T2 Units & consistency Check

Please check your diagnostic output to verify that

1. You have used the correct units as specified in the SF6 Protocol;

2. You have the right global integral mass of SF6 (within a few percent of the initial mass plus the time integral of the emissions field); and

3. The numbers you are submitting are self consistent in terms of defining a mass balance for SF6 in each grid cell.

The consistency check involves comparing the sum of all the tendencies (advective, convective, diffusive, etc) to the time rate of change of SF6 in a slab of air. A perfect balance is not likely because of errors in spatial interpolation and temporal averaging, but this is a good way to find gross errors. Please analyze the "middle troposphere" slab at each grid cell and check that the imbalance in the budget is small compared to he individual terms in the budget. This is easy because there are no surface emissions to worry about, and little topography to get in the way.

In general, the budget of each grid box is given by a horizontally and vertically integrated version of the tracer transport equation:

This is easy. The horizontal and vertical integrals were essentially done by your model and the diagnostic subroutines you wrote. All you need to do is calculate the advective tendency implied by the divergence of the zonal, meridional, and vertical fluxes supplied in the diagnostics, then compare these to the subgrid-scale tendencies. Please do this for just one month of the final year of your simulation.

If you find that the budget does not balance and that the residual is dominated by one term, then look for a units mistake in that term or perhaps another bug in your calculation. In this case, we can decide whether it is possible for me to correct your diagnostics (say, by multiplying your diffusive tendency by 1000), or whether you should recalculate that field and resubmit.

For a derivation of the method outlined above, see the discussion by Mark Holzer as an Acrobat, or PostScript file.

For some example code on how to calculate advective tendencies from your diagnostic output, click here (text document).

Please feel free to call or email for more information on how to do this important calculation.