The following files were replaced in the seaesf dataset. The replacements were needed to correct data and to extend these datasets to allow models to run to the end of 2003. Most originated at /home/ds/fms/forcings_input/to_2004 AEROSOLS : The file /archive/lwh/fms/INPUT.IPCC/slow_wetdep_bfix/aerosol.climatology.1859_2010_A1B.nc contains aerosol data (kg/m2) for the period 1859-2010. It is constructed as given below: The netCDF time variable is days since 1-1-1800, 0000Z. Data is (always) assumed to be on the 15th of the month. The calendar is "julian". The AM/CM models contain a module (in interpolator) to set this date for the actual mid-month, using a "NOLEAP" calendar, if such is desired. The year 1859 has the same data as the year 1860. This year is included for ease of interpolation in a model. Aerosol concentrations are obtained from calculations using the MOZART CTM. Calculations are made at 10-year intervals. Historical emissions are based on the EDGAR-HYDE emissions inventory. Results for 2000 and 2010 use the A1B scenario. These calculations include the "slow" wet deposition value (0.2*HNO3) for the aerosol species (sulfate, hydrophilic BC, hydrophilic OC, ammonium, nitrate). The calculations also include the black carbon biomass burning bugfix for 1860-1880. (NB: The ozone and aerosol files used in the original CM2.0/2.1 runs did not include this final bugfix.) aerosol.climatology.1859_2010_A1B.nc has been renamed aerosol.climatology.nc aerosol.1860.nc and aerosol.AM2.nc have been removed. WELL MIXED GREENHOUSE GASES : CH4 DATA : All CH4 datasets use values from Etheridge et al (as reported in IPCC WGI TAR and also in Pitts and Pitts) through 1992 (see below) Global mixing ratios calculated over the last 1000 years from Antarctic and Greenland measurements. Smoothing spline fit with 75 year cut off (Table 2, part 1, Etheridge et al., 1998) Also appears in IPCC WGI TAR Chapter 4 Figure 1 and Summary for Policymakers Figures 2 and 11, and as Figure 14-17 IN Chemistry of the Upper and Lower Atmosphere by Barbara J Finlayson-Pitts and James N. Pitts Jr. Academic Press, 2000. ch4_gblannualdata_hist_to2004 uses CMDL data for 1993.5 through 2004.5. The CMDL data from ~1986.5 through 1998.5 are actually quite close to the IPCC values (they were the primary datasource for the IPCC) but differ significantly at 2000.5 and 2001.5 from the scenario values (let alone in later years when scenario has a continuing increase of 11.1 ppbv). NOTE: actually the 1993.5 - 2004.5 data were developed using the new "NOAA04" scaling (see below). the data used here employ the scaling factor (dividing by 1.0124) required to return to the previous scaling. The datasets described above employs the previous scaling of lab CH4 measurements. Dlugokencky has developed a new "NOAA04" scaling for experimental data which has the effect of increasing ALL measured data (from 1000 onward) by ~1.2 percent (~20 ppbv). The following is from Ed Dlugokencky (CMDL): (questions of Dan Schwarzkopf, edited for clarity) Does this imply that the historical dataset (from year 1000) is also biased by ~1.2 percent? or is there a year before which the rescaling has no effect? The 1000-year record of CH4 global averages in the IPCC 3rd assessment report was from Etheridge et al. (1998) at CSIRO. At that time, they (CSIRO GASLAB) were using the same CH4 standard scale as we were. To convert their CH4 mole fractions to our NOAA04 scale, multiply by 1.0124 (1.24%). The error will be about 20 ppb at the end of their record and roughly half that at the beginning of their 1000-year record. Is the rescaling an experimental correction or a change in computing global averages from the individual stations? Methane, and most long-lived trace gases, are measured with relative analyical techniques. In the case of CH4, it is a gas chromatograph with flame ionization detection. The response of the analytical instrument has to be calibrated against a standard. We have developed a new standard scale (essentially cylinders of natural air with well-known CH4 abundances in them) that gives CH4 molar mixing ratios that are a factor of 1.0124 greater than the previous scale. So, the short answer is that this was an experimental error. The absolute uncertainty in the new scale is +/-0.2% (about 3 ppb). Ed ch4_gblannualdata_hist_to2004 has been renamed ch4_gblannualdata CO2 DATA The co2 data is in co2_gblannualdata_to2004 and co2_gblannualdata_hist_to2004. The data in co2_gblannualdata_to2004 are from the c4mip source: http://www.c4mip.cnrs-gif.fr/protocol.html. Data after 1960 is a spline fit of global data which is taken to be: 0.25*(S Pole) + 0.75*(Mauna Loa). The c4mip data ends at 1999 (1999.5). Data after 1999.5 are from Mauna Loa and S. Pole only. The global-mean is obtained from the above formula. Mauna Loa and S Pole data exists through 2004. (12/2004). at http://cdiac.esd.ornl.gov/by_new/bysubjec.html#atmospheric (CDIAC) . See Atmospheric Carbon Dioxide and Carbon Isotopes In this file the data for 2000.5 and 2001.5 are the scenario (A1B) values. this is to permit extension of current historical runs. The data in co2_gblannualdata_hist_to2004 are the same as in co2_gblannualdata_to2004 except that the 2000.5 and 2001.5 CDIAC (actual) data are used. This implied that runs using this dataset MUST begin no later than 1999.0. co2_gblannualdata_hist_to2004 has been renamed co2_gblannualdata HALOCARBON DATA The files f11(f12,f113,f22)_gblannualdata_to2004 are obtained using the data and procedures outlined below. The pre-1975 values for f11,f12 and f113 are obtained using a formula obtained by Prinn(2000) (rounded to the nearest pptv). It is assumed that pre-1950 values are zero. Also, the f11 value for 1950 is set to zero (formula breaks down). The 1978 - 1998 values for f11,f12,f113 are taken from IPCC reports and Ozone trends reports. the 2000.5 and 2001.5 are from SRES scenario (2001.5 being interpolated) this allows extension of existing runs beyond 2000. The 1986 halocarbon values are from table 8.1, ozone trends panel report (1988). The 2002 - 2004 values for f11, f12 are from Agage. as in the n2o case, as of 7 aug 2005 these values exist through 9/2004. so the apr-sep average is used, with no correction. global-mean values are an average of cape grim and mace head. For f113: Agage provided 2002 values. previously there were 2003 values but these were "retracted" due to calibration errors. There are no 2004 values. I have decided to take 2003.5 = (2002.5) - 0.6 and 2004.5 = (2002.5) - 1.2 as this trend is consistent with measured trends for previous years (and with the retracted 2003 value). It is a larger trend than scenario (which is -0.4 pptv/yr). The f22 values before 1978 are Dan Schwarzkopf's extrapolations. After 1978 they are from IPCC/Ozone trends reports. The 1992 value is 100 not 105 because the 1998 ipcc report gives that value and it seems to fit the timeline better. The 2000 value is the is92 scenario value (as with the other halocarbons). The f22 values for 2002-2004 are GCMS measuremaents (part of the Agage program) - as with f11,f12,f113 the 2004 values exist through 9/2004 and the Apr-Sep average is used for that year. The files f11(f12,f113,f22)_gblannualdata_hist_to2004 are the same as f11(f12,f113,f22)_gblannualdata_to2004 except for the following: f11 - Agage data are used for 1995.5 through 2004.5, with the proviso about the 2004 data as above. The 1994.5 data point is removed. This means runs with the new data MUST start no later than 1992.0 to retain compatibility with the old dataset. f12 - Agage data are used for 1995.5 through 2004.5, with the proviso about the 2004 data as above. This means runs with the new data MUST start no later than 1994.0 to retain compatibility with the old dataset. f22 - GCMS data (part of Agage) are used for 1999.5 through 2004.5, with the proviso about the 2004 data as above. This means runs with the new data MUST start no later than 1998.0 to retain compatibility with the old dataset. f113 - Agage data are used for 1995.5 through 2004.5, with the proviso about the 2003 and 2004 data as in the previous notes for f113. This means runs with the new data MUST start no later than 1994.0 to retain compatibility with the old dataset. f113_gblannualdata_hist_to2004 has been renamed f113_gblannualdata f11_gblannualdata_hist_to2004 has been renamed f11_gblannualdata f12_gblannualdata_hist_to2004 has been renamed f12_gblannualdata f22_gblannualdata_hist_to2004 has been renamed f22_gblannualdata N2O DATA : The file n2o_gblannualdata_to2004 uses IPCC data through 1998.5 and uses the A1B scenario data through 2001.5. This ensures that current runs through 2000 can use these data to extend runs. Data for 2002.5 and 2003.5 are global-annual-mean data from the Agage network. The global mean is the average of the Mace Head and Cape Grim values. As of 7 aug 2005, data for 2004 were available only through 9/2004. So the 2004.5 value is obtained as an average of the April through September values. In earlier years, this appears to underestimate the annual average by ~0.1 ppbv, but no correction was applied. The file n2o_gblannualdata_hist_to2004 is the same as the file n2o_gblannualdata_to2004, but Agage data is used for 2000.5 and 2001.5. This file is thus independent of any scenario values. n2o_gblannualdata_hist_to2004 has been renamed n2o_gblannualdata SOLAR SPECTRAL DATA : Solar spectral data are supplied by Judith Lean. The data have been massaged to conform to the spectral intervals used in the shortwave radiative transfer package. It is assumed ( for lack of data) that the 1860-1881 spectral (and total) irradiance is constant. The values used in that period are the 1882 average. These are represented in the dataset by the line (1882 0 ...). Data then exists through 2004/12. The final line (2004 0 ...) implies a constant spectral and total irradiance for 2005 and subsequent years. The value is the average of 2004. lean_solar_spectral_data_1882-2004.b18 has been renamed lean_solar_spectral_data.dat VOLCANIC AEROSOL PROPERTIES : The files /archive/ds/fms/volcano_input_netcdf/asmsw_1850_2004.nc extlw_1850_2004.nc extsw_1850_2004.nc omgsw_1850_2004.nc contain asymmetry factor, lw ext coefficient, sw ext coefficient and sw single-scattering albedo for stratospheric (volcanic) aerosols for the period 1850-2004. The historical data (1850-1999) have been supplied by G. Stenchikov using, in part, results from Sato. The values for 2000-2004 are equal to the December 1999 values, following a suggestion from Stenchikov. NOTE: the 2000-2004 values differ from those used in CM/AM models in which the 2000 et seq. values assumed zero volcanic influence (ie, extlw/sw = 0). They also differ from values used by J. Austin for Ozone assessment calculations (which use the 1997-99 average for 2000 et seq.) asmsw_1850_2004.nc has been renamed asmsw_data.nc extlw_1850_2004.nc has been renamed extlw_data.nc extsw_1850_2004.nc has been renamed extsw_data.nc omgsw_1850_2004.nc has been renamed omgsw_data.nc OZONE : The file /archive/ds/fms/ozone_input_netcdf/slow_wetdep_bfix_randel_mar05/o3.mon_1859_2004.nc contains tropospheric and stratospheric ozone data (mmr) for the period 1859-2004. It is constructed as given below: The netCDF time variable is days since 1-1-1800, 0000Z. Data is (always) assumed to be on the 15th of the month. The calendar is "julian". The AM/CM models contain a module (in interpolator) to set this date for the actual mid-month, using a "NOLEAP" calendar, if such is desired. The year 1859 has the same data as the year 1860. This year is included for ease of interpolation in a model. This ozone climatology is a blend of four datasets 1) MOZART CTM (surface to linear tropopause) 2) Stratospheric ozone < 50km (Randel, March 2005) 3) 51-56km (Fortuin-Kelder 1990) 4) > 60km (SKYHI climatology) A linear transition is used in the tropopause and 56-60km regions. Tropospheric ozone is obtained from calculations using the MOZART CTM. Calculations are made at 10-year intervals for the period 1860 to 2010. Historical emissions are based on the EDGAR-HYDE emissions inventory. Results for 2000 and 2010 use the A1B scenario. These calculations include the "slow" wet deposition value (0.2*HNO3) for the aerosol species (sulfate, hydrophilic BC, hydrophilic OC, ammonium, nitrate). The calculations also include the black carbon biomass burning bugfix for 1860-1880. (NB: The ozone and aerosol files used in the original CM2.0/2.1 runs did not include this final bugfix.) Linear interpolation between decadal values is made for the years 1979, 1981-9, 1991-9 and 2001-4. This is because stratospheric ozone values exist separately for all years 1979-2004 and the tropospheric and stratospheric values need to be combined to produce a dataset for each year. The tropospheric ozone values are used up to the linear tropopause. The linear tropopause is defined here as the 300 hPa level at the equator , 100 hPa at the poles, with a linear (in latitude) variation between equator and poles. Then a transition to stratospheric ozone values is made in a layer situated at the trop-strat boundary (essentially the pressure-weighted average of the ozone values from the strat and trop data). The stratospheric ozone values below 50 km are data from Randel as obtained in Mar 2005. Data are for the 1979-2004 period. In the 51-56 km range, a time-invariant ozone value is used, obtained from climatological ozone data (Fortuin-Kelder) for 1990. Above 56 km there is a linear transition to older climatological values previously used in SKYHI calculations. SKYHI values are used exclusively above 60 km. IMPORTANT ASSUMPTION: Stratospheric ozone for 1860-1978 is the same as the 1979 value. In contrast, the dataset used for previous AM and CM calculations used values for 1860-1975 from an older dataset. This produced a "jump" between 1975 and 1979 which compromised the interpretation of the effects of strat ozone change in the 1970s. o3.mon_1859_2004.nc has been renamed o3.climatology.nc o3.1860_1990.nc and o3.AM2.nc have been removed.