Genetic differences in δ¹³C (isotopic composition of dry matter carbon) have been evidenced among poplar genotypes at juvenile stages. To check whether such differences were maintained with age in trees growing in plantations, we investigated the time course of δ¹³C as recorded in annual tree rings from different genotypes growing at three sites in southwestern France and felled at ∼15-17 years. Wood cores were cut from tree discs to record the time course of annual basal area increment (BAI). The isotopic ratio δ¹³C was recorded in bulk wood and in extracted cellulose from the annual rings corresponding to the period 1996-2005. Discrimination against ¹³C between atmosphere and tissues (Δ¹³C) was computed by taking into account the inter-annual time course of δ¹³C in the atmosphere. Annual BAI increased steadily and stabilized at about 8 years. An offset in δ¹³C of ∼1‰ was recorded between extracted cellulose and bulk wood. It was relatively stable among genotypes within sites but varied among sites and increased slightly with age. Site effects as well as genotype differences were detected in Δ¹³C recorded from the cellulose fraction. Absolute values as well as the genotype ranking of Δ¹³C remained stable with age in the three sites. Genotype means of Δ¹³C were not correlated to annual BAI. We conclude that genotypic differences of Δ¹³C occur in older poplar trees in plantations, and that the differences as well as the genotype ranking remain stable while trees age until harvest.

& Key message The diversity of forest management systems and the contrasted competition level treatments applied make the experimental networks of the GIS Coop, a nationwide testing program in the field of emerging forestry topics within the framework of the ongoing global changes. & Context To understand the dynamics of forest management systems and build adapted growth models for new forestry practices, long-term experiment networks remain more crucial than ever. & Aims Two principles are at the basis of the experimental design of the networks of the Scientific Interest Group Cooperative for data on forest tree and stand growth (GIS Coop): contrasted and extreme silvicultural treatments in diverse pedoclimatic contexts. & Methods Various forest management systems are under study: regular and even-aged stands of Douglas fir, sessile and pedunculate oaks, Maritime and Laricio pines, mixed stands of sessile oak, European silver fir, and Douglas fir combined with other species. Highly contrasted stand density regimes, from open growth to self-thinning, are formalized quantitatively. & Results One hundred and eighty-five sites representing a total of 1206 plots have been set up in the last 20 years, where trees are measured regularly (every 3 to 10 years). The major outputs of these networks for research and management are the calibration/ This article is part of the Topical collection on Mensuration and modelling for forestry in a changing environment.

Reliable estimates of poplar plantations area are not available at the French national scale due to the unsuitability and low update rate of existing forest databases for this short-rotation species. While supervised classification methods have been shown to be highly accurate in mapping forest cover from remotely sensed images, their performance depends to a great extent on the labelled samples used to build the models. In addition to their high acquisition cost, such samples are often scarce and not fully representative of the variability in class distributions. Consequently, when classification models are applied to large areas with high intra-class variance, they generally yield poor accuracies. In this paper, we propose the use of active learning (AL) to efficiently adapt a classifier trained on a source image to spatially distinct target images with minimal labelling effort and without sacrificing classification performance. The adaptation consists in actively adding to the initial local model, new relevant training samples from other areas, in a cascade that iteratively improves the generalisation capabilities of the classifier, leading to a global model tailored to different areas. This active selection relies on uncertainty sampling to directly focus on the most informative pixels for which the algorithm is the least certain of their class labels. Experiments conducted on Sentinel-2 time series showed that when the same number of training samples was used, active learning outperformed passive learning (random sampling) by up to 5% of overall accuracy and up to 12% of class F-score. In addition, and depending on the class considered, the random sampling required up to 50% more samples to achieve the same performance of an active learning-based model. Moreover, the results demonstrate the suitability of the derived global model to accurately map poplar plantations among other tree species with overall accuracy values up to 14% higher than those obtained with local models. The proposed approach paves the way for national-scale mapping in an operational context.

To anticipate European climate scenarios for the end of the century, we explored the climate gradient within the REINFFORCE (RÉseau INFrastructure de recherche pour le suivi et l’adaptation des FORêts au Changement climatiquE) arboreta network, established in 38 sites between latitudes 37° and 57°, where 33 tree species are represented. We aim to determine which climatic variables best explain their survival and growth, and identify those species that are more tolerant of climate variation and those of which the growth and survival future climate might constrain. We used empirical models to determine the best climatic predictor variables that explain tree survival and growth. Precipitation-transfer distance was most important for the survival of broadleaved species, whereas growing-season-degree days best explained conifer-tree survival. Growth (annual height increment) was mainly explained by a derived annual dryness index (ADI) for both conifers and broadleaved trees. Species that showed the greatest variation in survival and growth in response to climatic variation included Betula pendula Roth, Pinus elliottii Engelm., and Thuja plicata Donn ex D.Don, and those that were least affected included Quercus shumardii Buckland and Pinus nigra J.F.Arnold. We also demonstrated that provenance differences were significant for Pinus pinea L., Quercus robur L., and Ceratonia siliqua L. Here, we demonstrate the usefulness of infrastructures along a climatic gradient like REINFFORCE to determine major tendencies of tree species responding to climate changes.

malgré une surface représentant seulement 2 % de la forêt française, le Peuplier a sa place au sein de la production forestière française grâce à un niveau de productivité sans égal, ou presque. Le Peuplier tient bien son rang au sein des feuillus, face aux Chênes et au Hêtre. Très étudié par les scientifiques et malgré sa technicité, il est populaire et ne laisse pas indifférent. mais pourquoi s'inquiéter de son avenir ? L' avenir du Peuplier questionne, car depuis vingt ans les bouleversements se succèdent pour arriver actuellement à un niveau de morosité que la populiculture française n'a pas connu par le passé. mais l'avenir reste prometteur car le Peuplier a plus d'un atout sous son écorce. de noMbreuX bouleVerseMents depuis 20 ansLa populiculture française a vécu depuis 20 ans un nombre important de bouleversements à tous les points de vue : génétique, sanitaire et climatique, environnemental, social et industriel. La révolution génétiqueElle a marqué fortement la filière par des gains de productivité significatifs. L' arrivée des hybrides interaméricains belges (Beaupré, Boelare, Unal, Raspalje) a permis de réduire de 5 à 10 ans la révolution de la peupleraie, en fonction des conditions pédo-climatiques. Avant cette période, la révolution (1) était couramment de l'ordre de 20 à 25 ans, voire plus, alors que maintenant elle se situe entre 15 et 20 ans, selon les conditions, avec une moyenne de l'ordre de 18 ans. Les hybrides interaméricains n'ont pas été les seuls à apporter ce gain de productivité. De nouveaux hybrides euraméricains, aussi bien hollandais (Dorskamp, Koster, etc.) qu'italiens (I 45/51, Triplo, etc.), ont depuis confirmé ces gains de productivité. De nouvelles générations d'hybrides belges, italiens et français, en expérimentation actuellement, entérinent définitivement qu'une révolution de Peuplier, selon les pratiques actuelles, se fait en moins de 20 ans, avec des révolutions de 15 ans qui ne sont plus exceptionnelles. À ces gains de productivité, il convient d'ajouter un gain de forme des arbres (rectitude et branchaison). Les cultivars arrivant à présent sur le marché sont de plus en plus droits et présentent généralement des branchaisons plus faciles à suivre en élagage ou en taille de formation. Enfin, les études récentes sur les qualités du bois des cultivars de Peuplier montrent aussi des performances de plus en plus satisfaisantes du point de vue des qualités mécaniques, de la déroulabilité, et de l'homogénéité de la couleur du bois (Berthelot et al., 2009 ; Berthelot et al., 2013).(1) Une révolution correspond ici à la durée nécessaire pour atteindre une circonférence moyenne à 1,30 m de 140 cm.

No abstract

Poplar (Populus spp.) is a fast-growing tree planted to meet the growing global demand for wood products. In France, the country with the largest area planted with poplar in Europe, accurate and up-to-date maps of its spatial distribution are not available at the national scale. This makes it difficult to estimate the extent and location of the poplar resource and calls for the development of a robust and timely stable approach for mapping large areas in order to ensure efficient monitoring. In this study, we investigate the potential of the Sentinel-2 time series to map the diversity of poplar plantations at the French countrywide scale. By comparing multiple configurations of spectral features based on spectral bands and indices over two years (2017 and 2018), we identify the optimal spectral regions with their respective time periods to distinguish poplar plantations from other deciduous species. We also define a novel poplar detection index (PI) with four variants that combine the best discriminative spectral bands. The results highlight the relevance of SWIR followed by red edge regions, mainly in the growing season, to accurately detect poplar plantations, reflecting the sensitivity of poplar trees to water content throughout their phenological cycle. The best performances with stable results were obtained with the PI2 poplar index combining the B5, B11, and B12 spectral bands. The PI2 index was validated over two years with an average producer’s accuracy of 92% in 2017 and 95% in 2018. This new index was used to produce the national map of poplar plantations in 2018. This study provides an operational approach for monitoring the poplar resource over large areas for forest managers.

Reliable estimates of poplar plantations area are not available at the French national scale due to the unsuitability and low update rate of existing forest databases for this short-rotation species. While supervised classification methods have been shown to be highly accurate in mapping forest cover from remotely sensed images, their performance depends to a great extent on the labelled samples used to build the models. In addition to their high acquisition cost, such samples are often scarce and not fully representative of the variability in class distributions. Consequently, when classification models are applied to large areas with high intra-class variance, they generally yield poor accuracies. In this paper, we propose the use of active learning (AL) to efficiently adapt a classifier trained on a source image to spatially distinct target images with minimal labelling effort and without sacrificing classification performance. The adaptation consists in actively adding to the initial local model, new relevant training samples from other areas, in a cascade that iteratively improves the generalisation capabilities of the classifier, leading to a global model tailored to different areas. This active selection relies on uncertainty sampling to directly focus on the most informative pixels for which the algorithm is the least certain of their class labels. Experiments conducted on Sentinel-2 time series showed that when the same number of training samples was used, active learning outperformed passive learning (random sampling) by up to 5% of overall accuracy and up to 12% of class F-score. In addition, and depending on the class considered, the random sampling required up to 50% more samples to achieve the same performance of an active learning-based model. Moreover, the results demonstrate the suitability of the derived global model to accurately map poplar plantations among other tree species with overall accuracy values up to 14% higher than those obtained with local models. The proposed approach paves the way for national-scale mapping in an operational context.