The best soil-site model for the Montana and Idaho
Douglas-®r stands developed by Monserud et al.
(1990), explained only 16% of the total variation in
site index, when only edaphic factors were included.
For the same area, the best model, including aboveground
properties like elevation, habitat series, precipitation
and longitude, explained 42% of variation in
site index (Monserud et al., 1990). The best model on
better sites in Scotland explains 45.5% of the variation
in General Yield Class of Douglas-®r by means of
spring temperature, topographical exposure, crop age,
winter rainfall and site drainage (Tyler et al., 1996):
Table 4
Estimates, standard errors and t values for the final model
b estimate Standard error t
Constant ÿ6.9 7.42 0.92
EWASP ÿ1.78 0.46 3.8
SURPHY ÿ0.04 0.01 3.75
PCLAY ÿ0.21 0.03 5.79
CALCA ÿ0.77 0.16 4.69
ARAIN 0.05 0.11 4.35
the unusual `age effect' was likely to be associated
with improvements in silviculture and/or genetic stock
through time. The best analytical model by Klinka and
Carter (1990) explained 72% of the variation in site
index of immature coastal Douglas-®r stands: it is
based on factors not readily assessable in the ®eld, like
actual evapotranspiration during May and June and
mineralizable-N in forest floor and soil.