Spent some time this evening with an old paper by Tom Swetnam and Julio Betancourt on climate variability and its effects on ecosystems in the southwest. It was published in the Journal of Climate in 1998, which I don't think is on line, though there seems to be a short version here. Reading it years ago marked the first time I really got the linkage between decadal climate variability, drought, and widespread wildfire.

Here's the part that got my attention:

Based on tree ring records (fires leave nice, measurable and datable scars) fire frequency in any one spot averaged about once in every 7.5 years. Swetnam and his buddies in the tree ring mafia have dated scads of these things all over the southwest in order to come up with what everyone seems to agree are pretty good numbers - 63 separate sites with fire records covering 1700 to 1900. (1900's the start of fire suppression, so you have to cut off the study there and significantly rethink the way things have played out since).

If the statistics were essentially random, you'd expected widespread fire years - a whole bunch of sites being burned simultaneously during a single year - to be relatively rare. Here's how they put it: "At this frequency, strictly by chance, we would expect about one coincidence of the same fire date in 21 of the 63 sites (one-third) in about a 35,000 year period." Instead, they found that happened 15 different years during that time. One year (1748) saw 41 of the 63 sites burn ("one chance in billions", they say).

They then compared those widespread fire years with a reconstruction of the Palmer Drought Severity Index for the region and found a strong correlation. Tack on a lag with antecedent moisture conditions and the thing is rock solid. Wet years allow stuff to grow, then drought sets in and stuff burns like hell, all over the southwest.