CMIP5 Coupled Model Intercomparison Project
WCRP    World Climate Research Programme

CMIP5 - Modeling Info - Forcing Data


The following forcing data is available:

1. Recommended CMIP5 solar forcing data.
2. Greenhouse gas concentration data 08 Dec 2009.
3. CO2 emissions data. Nov 2010.
4. Emissions data for other chemical species. 14 Jan 2010.
5. Land-use data. 31 Aug 2009.
6. The AC&C/SPARC ozone database. New extended data 18 Nov 2010.
7. The AMIP sea surface temperature and sea ice datasets.
8. Specifications for CFMIP-inspired experiments.



1. Recommended CMIP5 solar forcing data.


See SOLARIS website

http://sparcsolaris.gfz-potsdam.de/cmip5.php

for information and to download the solar irradiance.



2. Greenhouse gas concentration data. 08 Dec 2009

 

The pre-industrial, historical and future (RCP) concentrations of anthropogenically affected greenhouse gases are available as global mean time series from the website at IIASA http://tntcat.iiasa.ac.at/RcpDb/

 

Click on the "download" tab at the top of the page. You will have to register and then you will see on the lower half of the page in bold red font: "CMIP5 recommended data " and under that "RCP concentrations" (which also includes the historical data). This is the data that should be used to drive your runs.

 

The recommended aerosol concentration data, which might be required by some models, is also available at the IIASA website.



3. CO2 emissions data. Nov 2010

 

CO2 emissions are available for the historical period and for each of the RCPs.


Historical CO2 emissions:


Updated Nov 2010

Two gridded CO2 emissions datasets for forcing historical runs of CMIP5 Earth Systems Models (ESMs) are provided (in netCDF format):

  • 1) fossil fuel CO2 emissions (including emissions from cement and gas flaring), and
  • 2) CO2 emissions associated with historical land use changes.
The fossil fuel dataset of monthly gridded emissions (units gC/m**2/s) can be downloaded from the IPSL site at http://dods.ipsl.jussieu.fr/cpipsl/ANDRES/ under file name CMIP5_gridcar_CO2_emissions_fossil_fuel_Andres_1751-2007_monthly_SC_mask11.nc
An ascii file named CMIP5_gridcar_CO2_emissions_fossil_fuel_Andres_1751-2007_monthly_SC.txt that contains the globally integrated emissions for every month is also provided on this IPSL site. Note that the units are PgC/month, which differs from the gridded dataset.

Annual-mean gridded CO2 emissions (units gC/m**2/s) from historical land use change can be found at the MPI site: http://www.mpimet.mpg.de/en/wissenschaft/land-im-erdsystem/wechselwirkung-klima-biogeosphaere/landcover-change-emission-data.html under file name carbon_emissions_landuse_20person.nc
An ascii file containing the globally integrated land use CO2 emissions for every year (in units of Pg/year) is 'carbon_emissions_landuse_20person_global.txt' .

Important Technical Issues   - click to see details

1) Fossil fuel CO2 emissions dataset:

The current files replace the dataset that was distributed in 2009 (by P. Cadule of IPSL), which did not contain CO2 emissions from cement and gas flaring (now provided by R. Andres of ORNL), and thus underestimating the global totals. The current dataset also employs a better methodology to estimate historical fossil fuel emissions.

However, the global CO2 emissions values in the 2009 dataset are very close to those in the current dataset (a difference of only ~ 0.05 PgC/yr, yielding a cumulative difference of less than 1 PgC). For a modeling group that has already started its historical simulation using the 2009 dataset, it is thus recommended that the simulation should be continued, but with the current fossil fuel CO2 emissions substituted as forcing, if this is still possible. That is, a rerun of the historical simulation should not be necessary.

Note, the fossil fuel emissions provided here are slightly lower than the ones provided in the RCP database maintained by IIASA at
http://tntcat.iiasa.ac.at/RcpDb/
This is because the RCP data at IIASA are based on production statistics, while the data provided at IPSL are based on national consumption statistics (as is needed for a gridded file). It is thus recommended that only the files provided here be used for a historical run, without attempting to rescale or merge these data with the RCP dataset. There will only be a very small discontinuity in emissions rate between the historical period and the future, so no special smoothing is necessary.

For those modelling groups that only need global CO2 emissions (i.e. in the absence of atmospheric transport of CO2 in their models), the ascii file provided by IPSL is the one that should be used,in order to avoid forcing inconsistencies among CMIP5 groups.

2) Land use CO2 emissions dataset:

The dataset provided at the MPI site also differs from the one provided in the RCP database at IIASA, but not because different methodologies are used (both datasets employ the same global estimates of R. Houghton of Woods Hole). The differences arise because the RCP data community decided to scale down (by a factor of ~ 0.8, or ~ 0.27 PgC in 2005) the entire historical land use dataset in order to match the values simulated by the integrated assessment models for the year 2005.

It is thus recommended that the MPI land use dataset be used for the historical CMIP5 simulations through the year 2000, but that the MPI land use CO2 emissions be scaled as follows for the years 2001-2005 in order to smoothly transition to the 2005 RCP data:

2001: 0.967
2002: 0.908
2003: 0.867
2004: 0.818
2005: 0.815

Following this scaling procedure will yield a global land use CO2 emission of ~1.2 PgC/yr for 2005, thereby matching the RCP value in that year. Note, this scaling has already been applied to the the 2001-2005 globally integrated values in the ascii file global_emissions_landuse_20person_global.txt.

Finally, for both fossil fuel and land use CO2 emissions, it is recommended that modelling groups check the global totals when the data have been adapted to their model grid and calendar. If needed, the gridded values should then be rescaled in order to match the globally integrated emissions provided in the two ascii files mentioned here. As noted above, a rescaling is required for the land use emissions in years 2001-2005 in all CMIP5 models.

The land use emissions (LUE) are given as net fluxes (including, for example, releases of CO2 from deforestation and logging and uptake due to regrowth). These emissions should be imposed only in models that do not themselves simulate LUE internally using the in CMIP5 land cover dataset (http://cmip-pcmdi.llnl.gov/cmip5/forcing.html?submenuheader=2#land-use_data).

For models that do not transport CO2 in the atmosphere, there is no need for gridded emissions, so the global means should be used.

Contacts:

Patricia Cadule <Patricia.Cadule@lsce.ipsl.fr>, for queries on the fossil fuel dataset;
Pierre Friedlingstein <pierre.friedlingstein@lsce.ipsl.fr> or Thomas Raddatz <thomas.raddatz@zmaw.de>, for queries on the land use dataset.




RCP CO2 emissions:


For Earth System Models, CMIP5 calls for an RCP 8.5 simulation with prescribed anthropogenic CO2 emissions (rather then concentrations). The emissions can be obtained from the IIASA website. On that site click on "compare" tab at the top of the page. Then select "Regions=world", "Scenarios=RCP8.5", "Variables=CO2 emissions - Total". You may either download an Excel spreadsheet at the bottom of the page or copy the global mean emissions appearing in the table below the graph.

The future CO2 emisisons are not available on a grid.



4. Emissions data for other chemical species. 14 Jan 2010

 

The pre-industrial, historical and future (RCP) emissions of other greenhouse gases and of anthropogenic aerosols are available at two sites in slightly different formats and with one site resolving the annual cycle. The basic data are supposed to be the same. Here are the site URLs:

 

IIASA: http://tntcat.iiasa.ac.at/RcpDb/

Juelich: ftp://ftp-ipcc.fz-juelich.de/pub/emissions/gridded_netcdf/tarfiles/

[ On some browsers clicking on this link may not work. If it doesn't, try to reach the site from a terminal window:    ftp ftp-ipcc.fz-juelich.de     (user: anonymous,   passwd: {your_email} )]

 

There are no emissions data currently available for natural aerosols (e.g., volcanic, sea salts, dust).



5. Land-use data. 31 Aug 2009.


The Land-use Harmonization team has prepared a harmonized set of land-use scenarios that smoothly connects historical reconstructions of land-use, based on HYDE data and national wood harvest statistics, with future projections of land use based on the Integrated Assessment Model (IAM) implementations of the RCPs. The harmonization method assures better continuity over the transition between the end of historical reconstruction and IAM initial conditions, and it preserves the future changes depicted by the IAMs at the grid level. Fractional land-use patterns (e.g. crop, pasture, urban, primary, secondary) and underlying land-use transitions (e.g. agricultural expansion/contraction, wood harvest, and shifting cultivation) are provided annually for the time period 1500-2100 at 0.5° x 0.5° resolution. The reconstruction starting in year 1500 is recommended over one that may also be available beginning in year 1700.


The land use data is available at: http://tntcat.iiasa.ac.at/RcpDb/


If you are unable to access the above site, the same datasets are available at http://luh.unh.edu, where you will also find additional information describing how the datasets were derived. On this alternative site, the following files are the ones you'll need for CMIP5:
1) If your model includes transitions from/to "urban land" as a category, you should download:

LUHa_u2.v1
LUHa_u2.v1_image.v1.1 (for RCP2.6)
LUH_u2.v1_minicam.v1 (for RCP4.5)
LUHa_u2.v1_aim.v1.1 (for RCP6.0)
LUH_u2.v1_message.v1 (for RCP8.5)

2) If your model omits transitions from/to "urban land" as a category, you should download:

LUHa.v1
LUHa.v1_image.v1.1 (for RCP2.6)
LUH.v1_minicam.v1 (for RCP4.5)
LUHa.v1_aim.v1.1 (for RCP6.0)
LUH.v1_message.v1 (for RCP8.5)

Note that the data files listed above are based on reconstructions starting in 1500, and the alternative reconstructions starting in 1700 (i.e., files with "t1" in the filenames) should not be used for CMIP5 boundary conditions.


6. The AC&C/SPARC ozone database. (Extended data 18 Nov 2010)


Nov 2010: The additioinal ozone RCP 6.0 scenario is now available!

   Click here to see details about the new dataset

The standard AC&C/SPARC ozone database is provided on pressure levels between 1000 and 1 hPa. For use in high-top models, extrapolation of the data to ~0.01 hPa or higher may be required.

The UK National Centre for Atmospheric Science (NCAS) thus has produced an updated/extended version of the SPARC ozone dataset suitable for high-top models, as follows:

(1) A multiple-linear regression of the historical raw pressure-level data between 1000-1 hPa, consistent with the Randal and Wu method used to construct the timeseries, was performed. The time series of ozone was represented as:

O3(t) = a*SOL + b*EESC + seasonal_cycle + residuals

For consistency, the 11-yr solar cycle (SOL) and EESC indices were identical to those used to prepare the original dataset (where the 180.5nm time series of the SOL index was provided by Fei Wu of NCAR.)

(2) The seasonal cycle and residual fields then were extrapolated and reduced above 1 hPa:

O3(z)=O3(1hPa)*exp(-(z-z(1hPa))/H, where H = 7 km

The seasonal cycle field provided a smooth ozone climatology up to the model top. This was further smoothed by weighting by the cosine of latitude.

(3) To minimize the 11-yr solar cycle and trend components in the mesosphere, the a and b regression coefficients were extrapolated and reduced twice as much as the seasonal cycle above 1 hPa, i.e. applying the same method as in (2) above, but with H set to 3.5 km.

(4) The full ozone time series was then reconstructed as in (1) above, but the ozone climatologies were unaffected.

(5) The standard SPARC ozone dataset which extends into the future does not include post-2009 solar cycle variability. For production of future data that include an 11-yr ozone solar cycle, the solar regression index was used to build a time series consistent with a repeating solar irradiance compiled by the UK Met Office. It was modelled as a sinusoid with a period of 11 years, and with mean and max-min values corresponding to solar cycle 23

(Veronika Eyring, Gareth Jones,personal communication).


The AC&C/SPARC ozone database covers the period 1850 to 2100 and can be used as forcing in climate models that do not include interactive chemistry. The historical part of the AC&C/SPARC ozone database spans the period 1850 to 2009 and consists of separate stratospheric and tropospheric data sources: (1) A multiple linear regression analysis of SAGE I+II satellite observations and polar ozonesonde measurements is used for the stratospheric dataset during the well-observed period from 1979 to 2005. The regression includes terms representing equivalent effective stratospheric chlorine (EESC) and the 11-year solar cycle variability. The EESC regression coefficients are used to extrapolate that data back in time, and form a stratospheric ozone time series backward to cover the entire historical time period 1850-2006. (2) Tropospheric data are derived from the chemistry-climate models Community Atmosphere Model (CAM) version 3.5 and the NASA-GISS PUCCINI model. Both models simulate tropospheric and stratospheric chemistry with feedback to the radiation and were driven by the recently available historical (1850-2000) emissions succinctly described in Lamarque et al. (ACP, 2010).

 

The future part of the AC&C/SPARC ozone database covers the period 2010 to 2100 and seamlessly extends the historical database. The future ozone timeseries also combines separate stratospheric and tropospheric data sources: (1) The stratospheric ozone projections are taken from the future reference simulations (REF-B2) of the 13 CCMs that performed a future simulation until 2100 under the SRES A1b GHG scenario and the A1 adjusted halogen scenario in CCMVal-2 (Chapter 9 of the SPARC CCMVal Report available at http://www.atmosp.physics.utoronto.ca/SPARC). In the stratosphere, the multi-model mean of the REF-B2 simulations is used in all RCP scenarios. (2) The future tropospheric ozone time series continues the historical CAM3.5 simulation until 2100 using the four different Representative Concentration Pathways (RCPs). The stratospheric and tropospheric data are combined by simply merging the two data sets across the climatological tropopause, to produce a smooth final data set.

Citation: Cionni, I., V. Eyring, J. F. Lamarque, W. J. Randel, D. S. Stevenson, F. Wu, G. E. Bodeker, T. G. Shepherd, D. T. Shindell, and D. W. Waugh, Ozone database in support of CMIP5 simulations: results and corresponding radiative forcing, Atmos. Chem. Phys. Discuss., 11, 10875-10933, doi:10.5194/acpd-11-10875-2011, 2011.

Interactive Public Discussion on the paper until 02 Jun 2011: http://www.atmos-chem-phys-discuss.net/11/10875/2011/

 

Contact for questions: Irene Cionni, Veronika Eyring, Jean-Francois Lamarque, Bill Randel

 

A short documentation is included in the header of the files and some example plots are available on the CCMVal website (http://www.pa.op.dlr.de/CCMVal/).

 

The files are in CF compliant netCDF format and are split up into 16 files for the historical part and 9 files for each RCP scenario; each file includes one decade and has a size of 30 Mbytes.


List of filenames

    Click to open the panel and see file names.
   Historic Data ( released 24 Sep 2009 )

Notice files name change (24 Sep 2009).
Ozone_CMIP5_ACC_SPARC_1850-1859_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1860-1869_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1870-1879_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1880-1889_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1890-1899_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1900-1909_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1910-1919_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1920-1929_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1930-1939_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1940-1949_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1950-1959_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1960-1969_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1970-1979_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1980-1989_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_1990-1999_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2000-2009_T3M_O3.nc

   The RCP2.6, RCP4.5, RCP8.5 Data ( released 25 Mar 2010 )

Ozone_CMIP5_ACC_SPARC_2010-2019_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2020-2029_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2030-2039_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2040-2049_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2050-2059_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2060-2069_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2070-2079_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2080-2089_RCP2.6_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2090-2099_RCP2.6_T3M_O3.nc

Ozone_CMIP5_ACC_SPARC_2010-2019_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2020-2029_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2030-2039_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2040-2049_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2050-2059_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2060-2069_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2070-2079_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2080-2089_RCP4.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2090-2099_RCP4.5_T3M_O3.nc

Ozone_CMIP5_ACC_SPARC_2010-2019_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2020-2029_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2030-2039_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2040-2049_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2050-2059_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2060-2069_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2070-2079_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2080-2089_RCP8.5_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2090-2099_RCP8.5_T3M_O3.nc

   The RCP6.0 Data ( released 18 Nov 2010 )

Ozone_CMIP5_ACC_SPARC_2010-2019_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2040-2049_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2070-2079_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2020-2029_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2050-2059_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2080-2089_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2030-2039_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2060-2069_RCP6.0_T3M_O3.nc
Ozone_CMIP5_ACC_SPARC_2090-2099_RCP6.0_T3M_O3.nc


Downloading data

The ozone data are available on the old (CMIP3) ESG ftp server. Please do note, that this is an old CMIP3 ftp site and you need to use the CMIP3 ESG account. To access them:


(1) Register at the CMIP3 ESG site:
https://esg.llnl.gov:8443/security/accountRequestData.do

Existing ESG users can use their current username / password.


(2) ftp to the ESG ftp server:

% ftp ftp-esg.ucllnl.org


(3) The ozone data files are in cmip5/ozone:

ftp> cd cmip5/ozone



7. The AMIP sea surface temperature and sea ice datasets.


Two distinctly different data sets are available. The first is the observed SST and sea ice monthly mean data and it has been prepared and updated by NCAR following a procedure described by Hurrell et al. (2007) in a draft journal article.

 

The second data set, known as the AMIP SST and sea ice boundary conditions, should be prescribed in all AMIP simulations. This data set is derived from the observational data set and if the mid-month values in this dataset are interpolated linearly in time to obtain daily values, then the user is certain that a subsequent monthly mean of these daily values will be the same as the observed. Note that monthly averages are not preserved when daily values are obtained by linearly intepolated between monthly mean observations. When the boundary condition data (as opposed to the observed monthly means) are used for an AGCM simulation, it has been shown that in addition to preserving the observational monthly means, the seasonal cycle and the interannual variance are more accurately represented.

 

The data and documentation are available at (PLEASE USE THE SECOND DATASET FOR YOUR SIMULATIONS):

 

1) Observed AMIP SST and Sea ice: documentation and data
2) AMIP SST and Sea ice Boundary Conditions: documentation and data

 



8. Specifications for CFMIP-inspired experiments.


The Cloud Feedback Model Intercomparison Project (CFMIP) experiments are described in Section IV of the CFMIP-2 experimental design and appear in Table 6 of the CMIP5 experimental design.

 

A patterned SST forcing dataset is required for Expt. 6.6. (Thanks to Brian Soden for producing this).

 

For information on setting up the CFMIP aquaplanet experiments, please refer to Brian Medeiros's CFMIP-2 Aquaplanet information.

 

The finalised list and plot of locations for high frequency CFMIP outputs is also now available.

 

CFMIP-2: CMIP5 web page.

COSP Version 1.1 (CFMIP observational simulator package).