List of MERRA met fields

From Geos-chem
Jump to: navigation, search

Obsolete.jpg

The MERRA reanalysis product ceased in early 2016 and is now obsolete. We recommend that you use either GEOS-FP, which is the current operational GMAO met data product, or MERRA-2, which is the current GMAO reanalysis product. MERRA will be de-supported in GEOS-Chem v11-02.

On this page we list the GMAO MERRA met fields that are currently being archived for use with GEOS-Chem. This is only a subset of the entire MERRA met field data archive.

For more information about the MERRA data archive, please consult the following resources:

  1. MERRA File Specification Document.
  2. MERRA data subset requests (you can request data fields that have been "cut down" to size for your desired region)
  3. MERRA grid structure
  4. MERRA file naming convention
  5. MERRA data file collections (aka file types)
  6. MERRA time archiving
  7. MERRA mainstream and spinup data
  8. Details about the implementation of MERRA into GEOS-Chem
  9. GEOS-Chem projects using MERRA met data
  10. Version history of GMAO met data products


Constant fields

  • These are time-invariant constant fields, stored in the files 20000101.cn.2x25; 20000101.cn.4x5
  • For convenenice, a "fake" timestamp of 2000/01/01 0GMT has been assigned to each field in the file.
  • All constant fields are 2-dimensional (lon,lat)
  • For more information about the native MERRA grids, see the MERRA grid structure wiki post.
  • The GEOS-5 Field and GEOS-5 Units refer to the equivalent data fields (and units) in the GEOS-5 data archive for GEOS-Chem. If a field is missing it does not necessarily mean that it was not produced in GEOS-5; it just means that we did not regrid it for use with GEOS-Chem.


MERRA
Field
MERRA
Units
Description Native
Grid
Notes GEOS-5
Field
GEOS-5
Units
FRLAKE fraction Fraction of lake type in grid box Fx FRLAKE fraction
FRLAND fraction Fraction of land type in grid box Fx FRLAND fraction
FRLANDIC fraction Fraction of land ice type in grid box Fx FRLANDIC fraction
FROCEAN fraction Fraction of ocean in grid box Fx FROCEAN fraction
PHIS m2 s-2 Surface geopotential Fx PHIS m2 s-2

--Bob Y. 15:30, 11 August 2010 (EDT)

A1 fields

  • These are 1-hr time-averaged fields, stored in files YYYYMMDD.a1.2x25; YYYYMMDD.a1.4x5
  • All A1 fields are 2-dimensional (lon,lat)
  • Timestamps (i.e. center of averaging period): 00:30, 01:30, 02:30 ... 23:30 GMT
  • For more information about the native MERRA grids, see the MERRA grid structure wiki post.
  • The GEOS-5 Field and GEOS-5 Units refer to the equivalent data fields (and units) in the GEOS-5 data archive for GEOS-Chem. If a field is missing it does not necessarily mean that it was not produced in GEOS-5; it just means that we did not regrid it for use with GEOS-Chem.


MERRA
Field
MERRA
Units
Description Native
Grid
Notes GEOS-5
Field
GEOS-5
Units
ALBEDO fraction Surface albedo Nx ALBEDO fraction
CLDTOT fraction Total cloud fraction Nx CLDTOT fraction
EFLUX W m-2 Latent heat flux (positive upward) Nx EFLUX W m-2
EVAP kg m-2 s-2 Surface evaporation Nx EVAP kg m-2 s-2
FRSEAICE fraction Fraction of sea ice Nx Used to create the binned SEAICE** fields listed below - -
FRSNO fraction Fractional snow-covered area Nx - -
GRN fraction Vegetation greenness fraction Nx Currently not used GRN fraction
GWETROOT fraction Root zone soil wetness Nx GWETROOT fraction
GWETTOP fraction Top soil (first 2 cm) soil wetness Nx GWETTOP fraction
HFLUX W m-2 Sensible heat flux (positive upward) Nx HFLUX W m-2
LAI m2 m-2 Leaf area index Nx LAI fraction
LWGNT W m-2 Surface net downward longwave flux Nx LWGNET W m-2
LWI unitless Land/water/ice flags Nx See below for detailed explanation
Needed for backwards compatibility
LWI unitless
LWTUP W m-2 Upward longwave flux at top of atmosphere (TOA) Nx Currently not used LWTUP W m-2
PARDF W m-2 Surface downward PAR diffuse flux Nx PARDF W m-2
PARDR W m-2 Surface downward PAR beam flux Nx PARDR W m-2
PBLH m Planetary boundary layer height above surface Nx PBLH m
PRECANV kg m-2 s-2 Surface precipitation flux from anvils Nx PRECANV kg m-2 s-2
PRECCON kg m-2 s-2 Surface precipitation flux from convection Nx PRECCON kg m-2 s-2
PRECLSC kg m-2 s-2 Surface precipitation flux from large-scale Nx PRECLSC kg m-2 s-2
PRECSNO kg m-2 s-2 Surface snowfall flux Nx PRECSNO kg m-2 s-2
PRECTOT kg m-2 s-2 Total surface precipitation flux Nx = PRECANV + PRECCON + PRECLSC PRECTOT kg m-2 s-2
QV2M kg kg-1 Specific humidity at 2 m above the displacement height Nx QV2M kg kg-1
SEAICE00 fraction Fraction of grid box that has < 10% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE10 fraction Fraction of grid box that has 10-20% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE20 fraction Fraction of grid box that has 20-30% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE30 fraction Fraction of grid box that has 30-40% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE40 fraction Fraction of grid box that has 40-50% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE50 fraction Fraction of grid box that has 50-60% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE60 fraction Fraction of grid box that has 60-70% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE70 fraction Fraction of grid box that has 70-80% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE80 fraction Fraction of grid box that has 80-90% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SEAICE90 fraction Fraction of grid box that has > 90% sea ice coverage Nx Created by separating FRSEAICE into bins - -
SLP hPa Sea level pressure Nx Converted from Pa --> hPa
for compatibility w/ GEOS-Chem
SLP hPa
SNODP m Snow depth Nx SNODP m
SNOMAS kg m-2 Snow mass Nx See below for detailed explanation
Also note: 1 kg m-2 = 1 mm liquid H2O equivalent
SNOMAS mm H2O
SWGDN W m-2 Surface incident shortwave flux Nx - -


SWGNT W m-2 Net surface downward shortwave flux Nx SWGNET W m-2
TROPPT hPa T based tropopause pressure Nx Converted from Pa --> hPa
for compatibility w/ GEOS-Chem
TROPP hPa
TS K Surface skin temperature Nx TS K
T2M K Temperature 2 m above displacement height Nx Used as a proxy for surface air temperature in GEOS-5 T2M K
U10M m s-1 Eastward wind 10 m above displacement height Nx U10M m s-1
USTAR m s-1 Friction velocity Nx USTAR m s-1
V10M m s-1 Northward wind 10 m above displacement height Nx V10M m s-1
Z0M m Roughness length, momentum Nx Z0M m

--Bob Y. 16:41, 10 March 2011 (EST)

A3 fields

  • These are 3-hr time-averaged fields, stored in files YYYYMMDD.a3.2x25; YYYYMMDD.a3.4x5
  • All A3 fields are 3-dimensional (lon,lat,alt)
  • Timestamps (i.e. center of averaging period): 01:30, 04:30, 07:30, 10:30, 13:30, 16:30, 19:30, 22:30 GMT
  • For more information about the native MERRA grids, see the MERRA grid structure wiki post.
  • The GEOS-5 Field and GEOS-5 Units refer to the equivalent data fields (and units) in the GEOS-5 data archive for GEOS-Chem. If a field is missing it does not necessarily mean that it was not produced in GEOS-5; it just means that we did not regrid it for use with GEOS-Chem.


MERRA
Field
MERRA
Units
Description Native
Grid
Notes GEOS-5
Field
GEOS-5
Units
CLOUD fraction Total cloud fraction in grid box Fv = CFAN + CFCU + CFLS;
Regridded w/ Hongyu Liu's algorithm
CLOUD fraction
CMFMC kg m-2 s-2 Upward moist convective mass flux Fe CMFMC kg m-2 s-2
DQRCU kg kg-1 s-1 Precipitation production rate – convective Cp P -> hybrid interpolation necessary - -
DQRLSAN kg kg-1 s-1 Precipitation production rate - large-scale+anvil Cp P -> hybrid interpolation necessary - -
DQIDTMST kg kg-1 s-1 Ice tendency from moist physics Cp P -> hybrid interpolation necessary DQIDTMST kg kg-1 s-1
DQLDTMST kg kg-1 s-1 Liquid water tendency from moist physics Cp P -> hybrid interpolation necessary DQLDTMST kg kg-1 s-1
DQVDTMST kg kg-1 s-1 Water vapor tendency from moist physics Cp P -> hybrid interpolation necessary DQVDTMST kg kg-1 s-1
DTRAIN kg m-2 s-2 Detrainment cloud mass flux Fv DTRAIN kg m-2 s-2
MOISTQ g kg-1 day-1 Overall tendency of QV N/A = DQIDTMST + DQLDTMST + DQVDTMST MOISTQ g kg-1 day-1
OPTDEPTH unitless Total in-cloud optical depth (visible band) Fv = TAUCLI + TAUCLW;
Regridded w/ Hongyu Liu's algorithm
OPTDEPTH unitless
PFICU kg m-2 s-1 Downward flux of ice precipitation – convective Cp P -> hybrid interpolation necessary - -
PFILSAN kg m-2 s-1 Downward flux of ice precip - large-scale+anvil Cp P -> hybrid interpolation necessary - -
PFLCU kg m-2 s-1 Downward flux of liquid precipitation – convective Cp P -> hybrid interpolation necessary - -
PFLLSAN kg m-2 s-1 Downward flux of liquid precip - large-scale + anvil Cp P -> hybrid interpolation necessary - -
QI kg kg-1 Cloud ice mixing ratio Fv QI kg kg-1
QL kg kg-1 Cloud liquid water mixing ratio Fv QL kg kg-1
QV kg kg-1 Specific humidity Fv QV kg kg-1
REEVAPCN kg kg-1 s-1 Evaporation of precipitating convective condensate Cp P -> hybrid interpolation necessary - -
REEVAPLS kg kg-1 s-1 Evaporation of precipitating LS & anvil condensate Cp P -> hybrid interpolation necessary - -
T K Air temperature Fv T K
TAUCLI unitless In-cloud ice cloud optical thickness (visible band) Fv Regridded w/ Hongyu Liu's algorithm TAUCLI unitless
TAUCLW unitless In-cloud liquid cloud optical thickness (visible band) Fv Regridded w/ Hongyu Liu's algorithm TAUCLW unitless
U m s-1 Eastward component of wind Fv U m s-1
V m s-1 Northward component of wind Fv V m s-1

--Bob Y. 16:41, 16 August 2010 (EDT)

I6 fields

  • These are 6-hr instantaneous fields, stored in files YYYYMMDD.i6.2x25; YYYYMMDD.i6.4x5
  • I6 fields are either 2-dimensional (lon,lat) or 3D (lon,lat,alt)
  • Timestamps: 00:00, 06:00, 12:00, 18:00 GMT
  • For more information about the native MERRA grids, see the MERRA grid structure wiki post.
  • The GEOS-5 Field and GEOS-5 Units refer to the equivalent data fields (and units) in the GEOS-5 data archive for GEOS-Chem. If a field is missing it does not necessarily mean that it was not produced in GEOS-5; it just means that we did not regrid it for use with GEOS-Chem.
MERRA
Field
MERRA
Units
Description Native
Grid
# Dims Notes GEOS-5
Field
GEOS-5
Units
OMEGA Pa/sec Vertical pressure velocity Cp 3D P -> hybrid interpolation necessary;
OMEGA is mostly used for offline analysis
OMEGA hPa
PS hPa Surface pressure Nx 2D Converted from Pa --> hPa
for compatibility w/ GEOS-Chem
PS hPa
PV K m2 kg-1 s-1 Ertel potential vorticity Cp 3D P -> hybrid interpolation necessary;
PV is mostly used for offline anaylsis
PV K m2 kg-1 s-1
RH unitless Relative humidity Cp 3D P -> hybrid interpolation necessary

RH is calculated with respect to liquid above 273.16 K, with respect to ice below 253.16 K, and is a weighted average of both at intermediate temperatures.

RH unitless

--Bob Y. 12:37, 15 March 2011 (EDT)

Special processing for certain fields

Fractional sea ice

The MERRA "FRSEAICE" data field is the fraction of each grid box that is covered by sea ice. This will be an important field for GEOS-Chem simulations for certain tracers (e.g. Hg) of the Arctic and Antarctic regions.

However, our normal regridding process from the MERRA native 0.5° x 0.666° grid to the GEOS-Chem 2° x 2.5° and 4° x 5° grids causes important information about sea ice leads (i.e. the "fingers" of ice that stretch out into open water) to be lost.

In order to preserve some information about fractional sea ice, we have binned the fractional sea ice coverage into deciles. We have created the new met fields:

  1. SEAICE00: Fraction of each grid box that has < 10% sea ice coverage
  2. SEAICE10: Fraction of each grid box that has 10-20% sea ice coverage
  3. SEAICE20: Fraction of each grid box that has 20-30% sea ice coverage
  4. SEAICE30: Fraction of each grid box that has 30-40% sea ice coverage
  5. SEAICE40: Fraction of each grid box that has 40-50% sea ice coverage
  6. SEAICE50: Fraction of each grid box that has 50-60% sea ice coverage
  7. SEAICE60: Fraction of each grid box that has 60-70% sea ice coverage
  8. SEAICE70: Fraction of each grid box that has 70-80% sea ice coverage
  9. SEAICE80: Fraction of each grid box that has 80-90% sea ice coverage
  10. SEAICE90: Fraction of each grid box that has > 90% sea ice coverage

Bob Yantosca has created a PowerPoint document that shows the daily evolution of each of these fractional sea ice fields. Please see:

LWI

The GEOS-5 contained a field named LWI, which were flags to denote land, water, and ice coverage. LWI was intended for backwards compatibility with older GMAO met field products. It is largely superseded by the FRLAKE, FRLAND, FRLANDICE, FROCEAN, FRSEAICE fields.

However, some GEOS-Chem routines (in particular the IS_LAND, IS_WATER, IS_ICE functions in dao_mod.f depend on LWI. Therefore, to facilitate backwards compatibility with these existing GEOS-Chem routines, we have created the LWI field from the MERRA FRLAKE, FRLAND, FRLANDICE, FROCEAN, and FRSEAICE fields as follows:

  1. On the 0.5° x 0.666° "Nx" grid:
    • All ocean boxes (where FROCEAN > 0) are assigned a value of 0.
    • All land boxes (where FRLAND + FRLANDICE + FRLAKE > 0) are assigned a value of 1.
    • All grid boxes with more than 50% ice coverage(i.e. where FRSEAICE > 0.5) are assigned a value of 2.
  2. Use the same algorithm for regridding the GEOS-5 LWI data to coarser resolution:
    • The LWI value on the coarse grid (e.g. 2° x 2.5° or 4° x 5°) will be the mode (i.e. most frequently occurring value) of the values of all the "fine" grid boxes that fit into the "coarse" grid box.
    • Tiebreakers" are handled as follows (by the F90 MAXLOC function):
      1. If a coarse grid box consists of exactly 50% water and 50% ice, it will be considered water (LWI =0).
      2. If a coarse grid box consists of exactly 50% land and 50% ice, it will be considered land (LWI=1).

Bob Yantosca has prepared a document containing plots of the GEOS-5 and MERRA LWI fields. Please see:

SNOMAS

The SNOMAS field in MERRA differs from that in the GEOS-5 operational data:

From the GEOS-5 File Specification Document:

SNOMAS: The mass of snow in per unit of land area in meters of liquid-water-equivalent depth (i.e., 10^3 kg/m2). In grid boxes with no land (FRLAND+FRLANDICE=0) it is set to _FillValue (= 1e15). Where FRLANDICE > 0.9 it is arbitrarily set to 4 meters. Over other land areas it represents an average over the non-glaciated part.

From the MERRA File Specification Document:

SNOMAS: The mass of snow per unit of ice-free land area (FRLAND), in kg/m2. In grid boxes with no land it is set to _FillValue (=1e15). Over other land areas it represents an average over the nonglaciated part.

Max Suarez wrote:

Early versions of GEOS had been writing SNOMAS in meters. This was changed to mm (or kg/m^2) in all recent versions, including 5_2, MERRA, and the current development tags. But the forward processing spec was not updated. The MERRA spec, however, is correct.
To further complicate matters, the variable called SNOMAS in MERRA comes from a very different part of the code that in 5_x. It is in a land collection intended to have representative values over the ice-free land portion of the grid box. This applies to all variables in that collection. SNOMAS in particular makes this clear in the glossary definition in the spec.
Forward processing (FP) puts out a grid averaged SNOMAS, including ice-covered areas, where the "SNOMAS" was arbitrarily set to 4000 mm. Neither FP nor MERRA includes ocean or freshwater regions with snow over ice.
To answer your question, to convert MERRA SNOMAS to 5.2 SNOMAS, use this equation:
       SNOMAS_5.X = ( SNOMAS_merra * FRLAND_merra  +  4000 * FRLANDICE_merra ) /
                    ( FRLAND_merra                 +  FRLANDICE_merra        )

Therefore, we shall implement the algorithm that Max Suarez described above. This will make the output SNOMAS field similar to GEOS-5, which will allow better backward compatibility w/ existing code.

Bob Yantosca has prepared a document containing plots of the GEOS-5 and MERRA SNOMAS fields. Please see:

--Bob Y. 12:54, 27 August 2010 (EDT)

Validation plots

"Sanity check" plots

Bob Yantosca has produced the following "sanity check" validation plots of GEOS-5 vs. MERRA met fields. These are not a detailed analysis but rather a quick way to determine if there were major errors in the MERRA data processing code.

--Bob Y. 15:38, 2 September 2010 (EDT)