How do we know what we know?

The WorldChanging blog has piece about monitoring 315 miles of the Hudson on a "minute-to-minute basis." While the work in itself is interesting and important, the author of the article mentions that "we are quickly reaching a point where anything that's measurable about the environment that we want to know, we will."

This raises some often overlooked questions about the environmental data we do have. Take global temperature data. It seems simple to take a temperature measurement, but all sorts of complicated issues creep into it. Where are you taking these measurements - rural or urban settings? If it is urban, do you want to correct for the Urban Heat Island effect? What if your temperature station started rural but, over the years, the urban area has built up to your doorway? These sorts of issues abound.

A big buzz-word in business is "data-based decision making." That's pretty hard to argue against. Of course data is good. But one needs to be critical of how that data was gathered and what systematic bias may be introduced.