The fact that, most of the site index variation is due
to two factors related to water balance (the annual
rainfall and the water surplus) corroborates the growth
modelling efforts to employ selected components of
the annual water balance. For instance, Grier et al.
(1988) developed a Douglas-®r productivity model
that explains approximately 60% of the variation in the
net primary production on the basis of leaf area,
components of the annual water balance and mineral
nutrients. The model by Wickramasinghe (1988),using only actual (effective) evapotranspiration,
explained between 40% and 60% of the variation in
the annual tree growth of Corsican and Scots pine. The
analytical site index/environmental factors model by
Klinka and Carter (1990), based on actual evapotranspiration
during May and June, the growing season
water de®cit and mineralizable-N in forest ¯oor and
mineral soil, explained 72% of variation in site index,
and even better results were obtained by categorical
models based on soil moisture and nutrient regimen
(84% of variation explained).