Book review: Storm World.

When I was growing up in New Jersey, hurricanes were "on the radar" for us, one of many possible (if infrequent) weather patterns during summer and fall. Later, in my first semester of college in Massachusetts, the morning of my first broadcast on the college radio station was made memorable by the landfall of Hurricane Gloria; I remember the name of the storm because I closed my show by playing the U2 song "Gloria" before signing off the air at 7 am. (The governor of the Massachusetts had just declared a state of emergency, although it wasn't until some 30 minutes later that the trustees of the college decided it might be a good idea to cancel classes.) The Atlantic coast, and the inland area not too far from it, was a site of weather events that commanded your attention.

For nearly two decades now, I've lived in California, in an area described as having climate, not weather. When I started reading Chris Mooney's new book Storm World, a part of me wondered whether the fact that hurricanes are no longer a part of my day to day life would make it harder for me to get into this book about the scientific efforts to understand and forecast hurricanes, and the political struggles around the science.

My worry was misplaced.

The big story here, and one which Mooney lays out clearly and compellingly, is the struggle between empiricists and theoreticians over the right way to understand phenomena. It's a struggle that plays out in many other scientific arenas (including the corner of chemistry I used to inhabit), but one which is likely less familiar -- or even surprising -- to the reader who is a non-scientist. One camp focuses on getting as much accurate data as possible about the phenomenon of interest, using that data to produce detailed descriptions of the phenomenon and, perhaps, to make careful projections forward on the basis of that data about what will happen next. The other camp strives to build models of the phenomenon of interest, isolating the essential factors that affect how the phenomenon manifests itself -- since being able to build a good theoretical model that nails down the causal contributions to the phenomenon means that you understand the phenomenon on a deeper level.

There is usually some tension in a scientific field between those who stick close to the empirical data and those who reach beyond the data to model the underlying dynamics. (Some will trace this tension back to Aristotle, who saw the essence of X as the kind of thing one should determine by close inspection of many particular specimens of X, and Plato, who thought every particular specimen of X was a necessarily imperfect material instantiation of the ideal form of X.) Around the weather, however, the tension is strong enough to have created distinct sub-tribes of science. The meteorological forecasters thought the right approach to hurricanes was to spot the trends in their voluminous data. To their way of thinking, the modelers abstracted away too many of the features of the real systems, so whatever they were "understanding" from their modeling, it couldn't have been real hurricanes. On the other hand, the modelers thought it was important to identify the factors that were causing the trends -- not least because this would be the best way of understanding what could happen with hurricanes if the background conditions of air and sea changed. Indeed, that very question -- which features of the environment are stable and which could change -- turns out to be a site of major disagreement for the forecasters and the modelers.

A lovely feature of this book is the way Mooney brings us into the scientists' world, tracing the historical origins of the meteorological debate and bringing us in contact with some of the personalities that make it difficult for the two camps to engage with each other. We see how the trajectory of one's personal training, the shifting "hot topics" within a field, and the changes in funding priorities set up conditions where it is hard for scientists to be receptive to or objective about differing approaches to the same phenomena.

Mooney also presents details that spark further rumination on the very nature of the scientific enterprise. For example, what should we make of Bill Gray, a scientist utterly dismissive of mathematical modeling but with a dead eye for the trends in large data sets? How much do we need to know about phenomena in order to judge whether our estimates about their properties are good ones or bad ones? Is stability a better assumption than flux? In cases where more recent data -- collected with better detection methods -- seems to depart from older data, is it more likely that things are really changing, or that the differences can be laid at the feet of the older detection methods? In the face of incomplete data, are there principled scientific reasons to prefer a high burden of proof (e.g., on the question of whether global climate change could increase the severity of hurricanes, or whether severe enough hurricanes might actually have effects on the climate) rather than opting for the "precautionary principle"? And, in wildly complex systems with lots of parameters, what properly counts as a cause of particular hurricanes? Of what can a change like global warming properly be counted as a cause?

For some readers, this glimpse into the workings of a field of science whose practitioners are divided will be most interesting when the influence of politics is added in. Mooney does not disappoint. He sets the scientific battles in their political context, and gives an especially interesting account of the ways communication between government scientists (supported with public money) and the public (often by way of journalists) changed as the Bush administration took pains not to be put in a position where the prospect of global climate change would demand any measures that might be inconvenient for corporate America. (It would be wrong to say the prospect of global climate change elicited no change in policy from the administration -- policies about granting interviews certainly changed.) For every scientist who has wished the public cared about her area of research, this glimpse of the ways political constraints impinge on the day to day activities of scientists may make her wish for something else.

Needless to say, Mooney also gives nice explanations of how hurricanes work, at least given our current understanding of things.

Much more could be said about this book (Josh gives his take here), but for me the interesting story is how preexisting disciplinary and methodological divides among scientists fed into political divides in the larger society, creating a perfect storm where the questions of whose knowledge is reliable, what knowledge necessitates action, and what hangs in the balance, swirl around with increasing speed and ferocity.