Abstract• Studies on wood density variations are necessary for estimating the forest carbon pool. Further, they can help predict the technological properties of wooden end-products. While there have been frequent reports on the relationships between wood density, cambial age, and ring width, there is little information about the historical trend in wood density for the last century, particularly in the context of global climate change.• In this study, different sources of variations in mean ring density (site, tree, ring age, ring width, and calendar date) were studied using an original sampling design. A total of 105 Norway spruce (Picea abies Karst.) trees were sampled in north-eastern France, from thirteen sites with trees of different ages growing at the same site and in similar conditions. X-ray densitometry measurements were performed on samples taken at breast height. The chronology of the mean ring density over the 20th century was estimated by using a statistical modelling approach based on linear mixed-effects models after accounting for the effect on the mean ring density introduced by different ring widths, cambial ages, sites, and trees.• The mean ring density of Norway spruce was seen to decrease by about 18 kg m −3 relative to the year 1900. The chronology showed no evolution between 1900 and 1950, a steep decline from 1950 to 1980 (reaching a maximum decrease of 30 kg m −3 ), followed by an increase from 1980 to 2000.• The observed decrease was consistent with the results of previous works and supports the hypothesis that this could indicate a global trend and that this trend is independent of the wood structure. Moreover, high inter-annual density variations were found. In future studies, the influence of climate on the wood density and within-ring properties must be clarified to identify the anatomic causes for wood density variations.

International audienceContext. While historical increases in forest growth have been largely documented, investigations on historical wood density changes remain anecdotic. They suggest possible density decreases in softwoods and ring-porous hardwoods, but are lacking for diffuse-porous hardwoods. Aims. To evaluate the historical change in mean ring density of common beech, in a regional context where a ring-porous hardwood and a softwood have been studied, and assess the additional effect of past historical increases in radial growth (+50% over 100 years), resulting from the existence of a positive ring size-density relationship in broadleaved species. Methods. 74 trees in 28 stands were sampled in Northeastern France to accurately separate developmental stage and historical signals in ring attributes. First, the historical change in mean ring density at 1.30m (X-ray microdensitometry) was estimated statistically, at constant developmental stage and ring width. The effect of past growth increases was then added to assess the net historical change in wood density. Results. A progressive centennial decrease in mean ring density of -55kg.m-3 (-7.5%) was identified (-10% following the most recent decline). The centennial growth increase induced a maximum +25kg.m-3 increase in mean ring density, whose net variation thus remained negative (-30kg.m-3). Conclusions. This finding of a moderate but significant decrease in wood density that exceeds the effect of the positive growth change extends earlier reports obtained on other wood patterns in a same regional context and elsewhere. Despite their origin is not understood, such decreases hence form an issue for forest carbon accounting

No abstract

<p style="text-align: justify;">Un modèle expérimental miniaturisé a été conçu dans le but de simuler "l'interface tonneau". Les cinétiques de l'imprégnation du liquide dans le bois et de l'évaporation de surface ont été mesurées en conditions de cave (15°C ; 90 p. cent HR) pendant 200 jours sur des petits disques à orientation tangentielle (Ø 63 mm; épaisseur 9 mm) à l'aide d'une méthodologie spécifique. Ces résultats expérimentaux ont été obtenus sur 4 espèces traditionnellement employées en tonnellerie (<em>Quercus</em> <em>robur </em>L., <em>Q. petraea</em> Liebl., <em>Castanea sativa</em> Mill., <em>Robinia pseudacacia</em> L.).</p><p style="text-align: justify;">L'analyse des cinétiques d'imprégnation et d'évaporation a permis de caractériser les régimes transitoire et permanent (état d'équilibre). Le régime transitoire est dominé par une forte imprégnation pendant les 40 premiers jours environ, qui décroît ensuite de façon asymptotique. Parallèlement, après un temps de latence appelé "retard à l'évaporation", le flux massique devient progressivement une fonction linéaire du temps. Le régime d'équilibre est atteint lorsque le niveau d'imprégnation de la pièce de bois est maximal et constant soit après 140 jours environ (profil d'humidité d'équilibre). Le fonctionnement en régime stationnaire se caractérise alors par un flux d'imprégnation à la face inteme équivalent au flux d'évaporation à la face exteme.</p><p style="text-align: justify;">Les variations quantitatives des phénomènes sont apparues liées aux conditions de transfert externe (HR de l'air ambiant) et aux conditions de transfert interne (propriétés de la pièce de bois). La comparaison d'échantillons toutes conditions égales par ailleurs a ainsi montré des différences significatives de comportements vraisemblablement liées à la variabilité anatomique du bois (interspécifique, interindividuelle et intraarbre).</p><p style="text-align: justify;">Cette variabilité pourrait avoir des conséquences sur les phénomènes oxydatifs et sur l'extraction des constituants du bois au cours de l'élevage des vins et des eaux-de-vie en fûts.</p>

No abstract