Sunday, December 19, 2004


Science for the Birds

A study just published in the Proceedings of the National Academy of Sciences estimates that ten-percent of the world's bird species will be extinct, and as many as 25-percent may be functionally extinct, by the year 2100. “A careful study of extinction rates so far, conservation measures underway and climate and environmental change,” the Stanford researchers told CNN, “shows that at least 1,200 species of birds will be gone by 2100. And that is a conservative estimate.” If only modeling nature was that easy… Substituting adjectives for analysis, the researchers followed in the tracks of an increasing number of scientists and described their study and its results as “careful" and "conservative" to cover up the inherently shaky nature of wildlife modeling. Two years ago, I studied wildlife population modeling from one of the top wildlife ecologists in the world, Oz Schmitz. Prof. Schmitz strives to teach his students the unspoken reality of wildlife modeling: expect faulty results due to inadequacies in the models' basic assumptions. There is rarely enough information to say anything reliable about wildlife populations, so most scientists cut a lot of corners and build inferences upon inferences to arrive at their predictions. They then call the predictions conservative by simply taking the most conservative number from the range that results from their study....

No comments: