Geoscience Reference
In-Depth Information
discrediting the science underlying climate change. Any scientific work perceived to bolster the case
for the reality and threat of human-caused climate change is likely to receive the unwanted attention of
the climate change denial machine. This applied to the early work of climate scientists Stephen
Schneider and James Hansen. It applied to Benjamin Santer and his important contributions to the
detection of human-caused climate change in the mid-1990s. And it now had come to apply to my
coauthors and me, and the hockey stick.
Scientific Give-and-Take
In the years following publication of the hockey stick, a healthy scientific debate did develop within
the paleoclimate research community. That debate played out in the peer reviewed scientific
literature. A number of different reconstructions were produced, and while they all appeared to come
to the same overarching conclusion—that recent warming was unusual in at least a millennial context
—they didn't agree on all the details: How cold was the Northern Hemisphere on the whole during
the Little Ice Age? How warm was it at the height of the medieval climate anomaly/medieval warm
period? To the extent that various reconstructions
1
differed on these points, were the differences due
to the differing mix of proxy data being used, the differing statistical methodologies, or both?
Some of the differences, my coauthors and I argued, could be attributed to the differing
seasonality or differing regional emphasis reflected in the various competing estimates. Our own
reconstruction was based on an averaging of temperatures over the full Northern Hemisphere—
tropics and extratropics, land and ocean—and over the entire year (winter and summer). Other
estimates, such as those of Jones et al. and Briffa et al., primarily reflected summer only and were
weighted more toward the extratropics and the continents only. Fluctuations in the pattern of the jet
stream could lead to temperature shifts over regions such as North America and Eurasia much larger
than those observed in other regions, for example. And, as noted earlier, opposing temperature shifts
in the tropical Pacific might tend to cancel out temperature changes in extratropical regions. Such
factors complicated comparisons of different proxy reconstructions that emphasized different regions.
We debated the implications of these various considerations with these other researchers in numerous
exchanges in the scientific literature.
2
These differences were about the details; none of this work
disputed the basic hockey stick finding that recent warmth is anomalous in a millennial context.
Other challenges to our work came from scientists reconstructing past temperatures from
borehole measurements. Henry Pollack and his team published a revised temperature reconstruction
in
Nature
in 2000.
3
They reiterated their original conclusion that recent warming was anomalous in a
long-term context (five hundred years), but they argued that the Little Ice Age during its peak was
roughly 0.5 °C cooler than reconstructions from “traditional proxies” (tree rings, ice cores, corals,
etc.) indicated. Pollack and colleagues asserted that the discrepancy might be due to the inability of
data from those proxies (in particular, tree ring data) to resolve very long-term variations. This claim
was certainly plausible. It was well established that there were potential limitations in reconstructing
long-term trends from tree ring data owing to something known as the “segment length curse”: When
joining together data from trees that lived at different times to form a single, long-term composite
measure, it can be challenging to retain climate information on timescales longer than the lengths of
the individual segments (i.e., the lifetime of the individual trees).
4
We were careful to acknowledge
