Stable Isotopes in Dendroclimatology: Moving Beyond ‘Potential’ - Dendroclimatology - page 163

and maximum densities of pine trees in northern Finland to recognize years with
very low summer rainfall, even though moisture stress did not limit the growth of
the trees. Gagen et al. ( 2004 ) have shown how proxies might be expected to covary
under different climate change scenarios, but so far stable isotope dendroclimatol-
ogy has not been used to produce chronologies that contain any climate parameters
that could not have been produced by using the traditional proxies. However, the
need for reduced data treatment and statistical detrending is an attractive char-
acteristic may provide access to a greater proportion of lower-frequency climate
information and represent a unique advantage of the isotopic proxies.

6.5.3 Tropical Isotope Dendroclimatology

Traditional dendrochronology has been unable to establish itself in large parts of
the tropics, and only extratropical species have contributed significantly to large-
scale paleoclimate reconstructions (Mann 2002 ) . Whilst tropical climatology may
prevent the formation of an annual ring boundary, the tropics are, in fact, the only
place where we potentially have a signal for the entire year, because growth does
not necessarily stop. Stable isotopes are capable of capturing this signal, even where
no visible ring is seen. Sites in the tropics may not exhibit the temperature season-
ality needed to form rings, but in the majority of areas, there is a seasonal cycle
in precipitation, arising either from wetter and dryer seasons, or from a shift in the
dominant trajectory of precipitation-bearing air masses, which seems to be recorded
in wood cellulose in a cyclical way. Evans and Schrag ( 2004 ) coined the term 'trop-
ical isotope dendroclimatology' and presented empirical and theoretical evidence
supporting the capture of rhythmical seasonal cycles in

18 O.
Seasonal cycles in precipitation isotopes may also arise from a fractionation
effect in the hydrological cycle, known as the 'amount effect'—the process by which
preferential rainout of the heavier water isotopes ( 18 O and 2 H) leaves subsequent
precipitation depleted. The

δ

18 O of precipitation is negatively correlated with rain-
fall amount, and this effect is one of the primary features of intra-annual variability
in the isotopic composition of rainfall in the tropics (Gat 1996 ) . Layered on top
of this seasonal cycle, the amount effect also captures a climate signal; the greater
the amount of precipitation within a season, the more depleted the

δ

18 O signal in
the wood products (Evans and Schrag 2004 ; Anchukaitis et al. 2008b ) . The amount
effect can be used to explain precipitation

δ

18 O over a range of timescales such that

δ

18 O values, and unusually dry years are
offset as high values (Evans and Schrag 2004 ) . It may also be possible to use cycles
in cellulose

unusually rainy years are offset as low δ

18 O derived from geographical shifts in dominant moisture source to
derive climate records.
Seasonal variations in cellulose

δ

13 C also retain an annual signal. The seasonal

δ

13 C of atmospheric CO 2 in the tropics should be between 0.4‰
and 0.6‰ (Mook et al. 1983 ) and a seasonal cycle in discrimination, within the
tree, will layer on top of this. Poussart et al. ( 2004 ) measured interannual

amplitude in δ

13 C

δ

18 O variability in Thai Podocarpus species, finding greater amplitude in 18 O

and δ δ