M. Messaoudene et L. Tessier, TREE-RING TO CLIMATE RELATIONSHIPS FOR PO PULATIONS OF QUERCUS-AFARESWILLD AND QUERCUS-CANARIENSIS POMEL IN ALGERIA, Annales des Sciences Forestieres, 54(4), 1997, pp. 347-358
Ring-width variations are analysed on 15 populations of Quercus afares
and Quercus canariensis in the Akfadou and Beni-Ghobri Forests (Alger
ia) (fig 1) which are submitted to temperate and cool variants of the
Mediterranean humid bioclimate. The populations selected are represent
ative of the different ecological conditions (substratum: exposure, al
titude) under which the two species are growing. After measuring ring-
width on three radii per tree and eight to 16 trees per site, a mean r
ing-width chronology is developed for each population (table I). In or
der to remove age trend and isolate interannual variations of radial g
rowth, ring-width series are modelized using ARMA processes, For each
population, a mean chronology of residuals is established. Response fu
nction is then calculated, involving the mean residual chronology as d
ependent variable and, in a first step, the 24 monthly parameters of a
nnual climate as regressors (P-T-max and P-T-min). The number of regre
ssors is then reduced by grouping together the successive monthly valu
es of the same sign. Response function is calculated for the period 19
18-1951, where valuable meteorological data are available. The reliabi
lity of response function is tested using the bootstrap method. Result
s obtained from modelization (table II) show that radial growth is lar
gely predetermined for both species. The amount of residual variance t
o be correlated with climate is low; this amount is lower for Q afares
than for Q canariensis. Statistical analysis reveals that neither alt
itude nor exposure or substratum (table III) can explain the differenc
es between populations. Comparison of the results obtained with respon
se functions (table IV) reveals that, for the two species, precipitati
on has a greater effect than temperature, the action of which concerns
only minimal values. The analysis of global significance of response
functions shows that Q afares does not react to climate whereas Q cana
riensis reacts significantly to the interannual fluctuations of climat
ic parameters. As the region is submitted to the same regional climate
, differences between populations can be attributed to differences in
habitats. However, just as with the growth models, it is not possible
to establish a significant relationship between the describers of the
habitats (altitude, substratum, exposure) and the response functions,
In conclusion, the endemic Q afares appears to be in better equilibriu
m with climate variations than Q canariensis; however, as the two spec
ies are highly in competition in the same habitat, this equilibrium co
uld be perturbated by a small change in climate.