On uncertainty and water modeling

As I wrote earlier, it looks like any model that is developed to help understand Upper County groundwater is likely to come with a fair amount uncertainty, in spite of the USGS's powerful computer model. This is because Upper County geology is complicated, and because rainfall amounts change rapidly with elevation and distance, not to mention the unknown but real future effects of global climate change.

I was trained as a scientist, and I taught science in college for many years. The best definition I ever came up with for science (I was actually rather proud of this) is that science is the management of uncertainty.

Scientists know that it is impossible to eliminate uncertainty, which is why they start to squirm when people start talking about "proving" anything. To a scientist, there's no such thing. Mathematicians use the word "proof" and mean it, but even in that field there can be room for judgment and reasonable disagreement.

Uncertainty can be reduced and managed, but never completely eliminated. Reducing and managing uncertainty can be difficult and expensive, and eventually one has to decide that enough time, effort, and money have been spent, and just accept the uncertainty.

Even so, when extended to the Upper County, the model will be useful. What computer models are really useful for is testing different management scenarios, which then can be used, carefully, for taking action and setting policy. It is uncertain when this will become possible, but it should be within months or a very small number of years.