Eivind Handegard
Forsker
Biografi
Min forskning dreier seg rundt Miljøregistrering i Skog og det å bruke bevaringsbiologi for å gjøre det mulig for forvaltere å ta informerte beslutninger. Derfor innebærer en betydelig del av arbeidet mitt å ta opp anvendte økologiske utfordringer. Jeg er spesielt interessert hvordan tid påvirker beta-diversitet, både i kulturskogen og i den eldre skogen. Arbeidet mitt har vært rettet mot betydningen av skogens alder og metoder for å identifisere gamle trær.
Forfattere
Stefano Puliti Binbin Xiang Maciej Wielgosz Eivind Handegard Nicolas Cattaneo Marta Vergarechea Terje Gobakken Juha Hyyppä Erik Næsset Mikko Vastaranta Tuomas Yrttimaa Rasmus AstrupSammendrag
Accurately determining the age of individual trees is important for understanding forest dynamics, tree growth, site productivity and describing ecological processes. Traditional methods, such as dendrochronological coring, are invasive, labor-intensive, and costly. This study investigates the use of deep learning (DL) to predict tree age from high-density laser scanning data as a scalable, non-invasive alternative. The dataset includes approximately 1700 tree point clouds from approx. 1 K trees across Norway, Sweden, and Finland, encompassing Norway spruce (Picea abies) and Scots pine (Pinus sylvestris) and a broad range of tree age and developmental stages, from young seedlings (1 year) to old trees (∼350 years). Data were collected using terrestrial, mobile, and high-density airborne laser scanning platforms, enabling the development of sensor-agnostic models. We evaluated multiple modelling approaches, from linear regression to transformer architectures, using both training-from-scratch and fine-tuning strategies. Models fine-tuned starting from pre-trained weights from ForestFormer3D's U-Net as well as the transformer architecture (PointTransformerV3) trained from scratch, proved effective for age regression (RMSE ≤23 years). Although our analysis was limited to two tree species, we demonstrated that a single joint age-estimation model can be successfully trained for both species. We demonstrate that models trained on high-resolution data can generalize to lower-resolution, less costly inputs, provided that data augmentations that mimic reduced resolutions are included during training. This study presents a data-driven framework for estimating tree age without destructive sampling. The findings support the potential for AI-based methods to complement or replace traditional age estimation techniques in forest inventory and monitoring.
Sammendrag
1. The results of nature restoration efforts have been characterized as notoriously unpredictable. Many variables impact the trajectory of species communities towards recovery, and ecological theory that takes traits, habitat configuration and scale into account, can improve models. However, the most important questions regarding the predictability of species community restoration may be related to stochasticity. 2. We investigated the assembly of a cyanolichen community in a chronosequence consisting of 88 new forest patches (30–140+ years old) comprising today 0.4% of a 170 km2 former treeless heathland area in south-western Norway. Two complete inventories were carried out 12 years apart, and we (1) tested inferences on colonization status and recovery time based on the first inventory only; (2) investigated the recovery of the lichen community by changes in species richness, species density and composition at three different spatial scales; and (3) discussed how dispersal capacity and stochasticity affect community recovery in general. 3. Colonization of sites by lichen species exceeded extinctions in young sites but not in old sites, and in the second inventory, the richness of species weighed by occurrences no longer differed significantly between young and old sites at landscape scale. However, the differences between old and young sites depended on the spatial scale and method of measurement. 4. In accordance with inferences based only on the first inventory, colonization and extinction dynamics indicated that recovery of species richness in our study system will take 90–120 years at the landscape scale, whereas recovery of species composition was difficult to determine due to idiosyncratic development among sites. 5. Synthesis and applications. Using species composition as a template for the evaluation of restoration recovery in systems with a high degree of stochastic colonization and extinction is problematic, particularly at finer scales. Ideally, comparisons of restoration and reference communities should therefore be at large enough spatial scale to cancel out the major effects of stochasticity at finer scales. Furthermore, we suggest that a complete recovery of species numbers may not be needed as an indicator of restoration success if species richness measurements indicate that communities are en route to recovery.
Sammendrag
Det er ikke registrert sammendrag