84
R.D . D'Arrigo and G.c. Jacoby
mean is one way to assess whether recent trends are exceeding past
natural variation (e.g., Cook et aI., 1992). Other techniques include tests
for nonstationarity (long-term changes in the mean) and t-tests to evaluate
the significance of differences in the means between two time periods
(Graumlich, 1991).
CO 2 Fertilization
The issue of whether natural vegetation is currently responding to direct
CO 2 fertilization (i.e. , enhanced growth due to the increased concentration
of CO 2 for photosynthesis) is a matter of considerable debate. In particular, this subject is being examined for high-elevation sites in the
southwestern United States (e .g., Graumlich , 1991; Graybill & Idso,
1993; LaMarche, Graybill, Fritts, & Rose, 1984). These sites are of
interest because it has been suggested that alpine vegetation may be
particularly sensitive to increasing levels of atmospheric CO 2 , due to
decreased CO 2 partial pressure with altitude (Cooper et aI., 1986; Gale ,
1972; Korner & Diemer, 1987). This effect may decrease the diffusion
gradient and hence the availability of CO 2 into the leaf, resulting in a
greater CO 2 limitation (and , presumably, response) than at sea level.
LaMarche et al. (1984) used graphical comparisons between highelevation western U .S. tree-ring chronologies and temperature and precipitation data to conclude that CO 2 growth enhancement may be taking
place. Graumlich (1991) determined thatCOz-induced growth enhancement is not occurring in subalpine conifers in the Sierra Nev-ada. She
based this conclusion in part on residual diagnostic statistics to evaluate
trends in the residuals following response surface and regression analyses
of tree growth and climate. Graybill and Idso (1993) suggested that CO 2
fertilization in the western United States may be most detectable in trees
with a stripbark morphology, in which most new biomass is allocated to
cambial production. D'Arrigo and Jacoby (1993a, and 1993b; Jacob y &
D 'Arrigo, 1989) evaluated residual estimates from regression models
relating tree growth to climate at sites in northern North America. No
significant positive residual trends (based on the Kendall Rank correlation
coefficient test and other statistical tests) were detected that were not
modeled by climate and might thus be caused directly by increasing levels
of atmospheric CO 2 (D'Arrigo & Jacoby , 1993b). Similar studies can be
done to evaluate whether a CO 2 fertilization effect is detectable in highelevation and high-latitude South American tree-ring records.
Conclusions
A brief overview of some recent topics in dendroclimatic research has
been presented, with an emphasis on regions of potential interest to the
AMIGO project. We have also suggested several possible topics for
R.D . D'Arrigo and G.c. Jacoby
mean is one way to assess whether recent trends are exceeding past
natural variation (e.g., Cook et aI., 1992). Other techniques include tests
for nonstationarity (long-term changes in the mean) and t-tests to evaluate
the significance of differences in the means between two time periods
(Graumlich, 1991).
CO 2 Fertilization
The issue of whether natural vegetation is currently responding to direct
CO 2 fertilization (i.e. , enhanced growth due to the increased concentration
of CO 2 for photosynthesis) is a matter of considerable debate. In particular, this subject is being examined for high-elevation sites in the
southwestern United States (e .g., Graumlich , 1991; Graybill & Idso,
1993; LaMarche, Graybill, Fritts, & Rose, 1984). These sites are of
interest because it has been suggested that alpine vegetation may be
particularly sensitive to increasing levels of atmospheric CO 2 , due to
decreased CO 2 partial pressure with altitude (Cooper et aI., 1986; Gale ,
1972; Korner & Diemer, 1987). This effect may decrease the diffusion
gradient and hence the availability of CO 2 into the leaf, resulting in a
greater CO 2 limitation (and , presumably, response) than at sea level.
LaMarche et al. (1984) used graphical comparisons between highelevation western U .S. tree-ring chronologies and temperature and precipitation data to conclude that CO 2 growth enhancement may be taking
place. Graumlich (1991) determined thatCOz-induced growth enhancement is not occurring in subalpine conifers in the Sierra Nev-ada. She
based this conclusion in part on residual diagnostic statistics to evaluate
trends in the residuals following response surface and regression analyses
of tree growth and climate. Graybill and Idso (1993) suggested that CO 2
fertilization in the western United States may be most detectable in trees
with a stripbark morphology, in which most new biomass is allocated to
cambial production. D'Arrigo and Jacoby (1993a, and 1993b; Jacob y &
D 'Arrigo, 1989) evaluated residual estimates from regression models
relating tree growth to climate at sites in northern North America. No
significant positive residual trends (based on the Kendall Rank correlation
coefficient test and other statistical tests) were detected that were not
modeled by climate and might thus be caused directly by increasing levels
of atmospheric CO 2 (D'Arrigo & Jacoby , 1993b). Similar studies can be
done to evaluate whether a CO 2 fertilization effect is detectable in highelevation and high-latitude South American tree-ring records.
Conclusions
A brief overview of some recent topics in dendroclimatic research has
been presented, with an emphasis on regions of potential interest to the
AMIGO project. We have also suggested several possible topics for