This section describes the major projects completed between GMI’s inception in the 1990s and about 2007. These simulations are no more than a few years long. They are ‘pre-Hindcast’ because there was no multi-decadal reanalysis product like MERRA. As such they have limited application and are frankly pretty useless compared to more recent model versions, not only because newer meteorological products are much better but numerous mistakes in the older chemical codes have since been corrected (e.g., the FastJX ‘half-a-level off’ error, SYNOZ-produced O3tracer actually being added into the real ozone field). The order is reverse chronological. Some of this output may be obtained by anonymous ftp to Index of /datashare/dirac/gmidata2/users/mrdamon (nasa.gov) or Index of /datashare/dirac/gmidata2/users/steenrod (nasa.gov). If there’s a simulation listed here that you want but cannot find it under either of these directories, send inquiries to megan.damon@ssaihq.com or Stephen.d.steenrod@nasa.gov.
The output directories listed also contain ‘README’ and namelist files that give more detailed information about the simulation.
The first finite volume assimilation was the FVDAS. After that came GEOS-4-DAS, GEOS-5-DAS, then MERRA.
Combo Model - Aura Period (G5Aura)
Met fields: GEOS-5.1-Data Assimilation System (pre-MERRA) 2° lat x 2.5° lon x 72 levels (lid at 0.01 hPa) Time Period: Jan. 2005-Dec. 2007
These simulations are in deep archive. Please contact megan.damon@ssaihq.com or Stephen.d.steenrod@nasa.govfor assistance. (Old dirac directory was /pub/gmidata2/output/gmic/G5Aura)
Similar to Aura4 except with GEOS5.1 meteorological fields and higher vertical resolution (72 instead of 42 levels).
Combo Model - Aura Period (Aura4)
Met fields: GEOS-4-Data Assimilation System, 2° lat x 2.5° lon x 42 levels (lid at 0.01 hPa) Time Period: Feb. 2004-Dec. 2006 Aura4 Directory: Index of /datashare/dirac/gmidata2/users/mrdamon/Aura4-Final (nasa.gov)
There are 4 minor changes to the model compared to the version used for Aura3: 1) H2 mixing ratio is set to 500 ppb everywhere, 2) a minor coding error in the 'condense' subroutine was fixed, 3) the units for O3 produced from shipping NOx emissions were corrected, and 4) an updated Pickering and Allen lightning parameterization was used. This run differs from Aura4 primarily by the emissions used. Some of the new emissions include hourly NOx fossil fuel variation (Harvard), seasonality of Chinese emissions (Streets), and a fix in the GFEDv2 boreal emissions for 2004-5.
Combo Model – AURA Period (Aura3)
Met fields: GEOS-4-Data Assimilation System, 2° lat x 2.5° lon x 42 levels (lid at 0.01 hPa) Time Period: Feb. 2004-Apr. 2007 These simulations are in deep archive. Please contact megan.damon@ssaihq.com or Stephen.d.steenrod@nasa.govfor assistance.
This run was previously called ‘ap1.0HO2Aura2’. The met fields used are a 3-hr time averaged DAS product. This simulation has the new Pickering and Allen lightning parameterization with its original NOx profile. It includes the HO2 heterogeneous uptake reaction that was turned off in Aura2. Tropospheric aerosol inputs used are from 2004-2006 GOCART simulations. JPL06 reactions rates are used, however, photolysis cross sections were not updated. The previous run of the Aura period, ‘Aura2’, differs only in that it used the original GMI lightning parameterization and the HO2 uptake reaction was turned off. Schoeberl et al. [2007] compare tropospheric column O3 (TCO) from the aura2 run with an Aura OMI/MLS-derived TCO product.
Combo Model – AURA Period - Forecasts
Met fields: Forecasts from GEOS-4-DAS first-look analyses, 2° lat x 2.5° lon x 42 levels (lid at 0.01 hPa) Time Period: July 1, 2004-June 30, 2005These simulations are in deep archive. Please contact megan.damon@ssaihq.comor Stephen.d.steenrod@nasa.gov for assistance.
These simulations use the Aura2 version of the code (see above) with GEOS-4-DAS instantaneous forecast met fields (6-hr updates) that were generated from 6-hr time-averaged DAS analyses. ‘for12h’ used 18-30 hr forecasts, updating the forecast sequence every 12 hrs. ‘for24h’ used 12-36 hr forecasts, updating the forecast sequence every 24 hrs.
Combo – Model GCM Met fields 1994-1998 (FvgcmCombo)
Met fields: GEOS-4-GCM, 2° lat x 2.5° lon x 42 levels (lid at 0.01 hPa) Time Period: Jan. 1994-Dec. 1998 (years based on SSTs used in the GCM integration). These simulations are in deep archive. Please contact megan.damon@ssaihq.com or Stephen.d.steenrod@nasa.govfor assistance.
This simulation uses the same model build as ’Aura3’, which includes the new lightning parameterization and has HO2 uptake turned on. Met fields have 3-hr updates. Inputs include Harvard 1980-1990 biomass burning sources, ship emissions, fossil fuel and biofuel from Harvard (1995), and source gas boundary conditions for 1994-8. Publications that used this or a prior Combo-fvgcm integration include Schoeberl et al. [2006], Duncan et al. [2007], and Strahan et al. [2007].
Tropospheric Model – Hemispheric Transport of Air Pollution (HTAP) experiments
Met fields: GEOS-4-DAS, 2° lat x 2.5° lon x 42 levels (lid at 0.01 hPa) Time Period: Jan. 1, 2001-Dec. 31, 2001. These simulations are in deep archive. Please contact megan.damon@ssaihq.com or Stephen.d.steenrod@nasa.govfor assistance.
Eighteen simulations using the tropospheric model were integrated for the HTAP study of source-receptor relationships for ozone and its precursors (NOx, NMHC, CH4, and CO). The results were submitted to HTAP. This work was supervised by Bryan Duncan.
Aerosol Model – Hemispheric Transport of Air Pollution (HTAP) experiments
Met fields: GEOS-4-DAS, 2° lat x 2.5° lon x 42 levels (lid at 0.01 hPa) Time Period: Jan. 1, 2001-Dec. 31, 2001. These simulations are in deep archive. Please contact Stephen.d.steenrod@nasa.gov for assistance.
A modified version of the GMI aerosol model was used to integrate source-receptor experiments for 4 regions: N. America, Europe, E. Asia, and S. Asia. Both baseline (SR1) and perturbation runs (SR6) were submitted to HTAP (Hemispheric Transport of Atmospheric Pollutants). These results are used in the HTAP 2007 Interim Report, submitted in Dec. 2007. This work was supervised by Huisheng Bian.
Aerosol Model – Aerosol Sensitivity to Meteorology
Met fields: GEOS-1-Data Assimilation System (aka ‘DAO’), GEOS-4-GCM (aka FVGCM), and NASA GISS II’ GCM GEOS-1-DAS Resolution: 4° lat x 5° lon x 46 levels (lid at ~0.4 hPa) GEOS-4-GCM Resolution: 4° lat x 5° lon x 42 levels (lid at ~0.01 hPa) GISS II’ Resolution: 4° lat x 5° lon x 23 levels (lid at ~0.01 hPa) Mechanism: Univ. of Michigan aerosol module Time Period: DAS met fields Mar. 1997-Feb. 1998, other fields are from GCMs. Source gases for 1995. These simulations are in deep archive. Please contact Stephen.d.steenrod@nasa.gov for assistance.
Current global aerosol models use different physical and chemical schemes and parameters, different meteorological fields, and often different emission sources. Since the physical and chemical parameterization schemes are often tuned to obtain results that are consistent with observations, it is difficult to assess the true uncertainty due to meteorology alone. Three meteorological data sets are used to drive the same aerosol model. The differences and uncertainties in aerosol simulations (for sulfate, organic carbon, black carbon, dust, and sea salt) solely due to different meteorological fields are analyzed and quantified. See Liu et al. [2007].
Aerosol Model – Preindustrial Aerosol Distributions
These simulations are analogous to the above simulations, but use estimates of preindustrial levels of aerosol precursors in the aerosol calculation.
Tropospheric Model – IPCC AR4 model intercomparison study
Met fields: GEOS-1-Data Assimilation System (aka ‘DAO’), GEOS-4-GCM (aka FVGCM), and NASA GISS II’ GCM GEOS-1-DAS Resolution: 4° lat x 5° lon x 46 levels (lid at ~0.4 hPa) GEOS-4-GCM Resolution: 4° lat x 5° lon x 42 levels (lid at ~0.01 hPa) GISS II’ Resolution: 4° lat x 5° lon x 23 levels (lid at ~0.01 hPa) Time Period: DAS met fields Mar. 1997-Feb. 1998, other fields are from GCMs. Source gases for 1995. These simulations are in deep archive. Please contact Stephen.d.steenrod@nasa.gov for assistance.
These troposphere-only simulations (with synoz to represent stratospheric ozone) were run with 3 sets of met fields using 4 different present day (1995) and future (2030) emissions scenarios specified by the Accent AR4 photochemical intercomparison. Analyses were performed outside of GMI and references can be found in the IPCC-Accent publication list below.
Stratospheric Model – 'Hindcast,' 1975-2004
Met fields: Finite Volume GCM (FVGCM), 2° lat x 2.5° lon x 33 levels (lid at ~0.01 hPa) Time Period: Source gas boundary conditions from 1975-2004. These simulations are in deep archive. Please contact Stephen.d.steenrod@nasa.gov for assistance.
The credibility of a model used to assess future ozone trends can be evaluated by the model’s ability to ‘hindcast’ ozone over the period of existing measurements. We wish to bracket some of the range of possible model results by doing simulations representing a perpetually warm and a perpetually cold Arctic winter. Two years of FVGCM met fields were chosen based on their Arctic lower stratospheric temperatures. The hindcast simulations include variations due to chlorine and bromine sources, solar UV cycle, and volcanic aerosols.
Publications using results from the hindcast simulations are Waugh et al. [2007] and Douglass et al. [2006].
Stratospheric Model – Sensitivity of Antarctic ozone recovery to meteorology
Met fields: Finite Volume GCM (FVGCM) and Finite Volume DAS (FVDAS), 4° lat x 5° lon x 28 levels (lid at ~0.4 hPa) Time Period: FVDAS represents July 1, 1999-June 30, 2000. WMO source gas boundary conditions represent the period 1995-2030.
These simulations are in deep archive. Please contact megan.damon@ssaihq.com or Stephen.d.steenrod@nasa.govfor assistance. (Old Directory: /pub/gmidata/LLNL/GMI-S/FVCCM and ../FVDAS)
This experiment was designed to investigate the uncertainty due to meteorology in the prediction of ozone recovery. We chose meteorological input from a general circulation model and from a data assimilation system because they are know to have significant differences in their residual circulation. Source gas boundary conditions came from the WMO scenario ‘MA2’, which includes changing surface concentrations for 16 halocarbons, including CFCs, HCFCs, and halons. Methane and nitrous oxide boundary conditions increase by ~25% and ~11%, respectively, over the simulation period.
Age of air in the GCM was known to be much more realistic than in the DAS, however both meteorological fields were known to have some realistic transport characteristics. Simulations with these wind fields were also performed with the Goddard Chemistry and Transport Model and the results were graded with some of the original GMI objective grading criteria [Douglass et al., 1999]. Both simulations scored higher than any previous GMI simulations. The age of air differences indicated significant differences in residual circulations [Schoeberl et al., 2003].
Each GMI CTM simulation was integrated by repeating one year of meteorological input from the FVGCM and the FVDAS to simulate 1995-2030. Both met fields had a colder than average Arctic winter. Analyses of chemistry, transport, and polar processes in these simulations are reported in Douglass et al. [2004], Considine et al. [2004], and Strahan and Douglass [2004]. We found that ozone trends in the Antarctic were dominated by chlorine changes rather than transport differences (chemistry rules, transport drools).
Stratospheric Model – Sensitivity of Lower Stratospheric Chemistry and Transport to Meteorology
Met fields: GEOS-1-Data Assimilation System (aka ‘DAO’), NCAR Middle Atmosphere Community Climate Model Version 2 (MACCM2), and NASA GISS II’ GCM GEOS-1-DAS Resolution: 4° lat x 5° lon x 46 levels (lid at ~0.4 hPa) MACCM2 Resolution: 4° lat x 5° lon x 52 levels (lid at ~0.01 hPa) GISS II’ Resolution: 4° lat x 5° lon x 23 levels (lid at ~0.01 hPa) Time Period: DAS met fields Mar. 1997-Feb. 1998, other fields are from GCMs. Source gases for 1995. These simulations are in deep archive. Please contact Stephen.d.steenrod@nasa.gov for assistance. (Old Directory: /pub/gmidata/LLNL/GMI_tracer_runs (tracer experiments only))
This is the ‘original’ GMI study. Three sets of 1-yr met fields (GEOS-1-DAS, NCAR MACCM2, and GISS II’) were used to simulate a perturbed and a baseline stratosphere. Temperature and transport diagnostics were developed and used to evaluate the credibility of the stratosphere in CTM simulations with these met fields [Douglass et al., 1999]. The MACCM2 winds were identified as being the most realistic in the northern hemisphere lower stratosphere. Kinnison et al. [2001] used GMI-MACCM2 simulation to study the effects of exhaust injected into the lower stratosphere from a supersonic aircraft fleet on stratospheric ozone.
