Possible problems and points to be clarified

• pland and precipitation formation profile are very old data (use data from the parent model, ECMWF instead).
• methane loss rate factors (OH in TM in stratosphere is not good enough). Further note that ch4rds contains methane destruction rates for top 8 levels of TM3 (level dependent, see ch4relds). Solved according to Michiel ?
• program uars_ratio seems to extrapolate beyond 76 N and S.
• The software used by Sander Houweling to compile methane emissions needs to be included in this package.
• The software used by Sander Houweling to compile VOC emissions needs to be included in this package. Does it use the Geia terpene and NVOC emissions from [Guenther and et al., 1995]?
• The software used by Bram Bregman / Frans Alkemade to calculate the monthly mean HNO3 : O3 ratios from UARS should be included in this package.
• The calculation of the geometry of TM should be put in a separate external subroutine to the programs to facilitate use of other grids than the present one. For instance we may wish ot run on the reduced gaussian grid of ECHAM in the future.
• The program to make file NH3_conc.d at the right TM resolution seems absent.
• Replace dry deposition parameters from ECHAM by ECMWF ones. It would need to be checked that the dry deposition speeds remain similar.
• The program to calculate the heights of TM3 layers, used in compvolc is missing Perhaps also the vertical distribution of methane emissions should be adapted (relative to the orography). I also do not (yet) undestand the total emission by volcanoes.
• Before 1970 there may be problem in sources_sinks of TM3, because NRTCO is not initialised but used anyway:
  x2=(e_1co(i,j)+hisco(i,j)+bf3co(i,j)+nrtco(i,j))*cofac*year2dt
• Where does coseas come from ?
• Of the SO$_2$ anthropogenic emissions 2.5 % is emitted as sulfate, SO$_4$, see parameter frac_so4 in sources_sinks of TM3. Is this ok? Note that the volcanoes are excluded from this redistribution.
• Where does the yearly scaling of aircraft emissions come from. Cannot we use the annual kerosine numbers? We could also use the scaling given by IPCC special report Fa1 scenario, table 6.2 of the special report, see above.
• The calculation of DMS fluxes from the ocean in getDMS of TM3 should use ECMWF SST and seaice cover (for Sc). Note that [Boucher et al., 2003] suggest that [Nightingale et al., 2000] is a better choice for the seawater-air DMS flux calculation. They suggest to use
  k $_{600}=0.23 u_{10m}^{2} + 0.1 u_{10m}$
  which explains 81 % of the variance in their data set, or
  k $_{600}=0.222 u_{10m}^{2} + 0.333 u_{10m}$
  which explains 80 % of the variance in a larger combined data set. Remaining variance may be due to bubbles/breaking waves and/or surfactant films [Frew, 1997]. There's also still a need for further evaluations at high wind speeds, such as hurricanes. The value of the exponent of -0.5 for the dependence on Schmidt number, Sc, was confirmed by [Nightingale et al., 2000].

  Note also that in TM5 we should not use the suggested 2m wind but the 10m wind.

• Take into account that IC flashes produce more NO$_{x}$ than previously assumed.
• The contribution 1.85/4.8 for non-eruptive volcanic emissions of SO2 may be removed in TM3, as has been done in TM5.
• What to for interannual variability when using EDGAR3.2 ?