Its slightly more complicated than that since you also need a surface
- provided by the Polar Stratospheric Clouds during the winter and spring. Which also brings up the point that you probably don't really need UV protection when there's hardly any sunlight anyway. Ground UV levels may rise as a percentage of incident in the spring, but how much does this affect the peak UV flux during the summer or the total UV dose integrated over a full year? It seems to me those two numbers should be more important if you are worried about how it affects plants and animals on the ground, not how much ozone is gone in a cold atmosphere with a grazing angle incidence of sunlight (unless your real goal is to keep your research funded).
Certainly you can slant assumptions and pick and choose data to support whatever your pet view is. Requiring a cost/benefit analysis doesn't necessarily solve that, but it should allow everyone to see the assumptions and the development of the analysis so they can take shots at it. The linear dose/response model should be viewed as an assumption. It is certainly plausible that the body has some repair capability that should make it non-linear, however it is then incumbent on you to select one and provide a reasonably persuasive argument for preferring it over the simple linear one - which will probably be hard to do. The system also needs to enforce a high benefit/cost ratio requirement since you will invariably have hidden, unanticipated costs and it is certainly not unusual for advocates to overstate the benefits. This also means you really need to calibrate the required benefit/cost ratio by going back to previous decisions and attempting to assess what the real benefits and costs were. I suspect something like a factor of five would be marginal if you want
95% confidence that a given change would actually provide a break-even benefit. There are also opportunity costs that may be hidden as well.The Peer-Review process is the worst - except for all of the others. Science isn't perfect, either, but at least it is self-correcting over time.
Brad Hitch