FAQs
CMIP5 – ESM2M and ESM2G
- Where can I find the output?
- Should I expect there to be data gaps?
- Is snow cover fraction (“SNC”) data available?
- Why are there small negative values for monthly surface snow amount (“SNW”)?
- Why are there strange bounded characteristics in the time series of sea level pressure?
- Is the GFDL-ESM2M and GFDL-ESM2G model code publicly available?
- Are all the coordinate variables available for the 3-dimensional aerosols?
- The ocean variable “omlmax” is no longer found in the GFDL CMIP5
database ocean “day” table. Has this data been withdrawn from the
database or only temporarily not available? - How is evapotranspiration (ET) modeled in the GFDL ESMs?
- Intel Array Viewer returns a message “not a NetCDF file” when opening GFDL data.
- What is the difference between the two +1%/yr CO2 simulations?
EXPERIMENT DESIGN
- What was the experimental protocol?
- What is the integration time length of piControl (fixed everything at the 1860 level)?
- How was climate equilibrium determined for the piControl integration?
- How was branch timing for the historical runs handled?
- What historical misc runs have you performed?
- What is the difference between “Concentrations” and “Emissions” runs?
- How was land use treated?
- How are the historical variations in stratospheric volcanic aerosols extended after 2005 in the RCP simulations?
- Were water hosing experiments performed by GFDL as part of CMIP5?
MODEL INFORMATION
DATA/METADATA
Data from all participating modeling centers can be found through the official CMIP5 website above. GFDL data can also be found via GFDL?s local server – http://nomads.gfdl.noaa.gov:8080/DataPortal/cmip5.jsp.
At this time (April 2012), there should not be any gaps in the ESM2M and ESM2G time series on our servers. It is possible that there are still data gaps on the ESG servers. Please try to access our data via the GFDL server – http://nomads.gfdl.noaa.gov:8080/DataPortal/cmip5.jsp.
We decided not to attempt to define and report a snow cover fraction partially because of ambiguity in its definition. For turbulent exchange with the atmosphere, we treat any snow on ground as 100% coverage of ground, regardless of depth. For radiative exchange, we use a depth-dependence, which might differ across the GFDL model streams. There is also snow on the vegetation, which could be called snow cover, too.
The negative SNW values are a common coding issue arising when the snow disappears during a model time step. Small negative values can exist and a floor (i.e., 0) has not been placed on the SNW values in the GFDL CMIP5 output.
The sea level pressure field was normalized to 1013.25 mb in ESM2M and ESM2G. A plot of the time series may show strange behavior due to the precision of the calculation of the global mean which was performed with 32-bit precision.
The GFDL-ESM2M code is publicly available and can be found by following the links at https://mom-ocean.github.io/docs/quick-start-guide/. We are currently working on making GFDL-ESM2G code public but have no time table.
It has been noted that there are some coordinate supporting variables (i.e., a, b, a_bnds) missing for the 3-dimensional aerosols. We are currently working on publishing these supporting variables and correcting the metadata describing the hybrid sigma pressure coordinate.
- The ocean variable “omlmax” is no longer found in the GFDL CMIP5 database ocean “day” table. Has this data been withdrawn from the database or only temporarily not available?
The “day” table variable, omlmax, was removed from our CMIP5 database because it was unintentionally included in the “day” table. The variable is the maximum mixed layer thickness over the month, and should not be included in the “day” table. The variable is available in the “Omon” table.
ET is modeled with consideration of simultaneous energy balance among bare soil, canopy air space, and foliage, as constrained by water availability. Transpiration is limited by stomatal conductance, which is determined jointly by consideration of photosynthesis and transpiration. The most important controls on ET in any model are precipitation and surface irradiance, regardless of the details of ET parameterization.
It has been reported that the Intel Array Viewer returns an error when using GFDL CMIP5 NetCDF data. Please use other NetCDF applications. We apologize for the inconvenience.
There are two +1%/yr increase in atmospheric CO2 mixing ratio experiments for which data is available:
- In the “r1i1p1” experiment, atmospheric CO2 levels were prescribed to increase from their initial mixing ratio level of 286.15 ppmv at a compounded rate of +1 percent per year until year 70 (the point of doubling, 2xCO2). CO2 levels were held constant at 574.2349 ppmv from year 71 through the end of the 200 year long experiment. Initial conditions for this experiment were taken from 1 January of year 1 of the 1860 control model experiment named ESM2M_pi-control_C1.
- In the “r1i1p2” experiment, atmospheric CO2 levels were prescribed to increase from their initial mixing ratio level of 286.15 ppmv at a compounded rate of +1 percent per year until year 140 (the point of quadrupling, 4xCO2). CO2 levels were held constant at 1144.6 ppmv from year 141 through the end of the 300 year long experiment. Initial conditions for this experiment were taken from 1 January of year 10 of the 1860 control model experiment named ESM2M_pi-control_C1.
In both of these experiments, all non-CO2 forcing agents (CH4, N2O, tropospheric and stratospheric O3, tropospheric sulfates, black and organic carbon, dust, sea salt, and solar irradiance) are held constant at values representative of year 1860, there is no land use change (i.e., “potential vegetation”), volcanic aerosols are zero, and the direct effect of tropospheric aerosols is calculated by the model, but not the indirect effects.
EXPERIMENT DESIGN
This set of experiments follows the historical and projected Representative Concentration Pathways (RCP) protocol of the Fifth Coupled Model Inter-comparison Project (CMIP5):
http://cmip-pcmdi.llnl.gov/cmip5/ as described in Taylor et al (2009; http://cmip-pcmdi.llnl.gov/cmip5/docs/Taylor_CMIP5_design.pdf)
piControl began in year 2401 and 1001 of spin-up integrations for GFDL-ESM2M and GFDL-ESM2G, respectively. Additional details of this spin-up integration for GFDL-ESM2M and GFDL-ESM2G can be found in the comment section of the netCDF metadata.
GFDL-ESM2M: “For the first 248 years of the spin-up, the physical climate was allowed to achieve a quasi-radiative equilibrium with the constant 1860 radiative forcing using potential, static vegetation and a fixed pCO2 value of 286 ppm, while a sponge was applied to the actual atmospheric CO2 values back to 286 ppm to avoid drift. Physical climate equilibrium was established using four coarse criteria: TOA net radiative flux less than ± 0.1 W/m-2, global 2-m air temperature drift less than 0.1C per century, annual SST biases less than 10C and North Atlantic Deep Water (NADW) transport southward at 30N greater than 10 Sv. At year 249 of the spin-up, dynamic vegetation was activated and the model was integrated for an additional 200 years after which time the land photosynthesis routine was allowed to feel the model’s own atmospheric CO2. After another ~150 years, the soil carbon was set to a quasi-equilibrium state using an offline analytic procedure, and the model integration was continued until the land and ocean carbon stored reached equilibrium. A net carbon flux of each component of less than 20 PgC per century and leading to net atmospheric response of less than 10 ppm per century was the criteria used for determining equilibrium. Year 1 of the archived ESM2M_pi-control_C1 experiment data begins at the end of this adjustment period.”
GFDL-ESM2G: “For the first 248 years of this spin-up integration, the physical climate was allowed to achieve a quasi-radiative equilibrium with the constant 1860 radiative forcing using potential, static vegetation and a fixed pCO2 value of 286 ppm, while a sponge was applied to the actual atmospheric CO2 values, restoring them back to 286 ppm to avoid drift. The criteria used to determine equilibrium for the physical climate was a TOA net radiative flux of less than ± 0.1 W/m-2. At year 249 of the spin-up integration, dynamic vegetation was activated and the model was integrated for an additional 200 years after which time the land photosynthesis routine was allowed to feel the model’s own atmospheric CO2. The atmospheric pCO2 was still restored to 286 ppm with a restoring time scale of about 1 year. After another ~150 years, the soil carbon was set to a quasi-equilibrium state using an offline analytic procedure and the ocean. At year 601, the temperature, salinity, and biogeochemical tracers in the ocean were reset to observations (Levitus et al. 2005) and changes to the mixing and diffusion parameterizations were introduced in order to address a deep-ocean cold bias. At years 1101 and 1401, changes were incorporated into the land vegetation model that resulted in an equilibrium total global biomass on land of ~850 PgC. The model integration was continued until the land and ocean carbon stores reached equilibrium. A net carbon flux of each component of less than 20 PgC per century and leading to net atmospheric response of less than 10 ppm per century was the criteria used for determining carbon equilibrium. Year 1 of the archived ESM2G_pi-control_C2 experiment data begins at the end of this adjustment period.”
Physical climate equilibrium was established using four coarse criteria: TOA net radiative flux less than +/- 0.1 W/m-2, global 2-m air temperature drift less than 0.1C per century, annual SST biases less than 10C and North Atlantic Deep Water (NADW) transport southward at 30N greater than 10 Sv. Additionally, a net carbon flux of each component of less than 20 PgC per century and leading to a net atmospheric response of less than 10 ppm per century was the criteria used for determining equilibrium. These are broad scale expert judgments used in the Dunne et al. 2012 documentation based on the need to represent the ocean’s role in regional climate and heat transport, the precedent of past decisions to abandon free coupling in GFDL’s early models and apply flux adjustment in order to retain such features, and the goals set in the last round of CM2 development (along with the presence of ENSO) where flux adjustment was no longer applied (Delworth et al. 2006 reference therein) The 10C guideline is local. Globally, we set as a guideline for credibility a RMSE of less than 2C, and preferably less than 1.5C.
To get the terrestrial carbon storage terms correct for Jan 1 1861, we run a 1700 to 1860 integration, starting from our 1860 pre-industrial control (i.e, piControl). In the 1700 to 1860 integration, the only forcing applied is the land-use transitions from 1700 to 1860; all other radiative forcings are held fixed at their 1860 values (as in the 1860 control integration). The historical integrations (concentration- and emissiondriven) start from the end of this integration and the forcings (solar, volcanoes, GHG, aerosols…and land-use transitions) are time evolving. The 1860 control integration is a so-called potential vegetation integration using no land-use transitions. For more details, see Sentman et al. 2011: Time scales of terrestrial carbon response related to land-use application: Implications for initializing an earth system model. Earth Interactions, 15(30), DOI:10.1175/2011EI401.1.
The 1700-1860 land use only integration will be made public at some future date (not before the end of 2012). This unpublished run started from Jan 1, year 1 of the 1860 piControl. Therefore Jan 1, year 1861 of the historical runs corresponds to year 162 in the piControl concentration-driven integration.
Model | RIP | Forcing * | Metafor Name |
---|---|---|---|
GFDL-ESM2M | r1i1p2 | Ant (Ant includes GHG,SD,Oz,LU,SS,BC,MD,OC; GHG includes CO2, CH4, N2O, CFC11, CFC12, HCFC22, CFC113) |
historicalMiscHCH |
GFDL-ESM2M | r1i1p3 | GHG,SD,Oz,Sl,Vl,SS,BC,MD,OC (GHG includes CO2, CH4, N2O, CFC11, CFC12, HCFC22, CFC113) |
historicalMiscHCP |
GFDL-ESM2M | r1i1p4 | GHG,SD,Oz,Sl,Vl,SS,BC,MD,OC (GHG includes anthropogenic CO2 emissions, CH4, N2O, CFC11, CFC12, HCFC22, CFC113) |
historicalMiscHEP |
GFDL-ESM2M | r1i1p5 | SD,SS,BC,MD,OC | historicalMiscHOA |
GFDL-ESM2M | r1i1p6 | LU | historicalMiscHOL |
GFDL-ESM2M | r1i1p7 | Sl | historicalMiscHOS |
GFDL-ESM2M | r1i1p8 | Vl | historicalMiscHOV |
GFDL-ESM2G | r1i1p3 | GHG,SD,Oz,Sl,Vl,SS,BC,MD,OC (GHG includes CO2, CH4, N2O, CFC11, CFC12, HCFC22, CFC113) – CO2 Concentration-driven |
historicalMiscHCP |
GFDL-ESM2G | r1i1p4 | GHG,SD,Oz,Sl,Vl,SS,BC,MD,OC (GHG includes CO2, CH4, N2O, CFC11, CFC12, HCFC22, CFC113) – CO2 Emission-driven |
historicalMiscHEP
|
- For a complete listing of these abbreviations, refer to the CMIP5 Data Reference Syntax (DRS) and Controlled Vocabularies (http://cmip-pcmdi.llnl.gov/cmip5/docs/cmip5_data_reference_syntax.pdf).
In ?Concentrations? runs, the atmospheric concentration of CO2 is restored to an externally specified value. The restoring time scale is 1 year. In ?Emissions? runs, the atmospheric concentration of CO2 is free to respond to the specified fossil fuel emissions as well as interactive land and ocean fluxes. All other atmospherically active factors (greenhouse gases, aerosols) are treated as externally specified in these runs.
In perturbation experiments, the conversion of primary forest and grassland to crops, pasture, and secondary forest, their management, and the conversion rates between them was done after Shevliakova, et al., (2009: Carbon cycling under 300 years of land use change: Importance of the secondary vegetation sink. Global Biogeochemical Cycles, 23, GB2022, DOI:10.1029/2007GB003176.)
- How are the historical variations in stratospheric volcanic aerosols extended after 2005 in the RCP simulations?
The stratospheric volcanic aerosols are all set to zero in the RCP simulations (as is the case for the 1860 control simulations). The stratospheric aerosol values for 2000-2005 in the historical runs are constant, set equal to those for December 1999.
GFDL has not performed water hosing runs with our CMIP5 models to date.
MODEL INFORMATION
The short answer is that the difference between ESM2M and ESM2G is the type of ocean model used. ESM2M uses a depth-based vertical coordinate (Modular Ocean Model Version 4.1; MOM4.1). ESM2G uses a sigma coordinate mixed layer and density-based vertical coordinate in the interior (Generalized Ocean Layer Dynamics; GOLD).
A more detailed description of the physical components and simulation characteristics can be found in Dunne et al. (in press: GFDL’s ESM2 global coupled climate-carbon Earth System Models Part I: Physical formulation and baseline simulation characteristics. Journal of Climate.
DOI:10.1175/JCLI-D-11-00560.1. 4/12.
Both of these models originate from GFDL?s CM2.1 climate model of Delworth et al. (2006: GFDL’s CM2 Global Coupled Climate Models. Part I: Formulation and Simulation Characteristics. Journal of Climate, 19(5), DOI:10.1175/JCLI3629.1.) submitted as part of the third Coupled Model Intercomparison Project (CMIP3). A prototype for these Earth System Models, called ESM2.1, added land and ocean carbon cycles to the CM2.1 physical climate model, and has been used in a variety of published science studies.