The accuracy of climate modeling
I had a conversation on the university’s shuttle bus about a year ago with one of the climatologists here at Maryland. He’s one of the modeling guys; he develops models such as those used to make policy decisions about climate change and global warming.
The interesting thing is that he readily admitted that the models that the community uses are not particularly accurate — for instance, they can’t be verified beyond a very rough level of precision.
What I took away from the conversation was that all the hubbub over global warming is at best an argument over opinions, until we get better models. Our understanding of climate change comes largely from models whose accuracy is suspect. As a side note: the spectacularly cool work being done on Green Flash at LBL is meant to address just that.
Why am I mentioning all this? Seems it has now become a large-scale issue. From “Warming panel under attack, seeks review”:
The Intergovernmental Panel on Climate Change says it’s seeking some kind of independent review because of recent criticism about its four 2007 reports.
Critics have found a few unsettling errors […]
Kind of interesting. Personally, I have no idea whether human activity has caused global temperature increases … though I do remember the 1970s scare that the world was headed into a new ice age (no, really), and there seems to be plenty of evidence that the latest rise is both within historical variation limits and a natural phenomenon. Bottom line, I think we need better models and better validated models. Which means the guys at LBL (Oliker, Shalf, and Wehner) are working on precisely the right problem.
