Geoscience Reference
In-Depth Information
alone when it might have ended or begun. Some went so far as to attack us for only going back six
centuries, claiming we had intentionally stopped short to avoid running into the putative medieval
indicated: that with the data we had, we couldn't go back any further and still obtain meaningful
estimates. Why would anyone impugn our integrity so brazenly and make an allegation so false? Boy,
would I later be in for a lesson.
Shortly after our study was published, Phil Jones and collaborators published another attempt to
trace the Northern Hemisphere's mean temperature. They used the same composite approach as had
Bradley and Jones in 1993, but this time they extended the estimate back over the entire past
millennium.
21
We were a bit skeptical. Our own statistical tests, after all, had told us that it was not
possible to obtain a meaningful reconstruction further back than six centuries, using more or less the
We decided to examine more closely the two dozen proxy records we had that extended back six
centuries or more, which included those used by Jones and colleagues, and several others. I
performed a series of so-called sensitivity tests, in which various proxy records are removed or—to
use standard statistical terminology, “censored”—from the network, and the sensitivity of the results
to those records is gauged by noting how much of an effect their removal has on the result. I titled the
computer directory “censored” accordingly—a choice I would later regret.
The tests revealed that not all of the records were playing an equal role in our reconstructions.
Certain proxy data appeared to be of critical importance in establishing the reliability of the
reconstruction—in particular, one set of tree ring records spanning the boreal tree line of North
didn't extend any further back than
A.D.
1400, however, and that's when we could no longer achieve a
reliable reconstruction. This result actually made perfect sense since the pattern of twentieth-century
warming in the instrumental record showed particularly strong warming in western and northern
North America—precisely where these tree ring data were located. Moreover, a previous analysis of
long climate model simulations had suggested that western North America was a “sweet spot” for
estimating the average temperature of the Northern Hemisphere.
24
We had other proxy data from this region, and they extended considerably farther back in time, in
many cases more than a thousand years. There was a group of tree ring records from high-elevation
sites in the western United States (primarily California, Nevada, and Arizona) that should in principle
have reflected temperature variations in the region. Yet, despite the availability of these data, our
validation tests were telling us that we couldn't go further back than six centuries. Intrigued by this
apparent inconsistency, I undertook a closer comparison of the two different North American datasets.
Something rather remarkable emerged when the two datasets were laid on top of each other. The
two tracked each other almost perfectly from the beginning of their period of overlap (
A.D.
1400) until
the early nineteenth century. Then the two series began to diverge, with the western U.S. series
showing an almost exponential increase in tree ring growth relative to D'Arrigo and Jacoby's boreal
tree line series. Then, at the beginning of the twentieth century, the two series began tracking each
other again. As it turns out, the likely reason for this enigmatic observation had already been
discussed in the scientific literature.
It was known that high-elevation trees are often not limited in growth simply by climate
conditions such as growing season warmth, but also by carbon dioxide levels. Trees need carbon
