamps: see image below
3. A3 (3-hr time-averaged data)
4. Timestamps: See image below
   • Filenames
     • GEOS572.YYYYMMDD.A3cld.{RESOLUTION}.nc
     • GEOS572.YYYYMMDD.A3dyn.{RESOLUTION}.nc
     • GEOS572.YYYYMMDD.A3mstC.{RESOLUTION}.nc
     • GEOS572.YYYYMMDD.A3mstE.{RESOLUTION}.nc
5. I3 (3-hr instantaneous data)
   • Filenames GEOS572.20110101.I3.{RESOLUTION}.nc
   • Timestamps: 00:00, 03:00, 06:00, 09:00, 12:00, 15:00, 18:00, 21:00 GMT.

NOTE: At present (Aug 2013), the file names still reflect the older GEOS572 nomenclature. We will remove this at a future date.

In the filenames abaove, the {RESOLUTION} tag indicates the horizontal resolution of the file. Possible values are:

1. 4x5: GEOS_Chem 4° x 5° global grid
2. 2x25 GEOS-Chem 2&deg x 2.5° global grid
3. 025x03125.CH: GEOS-Chem 0.25° x 0.3125° China nested grid
4. 025x03125.NA: GEOS-Chem 0.25° x 0.3125° North America nested grid

The A1 and A3 files use the following bins for time-averaging. The file timestamps in the data files indicate the center of the time-averaging bins.

NOTE: GEOS-FP uses the same timing scheme as MERRA, with the exception that MERRA has 6-hour instantaneous data but GEOS-FP has 3-hour instantaneous data.

Please also see our List of GEOS-FP met fields wiki page for detailed information about which fields are stored in each of these file types.

--Bob Y. 17:13, 16 August 2013 (EDT)

Implementation into GEOS-Chem

Please see our GEOS-FP implementation details wiki page for more information about how we modified GEOS-Chem for use with the GEOS-FP met field product.

--Bob Y. 15:18, 7 November 2013 (EST)

References

1. GEOS-FP file specification document, Version 1.0 (11 Jun 2013)
2. Bacmeister, J. T., M. J. Suarez, and F. R. Robertson, 2006. Rain Re-evaporation, Boundary Layer Convection Interactions, and Pacific Rainfall Patterns in an AGCM. J. Atmos. Sci., 63, 3383-3403.
3. Bloom, S., L. Takacs, A. DaSilva, and D. Ledvina, 1996: Data assimilation using incremental analysis updates. Mon. Wea. Rev., 124, 1256-1271.
4. Derber, J. C., R. J. Purser, W.-S. Wu, R. Treadon, M. Pondeca, D. Parrish, and D. Kleist, 2003: Flow-dependent Jb in a global grid-point 3D-Var. Proc. ECMWF annual seminar on recent developments in data assimilation for atmosphere and ocean. Reading, UK, 8-12 Sept. 2003.
5. Koster, R. D., M. J. Suárez, A. Ducharne, M. Stieglitz, and P. Kumar, 2000: A catchment-based approach to modeling land surface processes in a GCM, Part 1, Model Structure. J. Geophys. Res., 105, 24809-24822.
6. Molod, A., L. Takacs, M. Suarez, J. Bacmeister, I. Song, and A. Eichmann, The GEOS-5 Atmospheric General Circulation Model: Mean Climate and Development from MERRA to Fortuna, Technical Report Series on Global Modeling and Data Assimilation, 28, Ed: Max J. Suarez, NASA GMAO, Goddard Space Flight Center, Greenbelt, MD, April 2012. (PDF)
7. Rienecker, M.M., M.J. Suarez, R. Todling, J. Bacmeister, L. Takacs, H.-C. Liu, W. Gu, M. Sienkiewicz, R.D. Koster, R. Gelaro, I. Stajner, and E. Nielsen, 2008: The GEOS-5 Data Assimilation System - Documentation of Versions 5.0.1, 5.1.0, and 5.2.0. Technical Report Series on Global Modeling and Data Assimilation 104606, v27.
8. Wu, W.-S., R.J. Purser and D.F. Parrish, 2002: Three-dimensional variational analysis with spatially inhomogeneous covariances. Mon. Wea. Rev., 130, 2905-2916.

Also see the instructions for the software that we used to process the "raw" GEOS-FP data into netCDF files for GEOS-Chem.

--Bob Yantosca (talk) 16:14, 12 February 2018 (UTC)