Just finished my poster for next week's AGU Fall Meeting in San Francisco (coauthored with Jean-François Lamarque, Andrew Conley, Doug Kinnison and Francis Vitt, all at NCAR).
We have run several 150 year (1850-2000) CAMChem simulations where one of several "forcings" (NOx, VOC, CO emissions, methane emissions, aerosols, CFC concentrations, and climate) was kept at its pre-industrial (PI) conditions, allowing the others to evolve to their present day (PD) conditions as normal. The goal was to isolate the contribution of a given forcing to (e.g.) the overall tropospheric ozone radiative forcing - i.e. how much of the Pi to PD ozone change is due to methane emissions?. This was done by comparing these simulations against one where everything was allowed to evolve. Shindell et al. (2009) did something similar to this, but they kept everything at PI conditions save the one forcing that they allowed to evolve to its PD values.
The bar chart shows the contribution of the different forcings to the overall PI to PD tropospheric ozone radiative forcing, with the insert comparing the results to those from Shindell et al. (and our simulations are detailed in table). Clearly, for quantifying the contribution of different forcings it matters if you change one in a PI background (Shindell et al.) or keep one at PI conditions and allow the others to change. E.g., increasing NOx emissions with PI VOCs is different to increasing NOx emissions with PD VOCs (...we are just finishing a simulation that looks at NOx emissions alone).
Overall, separating out the contributions of a given forcing to a chemically-active greenhouse gas like ozone is complicated!