We addressed the topic of longterm growth trends by using stand-level data
gained in a network of permanent plots measured since 1930 in 4 Oak forests
of northern France. These forests spread along a climatic gradient, from t
he atlantic sites (Normandie) to the more continental sites (Lorraine). In
each forest, 10 plots have been observed, in stands of various ages, and su
bmitted to different silvicultures. We analysed dominant height and stand b
asal area increments. The modeling of height-age curves revealed different
curve shapes between forests. In 2 forests out of 4, the residuals exhibit
a very strong structure with date: growth rate has steadily increased since
the 1930s. Basal area increment was modelled with much more details, takin
g into account the effects of age, stand density, periodic fluctuations and
different growth levels between plots. These effects are combined with a s
mooth trend function of date, the amplitude of which was fitted on a per fo
rest basis. Several models were compared and we discussed the sensitivity o
f trends with regard to the underlying model. Whatever the model, we could
estimate trends of considerable amplitude: most of them are increases of ba
sal area growth of +25 to +50% between 1930 and 1990. Our results confirm v
ery closely those gained by dendrochronological methods. Curiously, Oak sta
nds from Normandie have a more complex trend: decreasing productivity betwe
en 1930 and 1960, then increase until now.