A Climate Hurricane

A month after hackers broke into the CRU email server and released to the web email (it could have been a leak) correspondences between top climate researchers, what it all means is still being sorted. It apparently had little influence on the Copenhagen Summit as world leaders had momentum going in as well as other issues to sift through.

It is tempting to see the affair (dubbed climategate) as a small tempest on the otherwise tranquil sea of climate research--damaging perhaps to the scientists involved but lacking broader impacts. That would be a mistake. Limited in scope, though certainly broadcast widely, it reveals a suspicion scientists harbor about the research process that rarely gets articulated to a wider audience.

In my opinion the most important repercussion concerns scientific integrity. Climategate demonstrates that scientists can be quick to dismiss research ideas when they threaten their own. This can be relatively benign as rejecting/accepting a paper without careful review (editor's decide using multiple reviews) or worse when failing to cite the relevant literature undermining an essential scientific commitment to evaluating ideas on intellectual merit. It assumes a certainty of methods and ideas of one's own that's counter to the essential self-skepticism of the scientific enterprise. And it can be insidious when the behavior is passed on to a generation of students.

In basic fields, like particle physics, consequences of this type of behavior might decay rather quickly. In climate science where multiple plausible explanations are the norm as evidence is based on observations (not controlled experiments) and theory is incomplete or lacking, consequences have a much slower decay rate. And in a field with policy relevance, this can have a negative impact on the enterprise of science and thus on society as a whole.

The enduring lesson should be greater scientific integrity. Read and cite the relevant literature, analyze the data with proper tools, acknowledge the underlying assumptions, create information-rich graphs, write clearly, and most importantly, explain why you might be wrong.