Publikasjoner
NIBIOs ansatte publiserer flere hundre vitenskapelige artikler og forskningsrapporter hvert år. Her finner du referanser og lenker til publikasjoner og andre forsknings- og formidlingsaktiviteter. Samlingen oppdateres løpende med både nytt og historisk materiale. For mer informasjon om NIBIOs publikasjoner, besøk NIBIOs bibliotek.
2025
Sammendrag
The European Union Deforestation Regulation (EUDR) mandates traceability of timber that makes up wood products from its harvest site to the end product to ensure sustainable wood sourcing. This study proposes a cost-effective, image-based method for tracing logs using alphabetic codes printed onto logs at the harvest site. These codes are detected and interpreted through a two-stage system leveraging deep learning models. The detection stage employs YOLOv8 to locate tracking codes in images of log piles. It is trained and evaluated on a dataset of 125 images, achieving an F1-score of 0.811 on unseen images. The recognition stage, trained on 1,020 images, uses YOLOv8 models to detect individual characters and their positions within each code. On a set of unseen images, the interpretation stage is able to identify 92.8% of the individual logs despite the limited quality of the printer and degradation of the codes due to stem wetness. Analysis indicates that errors predominantly arise in the character detection step. Compared to existing traceability approaches, this method is more cost-effective than RFID tags and attains higher accuracy than image-based biomarker tracking methods.
Forfattere
João Carlos de Moraes Sá Rattan Lal Klaus Lorenz Yadunath Bajgai Carla Gavilan Ademir De Oliveira Ferreira Clever Briedis Thiago Inagaki Daniel Ruiz Potma Gonçalves JeanKleber BortoluzziSammendrag
No-till systems (NTS) predicated on the tenets of conservation agriculture principles are a viable agricultural paradigm to achieve net zero or net negative emissions. We assessed the carbon dioxide equivalent (CO₂e) emissions based on soil organic carbon (SOC) stock changes in 1-m depth by plow-based tillage (PBT) and the mitigation potential through a no-till system (NTS) across 26 sites in the Cerrado biome and 37 sites in the Atlantic Forest biome. These sites comprise 86,411 ha (ha), encompassing four climate zones in Brazil. The investigation revealed a range of CO2e emissions, with the lowest recorded value of 74.2 Mg CO2e ha−1 observed in the tropical equatorial climate zone and the highest recorded value of 470.1 Mg CO2e ha−1 detected in the subtropical humid climate zone. The total CO2e emissions in the tropical equatorial, tropical central, subtropical humid and subtropical temperate climate zones were calculated to be 5.51, 3.88, 3.21, and 4.20 Tg CO2e, respectively, with a cumulative value of 16.80 Tg CO2e with 6.7 % of uncertainty (i.e., 1.12 Tg CO2e). Adoption of NTS demonstrated a high capacity for offsetting CO2 emissions, achieving 5.40 Tg CO2e in the tropical equatorial zone (recovering 98 % of the total emissions), 2.57 Tg CO2e in the tropical central zone (68.7 %), 2.67 Tg CO2e in the subtropical humid zone (83.2 %), and 2.88 Tg CO2e in the subtropical temperate zone (68.6 %). The percentage of net zero and net negative emissions contributed by the SOC stock for 1-m depth was 73.63 % and 26.37 %, respectively, and it played a pivotal role in integrating agriculture as a part of the climate solution.
Forfattere
Tomáš Hlásny Roman Modlinger Jostein Gohli Rupert Seidl Paal Krokene Iris Bernardinelli Simon Blaser Gediminas Brazaitis Gailenė Brazaitytė Eckehard G. Brockerhoff György Csóka Laura Dobor Maarten de Groot Mihai‐Leonard Duduman Massimo Faccoli Margarita Georgieva Georgi Georgiev Wojciech Grodzki Henrik Hartmann Anikó Hirka Gernot Hoch Tomasz Jabłoński Hervé Jactel Mats Jonsell Marija Kolšek Markus Melin Slobodan Milanović Constantin Nețoiu Mats Nieberg Bjørn Økland Milan Pernek Michaela Perunová Nick Schafstall Martin Schroeder Gottfried Steyrer Jozef Vakula Thomas Wohlgemuth Tiina Ylioja Andrew M. LiebholdSammendrag
Ongoing shifts in climate and land use have altered interactions between trees and insect herbivores, changing biotic disturbance regimes. However, as these changes are complex and vary across host species, insect taxa, and feeding guilds, they remain poorly understood. We compiled annual records of forest insect disturbance from 15 countries in temperate and boreal Europe, spanning the period from 2000 to 2022. The dataset comprises 1361 time series characterizing the dynamics of 50 herbivorous insects. We used this dataset to test whether insect disturbance has systematically changed during the 23‐year period across host trees and feeding guilds, whether it varies along latitudinal and climatic gradients, and whether synchrony exists among species in the same guild or among species sharing the same host. Since 2000, borer disturbance was predominantly concentrated on gymnosperms, while defoliators impacted gymnosperms and angiosperms more evenly. While 85.8% of gymnosperm disturbance was inflicted by a single species, Ips typographus , the majority of disturbances to angiosperms were caused by six different species. Borer impact on gymnosperms has increased in the 21st century, while defoliator impact has decreased across both clades. In contrast to diverging temporal trends, disturbance was consistently greater in warmer and drier conditions across feeding guilds and host types. We identified significant synchrony in insect disturbance within host types and feeding guilds but not between these groups, suggesting shared drivers within guilds and host types. Increasing insect disturbance to gymnosperms may catalyze adaptive transformations in Europe's forests, promoting a shift from historical conifer‐dominated management to broadleaved trees, which are less affected by insect herbivores. Our findings reveal a diversity of trends in insect herbivory, underscoring the need to strengthen monitoring and research in order to better understand underlying mechanisms and identify emerging threats that may not be apparent in currently available data.
Forfattere
Even UnsgårdSammendrag
Det er ikke registrert sammendrag
Forfattere
Trygve S. AamlidSammendrag
Det er ikke registrert sammendrag
Sammendrag
Det er ikke registrert sammendrag
Forfattere
Trygve S. AamlidSammendrag
Det er ikke registrert sammendrag
Sammendrag
Det er ikke registrert sammendrag
Sammendrag
Det er ikke registrert sammendrag
Forfattere
Jing Zhou Qianyi Duan Nicole AndersonSammendrag
Accurate classification of grass seed crop species is essential for estimating seasonal field acreages, informing market strategies, promoting crop diversification, and establishing long-term cropping histories. Unlike major commodity crops, grass seed crops lack reliable datasets and mapping products. This study investigates the use of spaceborne imagery and artificial intelligence (AI)-driven computer vision to remotely classify grass seed crops. Our ground observation dataset comprises 15 grass seed species grown in Oregon, USA (2021-2023), covering over 4,000 data points. Satellite imagery was acquired from Sentinel-2 (S2) spanning January 1st to June 14th each study year. The imagery includes 12 bands across 400-2190 nm with a spatial resolution of 10 m pixel-1, collected at five-day intervals, totalling 34 time stamps. Statistical analyses identified the second and third weeks of May as the most critical temporal window for spectrally distinguishing among grass species using satellite imagery, coinciding with field inspection timing for crop purity. The near-infrared [835.1 nm (S2A) / 833 nm (S2B)], red edge [740.2 nm (S2A) / 739.1 nm (S2B)], and narrow near-infrared [864.8 nm (S2A) / 864 nm (S2B)] bands showed the highest spectral separability among major grass species. A U-Net Temporal Attention Encoder (U-TAE) model was trained to classify grass seed crop species, integrating temporal and spectral data. The overall classification accuracy - defined as the ratio of correctly classified samples to total samples - was 0.89 across all 15 grass species with high accuracies for four major species, including tall fescue (0.93) (Schedonorus arundinaceus (Shreb.) Dumort.), perennial ryegrass (0.90) (Lolium perenne L.), annual ryegrass (0.87) (Lolium perenne L. ssp. Multiflorum (Lam.) Husnot), and Kentucky bluegrass (0.83) (Poa pratensis L.) (0.83). Our findings provide actionable insights for industry stakeholders, enabling informed pricing, planting strategies, and reduced risk of cross-pollination. This work highlights the potential of AI and remote sensing in grass seed crop production, with future efforts focused at estimating field acreage and predicting production potential.