Geoscience Reference
In-Depth Information
modest increase in solar output might also have played a role in the early twentieth-century warming,
they concluded, but solar impacts could not explain most of the warming of recent decades.
Paleoclimatologist Jonathan Overpeck and collaborators reached similar conclusions in a subsequent
1997 study of Arctic temperature trends.
6
These more recent studies still suffered some limitations. The temperature reconstructions only
resolved decadal timescale variations, not individual years, so nothing could be said about, say, the
“year without a summer” of 1816 or the apparent mother of all El Niño events in 1791-1792.
Moreover, these studies reconstructed only a single time series representative of hemispheric mean
temperatures and thus could not establish precisely which regions were warm and which cold in a
given year; in other words, they didn't produce a spatial pattern of relative temperature. Finally, these
studies didn't provide any margin of error, or “error bars,” to indicate the uncertainty in the estimated
temperature changes given the imperfect and uncertain nature of the proxy records. While the recent
decades were nominally the warmest in these reconstructions, without knowing how much uncertainty
there was in the estimates, it was impossible to rule out, with any degree of confidence, that past
temperatures might have been as warm as today's.
It was at this point that I stepped onto the paleoclimate reconstruction scene. After defending my
Ph.D. dissertation in the spring of 1996, I was funded on a Department of Energy postdoctoral
(U. Mass) in Amherst, the town I grew up in. I would also be working closely with Ray's colleague
Malcolm Hughes from the University of Arizona, a specialist in the use of tree ring data.
After having left for college more than a decade earlier, I was quite literally returning home. I
moved into an upstairs apartment in the house I'd grown up in, just down the street from the center of
town. I now constantly ran into old friends and acquaintances, including some of my old high school
and elementary school teachers. My father had recently retired as a math professor from U. Mass, and
we only narrowly missed being colleagues. Indeed, some of my U. Mass colleagues were parents of
kids I'd grown up with. It was all a bit odd. But Amherst was a nice place to live as a “thirty-
something,” and I couldn't have been happier with my newfound academic home in the U. Mass
Department of Geosciences. It was a lively, friendly department. And it also happened to back up on
the U. Mass faculty tavern, a favorite gathering place on Friday afternoons and evenings.
The Game Begins
My postdoctoral research was aimed at developing and applying a new statistical approach to the
problem of proxy climate reconstruction. Seeking to improve upon previous efforts, I wanted to
reconstruct surface temperatures not just for individual decades, but also for individual years.
Moreover, I was interested in reconstructing the underlying spatial patterns of temperature variation,
not simply average trends over large regions like the Northern Hemisphere or the Arctic.
Reconstructing these spatial patterns would not only tell us where it was warm or cold in any
particular year, but also would give us insight into the workings of the climate system. It was these
patterns that could tell us about the long-term role of the El Niño phenomenon, for example, or the
wiggles in the track of the jet stream from year to year called the North Atlantic Oscillation (NAO).
How were these patterns influenced by external factors such as volcanic eruptions and changes in
