How Well Understood Are the Processes that Create Dendroclimatic Records? A Mechanistic Model of the Climatic Control on Conifer Tree-Ring Growth Dynamics - Dendroclimatology - page 64

0.5

a) b)

0.6

0.4

0.2

0

0

Simulation (d01)
Mohonk Residual

-0.2

Simulation (d01)
Mohonk Residual

-0.5 -0.4

MAY JUL SEP NOV JAN MAR MAY JUL SEP MAY JUL SEP NOV JAN MAR MAY JUL SEP

PREVIOUS YEAR GROWTH YEAR PREVIOUS YEAR GROWTH YEAR

3

c) d)

2

0.8

1

0.6

0

-1

0.4 G

G G G W

-2

0.2

-3

-4

JAN MAR MAY JUL SEP NOV JAN

1930 1940 1950 1960 1970

Year Month

Fig. 3.5 Simulation of Mohonk Lake tree-ring width chronology. Correlation of previous and cur-
rent year tree-ring widths with ( a ) temperature and ( b ) precipitation show similar patterns for both
simulated and actual chronologies. Dashed lines of the same color show the 95% two-tailed con-
fidence intervals from bootstrapping (1,000 draws with replacement; Biondi and Waikul [ 2004 ] ).
Values above these lines can be considered statistically significant, accounting for the number of
independent predictors ( c ) The simulation ( black line ) is correlated with the actual chronology
( gray line )at r

0.05). ( d ) Growth functions G T , G W , G E , and overall growth function
G from daily simulation output averaged for 1925-1973. See text, Section 3.8 . 1, for discussion

= 0.57( p <

based on 1925-1973 simulations (Fig. 3.5d ) . The modeling results suggest that
early season growth is strongly tied to the timing of early spring (March) warm-
ing, unless such warming is strong enough by early summer (May), in which case
warm conditions lead to growth limitation by moisture stress. The way in which
the VS model simulates such a phenomenon is described in Evans et al. ( 2006 ) .
Given these results, maybe in this specific case it's not emergence after all. Our
tentative conclusion is that current-generation multivariate linear regression models
may be unable to completely describe the environmental controls on tree-ring varia-
tions because of intraseasonal-interannual changes in the limiting factors controlling
tree growth (see also the example from Anchukaitis et al. [ 2006 ] , described below).
Forward modeling exercises like this can complement statistical model verification
procedures, assess the influence of such effects in linear paleoclimate inversions,
increase confidence in our interpretation of the data ('What do we expect to see?'),