Stable Isotopes in Dendroclimatology: Moving Beyond ‘Potential’ - Dendroclimatology - page 160

Hemisphere. This reconstruction certainly satisfies Hughes' (2002) query (and it
is based on an average sample depth of 14 cores from seven trees, measured as
a pooled record with a minimum n of 3 in its earliest portion); furthermore, it
contributes significantly to climatology through evidence that changes in precip-
itation patterns probably exceed natural variability over the recent past (Evans
2006 ) .
In addition to robust 'single-site' studies, 'network' stable isotope dendroclima-
tology projects are also providing exciting results. The European Union-funded
'ISONET' project brought together many laboratories to produce 400-year isotope
chronologies from across Europe, mostly based on pooled material. Those data are
being used to examine the spatial coherence of the isotope signals from tree rings
(Treydte et al. 2007 ) . Gagen et al. ( 2007 ) reconstructed the summer temperature of
northern Finland back to AD 1640 based on a well-replicated

13 C record, without
pooling, and were able to demonstrate strong calibration and verification statis-
tics. The EU-funded 'Millennium project' ( www.millenniumproject.net ) is currently
building stable isotope chronologies at 12 sites across Europe, all of which will
extend for at least 500 years, and most for more than 1,000 years. These isotope
records will be integrated with tree-ring width, density (or blue intensity: McCarroll
et al. 2002 ) measurements and with other proxies, including those obtained from
peat mires and lake sediments, to produce multiproxy climate reconstructions.

δ

6.5 Future Directions

Tree-ring width and density proxies have traditional geographical strongholds where
it will be hard for stable isotopes to contribute more, and probably not an efficient
use of resources to try. Correlations between ring width/density and climate are
famously high; for example, at the northern timberline (e.g., Linderholm and Eronen
2000 ; Linderholm and Gunnarson 2005 ) . Here, stable isotope records are unlikely
to be able to compete as a proxy for summer temperature. Although, stable carbon
isotopes do appear to display comparable results to the traditional proxies, even in
the stronghold regions of the latter (Fig. 6.4 ) , and also show equally high, sometimes
higher, common signal strengths (Robertson et al. 1997a , b; McCarroll and Pawellek
1998 ; McCarroll et al. 2003 ; Gagen et al. 2004 ; Table 6.1 ) . Robertson et al. ( 1997a ,
b) found that generally slightly fewer

13 C series are needed than ring width series
for equivalent r values, with high-frequency indices only. The same information
is not yet known for the lower-frequency signal in stable carbon isotopes. Signal
strength in

δ

13 C may well be variable both spatially and by frequency.
In the example shown in Fig. 6.4 , stable carbon isotopes from latewood cellulose
and maximum latewood density capture slightly different (but strongly autocorre-
lated) climatic variables in the southern French Alps, with density recording summer
temperatures most strongly and stable isotope ratios (in a relatively dry region)
responding most strongly to summer precipitation (see Gagen et al. 2004 for details).
The spatial pattern of correlation with gridded climate data reveals a comparably
strong climatic signal in the

δ

13 C series compared to the maximum latewood density

δ