Brage NMBU
Not a member yet
19188 research outputs found
Sort by
Parasitic wasp is unlikely to establish and spread in Norway
No full textIt is highly unlikely that Anagyrus vladimiri will be able to establish or spread in Norway. There are no native host organisms, and winter temperatures are too low. Therefore, it is likely that the parasitic wasp will not affect local biodiversity. Thus concludes the Norwegian Scientific Committee for Food and Environment (VKM). Background VKM has assessed the environmental risk of using the product Citripar in Norway. The risk assessment was carried out at the request of the Norwegian Food Safety Authority. Citripar, a product for biological control, is being sought for approval for use in Norway. The product contains the parasitic wasp Anagyrus vladimiri and is intended to be used against mealybugs, especially the species Planococcus citri and P. ficus, on fruits, berries, vegetables, and herbs in greenhouses and plastic tunnels, as well as on indoor plants. Conclusions There have been no reported observations of Anagyrus vladimiri in Norway. VKM assesses that Anagyrus vladimiri will not be able to establish and spread in Norway under current conditions. Anagyrus vladimiri will have no effect on biological diversity in Norway, as there are no known native host organisms that the wasp can parasitize. Individuals of what is now known as Anagyrus vladimiri were for many years identified as belonging to the species Anagyrus pseudococci. Anagyrus pseudococci and A. vladimiri belong to a complex of species that are almost impossible to distinguish from each other and are informally referred to as the Anagyrus pseudococci complex. The risk assessment is approved by VKM's Panel on Plant Health.Parasitic wasp is unlikely to establish and spread in NorwaypublishedVersio
Vindindusert lateral servicebarhet av fotgjengerbroer: En numerisk studie av den dynamiske responsen til forskjellige konstruksjonsformer gjennom parametrisk analyse
No full textEksperimentell modalanalyse og kalibrert numerisk modellert model av Årumfjellet gangbru i
tre avdekket en første vertikal modal frekvens som er utsatt for fotgjengerindusert eksitasjon
og potensielle problemer under forenklede simulert vindlast i henhold til EN 1991-1-4, noe
som nødvendiggjør en grundig vurdering av bruksgrensetilstanden (SLS) i henhold til
eksisterende retningslinjer og standarder for å sikre sikker og komfortabel bruk. Numeriske
parameterstudier viste at modifikasjoner som avstivning og betonglag generelt forbedret
dynamisk respons ved å øke naturlige frekvenser og redusere akselerasjoner, der en
kombinasjon av dekke og bue avstivning pluss betong påstøp viste seg mest effektiv for å
forbedre stivhet og lateral stabilitet, og dermed positivt påvirke kriterier for
bruksgrensetilstanden knyttet til vibrasjonsnivåer og forskyvninger, mens en simulert
ombygging til stålbue var mindre effektiv i å dempe vibrasjoner enn betong på grunn av
dempningsegenskaper, en viktig faktor i bruksgrensetilstanden. Analyse av en virtuell stål
fagverksbro fremhevet den betydelige innflytelsen av opplagerbetingelser på
bruksgrensetilstanden, der fritt opplagrede fagverk viste høyere vertikal stivhet, men større
lateral forskyvning, som potensielt overskrider grenseverdier for lateral bevegelse, noe som
ble forbedret med leddlager, og innføringen av avstivede midtspennsøyler økte stivheten
betydelig og reduserte både akselerasjon og forskyvning, noe som forbedret broen i
bruksgrensetilstanden.Experimental modal analysis and a calibrated numerical model of the Årumfjellet timber arch
pedestrian bridge revealed a first vertical mode frequency susceptible to pedestrian excitation
and potential issues under simplified simulated wind load according to EN 1991-1-4,
necessitating a thorough serviceability assessment according to existing guidelines and
standards to ensure safe and comfortable use. Numerical parametric studies showed that
modifications like bracing and concrete layers generally improved dynamic response by
increasing natural frequencies and reducing accelerations, with a combination of deck and
arch bracing with concrete proving most effective for enhancing stiffness and lateral stability,
thereby positively impacting serviceability criteria related to vibration levels and
displacement, while a simulated steel arch conversion was less effective at mitigating
vibrations than concrete due to damping considerations, an important factor in serviceability.
Analysis of a virtual steel truss bridge highlighted the significant influence of support
conditions on serviceability, where simply supported trusses exhibited higher vertical stiffness
but greater lateral displacement, potentially exceeding serviceability limits for lateral
movement, which was improved by pinned supports, and the introduction of braced mid-span
columns substantially increased stiffness and reduced both acceleration and displacement,
significantly enhancing the bridge's serviceability performanc
Barriers and Opportunities: Muslim Female Entrepreneurs in Norway and the Impact of Systemic Discrimination and Financial Limitations
Get PDFThis thesis explores how systemic barriers, including discrimination, financial exclusion, and religious prohibition, affect Muslim women entrepreneurs in Norway. The study uses a qualitative method to examine how nine women deal with being entrepreneurs while being religious, ethnic and female. It is based on three theories: Becker's theory of discrimination (Becker, 2010), intersectionality theory (Crenshaw, 1989), and Islamic finance theory (El.-Gamal, 2006). These theories help explain how social and institutional processes influence both access to finance and business decision-making.
To strengthen these findings, the analysis is triangulated with existing literature on discrimination and minority entrepreneurship, as well as using multiple theoretical perspectives. The absence of Islamic financial institutions in Norway further complicates matters for the participants. Because of religious beliefs such as the riba, which forbids interest, it is difficult for several people to access traditional loans. Instead, they use community-based saving groups or family networks for informal funding. Even though this option can support the short term, it rarely leads to long-term growth.
The findings align with international trends and reveal context-specific challenges to the Norwegian welfare and financial system. The study provides new information about how Muslim women entrepreneurs in Norway deal with religion and structural limitations. It concludes that policies that support an open financial system and culturally sensitive support systems are essential to ensure that everyone has the chance to be an entrepreneur
Ytre kjennetegn og indre kvistegenskaper til kvalitetsevaluerte stokker av bjørk (Betula pendula Roth og Betula pubescens Ehrh.)
Get PDFDenne masteroppgaven undersøker hvordan ytre kjennetegn, særlig kvistarr, kan brukes til å
forutsi indre kvistegenskaper i bjørk (Betula pendula Roth og Betula pubescens Ehrh.), med
mål om å forbedre sortering og utnyttelse av bjørk som sagtømmer. Til tross for at bjørk utgjør
en betydelig andel av det norske skogvolumet, er den sterkt underutnyttet som virke. En av
hovedutfordringene er at indre kviststruktur reduserer mekaniske egenskaper og visuelle
kvaliteten på trelast, og disse egenskapene blir først synlige etter felling. Studien er derfor
relevant for å utvikle kostnadseffektive, visuelle feltbaserte indikatorer som kan gi informasjon
om treets indre struktur før hogst.
Feltarbeidet ble utført i tre bestander i Sørøst-Norge med ulik bonitet og skjøtselshistorikk.
Totalt ni trær ble valgt ut for dybdestudier. Etter felling ble 6 meter stokkene inndelt i avsnitt
og stammeskiver, der både ytre og indre kvistkjennetegn ble målt. Kvistarrene ble registrert
visuelt og målt med hensyn til høyde, bredde og areal, og klassifisert i fem klasser.
Stammeskiver ble analysert ved radielle snitt for å registrere kvistegenskaper som lengde,
høyde, diameter og kvistens utviklingstrinn fra frisk kvist til okklusjon.
Analysene viste klare forskjeller i kviststruktur mellom bestandene. Trær fra den
lavproduktive, uskøttede bestanden (B63) hadde større kvister, lavere andel ren ved og tregere
okklusjon enn trær fra den høyproduktive og skjøttede bestanden (B65), som viste mindre
kvister, raskere okklusjon og høyere andel kvistfri ved. Dette peker på at skjøtsel og
vekstforhold har stor betydning for treets indre kvalitet.
Ved hjelp av statistiske modeller, inkludert både lineære og blandede regresjonsmodeller, ble
det påvist sterke sammenhenger mellom ytre mål på kvistarr og innvendige kvistegenskaper.
Høyde og areal av kvistarr var spesielt gode prediktorer for kvistens volum, lengde og diameter,
og flere av modellene oppnådde høy forklaringsgrad og lav residualspredning. Modellene
indikerer at enkle målinger av kvistarr i felt kan gi pålitelig informasjon om treets indre
struktur, og dermed forbedre sorteringen til ulike bruksformål.
Til tross for begrenset utvalg gir studien et bidrag til forståelsen av kvistsegenskaper i bjørk
viser hvordan ytre kjennetegn kan brukes som verktøy i praktisk skogbruk. For fremtidig
arbeid anbefales det å validere modellene på større og mer varierte datasett, samt å teste
digitale målemetoder som kan øke nøyaktigheten.This master's thesis investigates how external characteristics, particularly branch scars, can be
used to predict internal branch properties in birch (Betula pendula Roth and Betula pubescens
Ehrh.), with the aim of improving the evaluation of sorting birch as saw timber. Despite birch
comprising a significant portion of Norway’s forest volume, it is severely underutilized as an
industrial resource. One of the main challenges is that internal branch structure reduces the
mechanical traits and visual quality of sawn wood, and these features only become visible after
felling. Therefore, the study is relevant for developing cost-efficient and visually field-based
indicators that can provide information about a tree’s internal structure before harvest.
Fieldwork was conducted in three stands in Southeastern Norway with varying site productivity
and silvicultural history. A total of nine trees were selected for in-depth study. After felling, the
6m logs were divided into sections and stem disks, where both external and internal branch
characteristics were measured. Branch scars were recorded visually and measured in terms of
height, width, and area, and classified into five classes. Stem disks were analyzed through
radial cuts to register branch traits such as length, height, diameter, and the developmental stage
of the branch—from live to occluded.
The analyses revealed differences in branch structure between the stands. Trees from the lowproductive,
unmanaged stand (B63) had larger branches, a lower proportion of clear wood, and
slower occlusion than trees from the high-productive and managed stand (B65), which showed
smaller branches, faster occlusion, and a higher proportion of knot-free wood. This indicates
that silviculture and growing conditions have a significant impact on internal wood quality.
Using statistical models, including both linear and mixed-effects regression models, strong
correlations were found between external measurements of branch scars and internal branch
characteristics. The height and area of branch scars were particularly good predictors of branch
volume, length, and diameter, and several models achieved high explanatory power and low
residual variance. The models suggest that simple measurements of branch scars in the field
can reliably indicate internal tree structure and thus improve sorting for various end-uses.
Despite the limited sample size (nine trees), the study makes an important contribution to
understanding branch development in birch and demonstrates how external traits can be used
as a practical tool in forestry. For future work, it is recommended to validate the models on
larger and more varied datasets, as well as to develop digital measurement methods that can
increase accuracy and operationalize the use of branch scars in the field
Captive Chimps, Captured Curiosity: An Anthrozoogical Study of the Mutual Influences Between Zoo Chimpanzees and Their Human Visitors
No full textZoos are uniquely situated in performing a diverse host of roles, which includes conservation, research, leisure/recreation, and education. In contrast with other wildlife housing institutions, zoos are reliant on visitors, which poses the opportunities for human animal interactions. Chimpanzees (Pan troglodytes spp.) are particularly popular in zoos as they are our closest living relative, exhibit recognisable emotions, and similar behaviour to humans. The overall aim for this study is to shed light on the different ways humans and chimpanzees who interact in zoos influence each other in order to assess how the zoo performs their educational role and how visitor effects may impact animal welfare. The study took place at Kristiansand Zoo, using the group of ten chimpanzees which are housed there and their visitors as study subjects. As different research methods were deemed best suited to collecting and processing the data from the human perspective and the animal perspective, the study is divided into two parts. The effects of visitors on chimpanzee behaviour were measured quantitatively by observation of changes in social behaviour, behavioural states, and enclosure utilisation in response to differing visitor conditions. I hypothesised that increased visitor numbers and/or specific visitor behaviour would induce negative behavioural states such as increased agonistic behaviours and decreased enclosure utilisation. The impact of viewing captive wildlife on learning and emotions in visitors was studied through qualitative interviews with participants to the educational chimpanzee talk and visitors who had interactions with the chimps. The results of part one showed that guests have a significant association with changes in the chimps’ behaviour but there was no clear negative or positive impact. The results of part two found that visitors greatly value viewing chimpanzees live, and that this may notably strengthen guests’ conservation learning. Humans have a great desire to connect with captive wildlife, which zoos can facilitate, but they should ensure that no animal in their care is affected negatively by visitors, whilst they continue to perform their educational duties
Towards Human-Robot Interaction in Agriculture Using Large Language Models
Get PDFSpråklige grensesnitt drevet av store språkmodeller (LLM-er) er i ferd med å endre hvordan mennesker samhandler med autonome systemer, og gjør avansert robotteknologi mer tilgjengelig og intuitiv. I landbruket, hvor arbeidskraftmangel og brukervennlighetsutfordringer hemmer teknologisk adopsjon, utgjør slike grensesnitt en viktig bro mellom automatisering og praktisk feltarbeid.
Denne masteroppgaven undersøker hvordan LLM-er og visjon-språk-modeller (VLM-er) kan integreres i landbruksrobotikk for å gjøre det mulig for ikke-tekniske brukere å gi kommandoer og motta tilbakemeldinger i naturlig, samtalebasert språk. Det ble utviklet et modulært system som tolker skriftlige instruksjoner, planlegger og utfører robotoppgaver, samt benytter visuell input ved behov. Systemet ble testet i en simulert plattform, med eksperimenter som evaluerte kommandotolkning, objektdeteksjon og feltvurdering.
Resultatene viser at LLM-er kan generere forklarbare og kontekstsensitive robotatferder i sanntid, selv med maskinvarebegrensninger og miljømessige utfordringer. Ulike prompt-strategier, Direct, Chain-of-Thought, Few-Shot og Hybrid, ble undersøkt, hvor hybrid prompting ga de mest pålitelige kjøremønstrene, men ikke mest stabile. Videre viste VLM-modulene seg å være effektive i å detektere objekter og analysere visuelle scener, og produserte strukturerte, lettleste feltrapporter.
Ved å forenkle samspillet mellom mennesker og roboter, bidrar denne tilnærmingen til å senke terskelen for bruk av robotikk i landbruket, og legger grunnlaget for brukervennlige, intelligente automatiseringsverktøy som kan øke produktiviteten og fremme bærekraftig jordbruk.Natural language interfaces powered by Large Language Models (LLMs) are redefining how humans interact with autonomous systems, making advanced robotics more accessible and intuitive. In agriculture, where labor shortages and usability barriers hinder technological adoption, these interfaces offer a vital bridge between automation and real-world fieldwork.
This thesis explores the integration of LLMs and Vision-Language Models (VLMs) in agricultural robotics, focusing on enabling non-expert users to command robots and receive feedback through natural, conversational language. A modular system was developed to interpret written instructions, plan and execute robotic tasks, and incorporate visual input when needed. The system was evaluated in a simulated environment, with experiments assessing command interpretation, object detection, and field reporting.
Results show that LLMs can generate explainable, context-aware robot behaviors in real time, even under hardware and environmental constraints. Various prompting strategies, Direct, Chain-of-Thought, Few-Shot, and Hybrid, were tested, with hybrid prompting yielding the most reliable trajectory execution, but not the most consistent. Additionally, VLM-based modules successfully detected objects and analyzed visual scenes, producing structured, human-readable field reports.
By simplifying human-robot interaction, this approach lowers the entry barriers to robotics in agriculture, facilitating the adoption of user-friendly, intelligent automation tools that enhance productivity and promote sustainable farming practices
Uncertainty modeling with Kolmogorov-Arnold Networks and Gaussian Processes
Get PDFThis thesis introduces Gaussian Process Kolmogorov-Arnold Networks (GP-KANs), integrating the compositional structure of KANs with the probabilistic framework of Gaussian Processes. By replacing deterministic activations with GP neurons, we develop a structured deep probabilistic architecture that provides uncertainty quantification throughout the network. We implement and evaluate this approach on diverse landscapes, from synthetic benchmark functions to real-world topographical data. Results demonstrate that GP-KANs effectively model uncertainty in moderately complex scenarios, with uncertainty estimates meaningfully corresponding to prediction errors. Analysis of GP neuron configurations reveals that successful function decomposition can lead to improved prediction accuracy and better-calibrated uncertainty. While our architecture shows limitations with highly complex terrains, these findings establish the viability of combining structured compositional modeling with principled uncertainty quantification—addressing a critical gap in reliable machine learning systems that can acknowledge their own limitations
Risk of developing gastrointestinal side effects in dogs following administration of carprofen, meloxicam, and robenacoxib
Get PDFIn this systematic review, we have examined the risk of developing gastrointestinal side effects in dogs following administration of carprofen, meloxicam, and robenacoxib. We aimed to investigate whether varying degrees of COX-2 selectivity among these drugs could influence the risk. Robenacoxib is a newer NSAID that, in theory, should cause fewer side effects due to its high COX-2 selectivity compared to carprofen and meloxicam. Due to the limited amount of recent data (post 2010), it was difficult to draw any specific conclusions. Nevertheless, a slightly higher risk of lesion development could be associated with the use of meloxicam compared to robenacoxib. The study showed no significant increase in the risk of developing diarrhea or melena between the different treatment groups. Publications from before 2010 already show that carprofen is associated with an increased risk of developing gastrointestinal side effects.I denne litteraturstudien har vi sett på risikoen for utvikling av gastrointestinale bivirkninger hos hund ved administrasjon av karprofen, meloksikam og robenakoksib. Vi ønsket å undersøke om ulik grad av COX-2 selektivitet i de forskjellige legemidlene kunne påvirke risikoen. Robenakoksib er et nyere NSAID, som i teorien skal gi mindre bivirkninger grunnet dens høye COX-2 selektivitet, sammenlignet med karprofen og meloksikam. På bakgrunn av lite datagrunnlag i nyere tid (etter 2010), var det vanskelig å kunne konkludere med noe spesifikt. Likevel kan en noe høyere risiko knyttes til bruk av meloksikam for utvikling av lesjoner enn robenakoksib. Studien viste ingen signifikant økning i risiko for utvikling av diaré eller melena mellom de ulike behandlingsgruppene. Publikasjoner fra før 2010 viser allerede at karprofen er knyttet til økt risiko for utvikling av bivirkninger i mage-tarm.VE
Koblingen mellom overvannshåndtering og naturbaserte løsninger i by- og regionplanlegging
Get PDFSom et resultat av den globale oppvarmingen, øker klimautfordringene i Norge. Dette kommer til utrykk gjennom hyppigere og mer ekstreme nedbørsperioder. Økte regnmengder har ført til stadig flere forekomster av oversvømmelse og flom, særlig i de tettbygde urbane områdene. Dette fenomenet kan tilskrives flere faktorer, men blant de viktigste årsakene er mangel på naturlige overflater og vannveier som kan håndtere store mengder regnvann. I byer og tettsteder blir det naturlige landskapet erstattet med ugjennomtrengelige overflater, som asfalt og betong. Dette hindrer vannet i å infiltrere i bakken, noe som resulterer i økt overvann ettersom regnvann ikke får naturlige utslippspunkter, men i stedet samler seg på overflaten.
I tillegg til mangelen på naturlige vannveier, er økt urbanisering og fortetting av byens infrastruktur sentrale faktorer som forsterker disse utfordringene. Byutvikling og utbygging fører til en økning i tetthet og reduksjon av grøntområder, noe som reduserer områdets evne til å håndtere regnvann på en bærekraftig måte. Dette skaper et dilemma i by- og regionplanlegging: Hvordan imøtekomme behovet for byutvikling og samtidig bevare naturens funksjoner?
Naturbaserte løsninger viser til hvordan bærekraftige prinsipper kan integreres i urbane miljøer. Konseptet tar sikte på å reetablere og fremme naturlige prosesser, slik som vannets naturlige infiltrasjon og habitat for biologisk mangfold. Ved hjelp av tiltak som gjenoppretting av elver, bekker og våtmarker, etablering av grønne tak, regnbed, og sammenhengende grønnstruktur, kan overvann håndteres på en mer effektiv og helhetlig måte. Naturbaserte løsninger tar utgangspunkt i naturlige prosesser, og bidrar ikke bare til å redusere flomrisiko, men forbedrer også livskvaliteten for innbyggerne ved å skape grønnere og mer attraktive nabolag.
Målet med denne oppgaven er å bidra til økt kunnskap og synliggjøre potensialet for overvannshåndtering gjennom etablering av grønnstruktur og naturbaserte løsninger i by- og regionplanlegging. I den forbindelse tar oppgaven utgangspunkt i å undersøke hvordan norske kommuner i dag integrerer bærekraftige prinsipper, som naturbaserte løsninger og overvannshåndtering i planleggingsprosessen. Fokuset rettes mot hvordan løsningene implementeres i tidlige faser av planleggingen, samt hvilke konkrete tiltak som har blitt realisert på et regionalt og kommunalt nivå.As a result of global warming, climate challenges are increasing in Norway. This is expressed through more frequent and extreme periods of precipitation. Increased rainfall has led to a growing number of incidents of flooding, particularly in densely built urban areas. This phenomenon can be attributed to several factors, but among the most significant causes is the lack of natural surfaces and waterways capable of managing large volumes of rainwater. In urban planning, natural landscapes are often replaced by impermeable surfaces such as asphalt and concrete. This prevents water from infiltrating the ground, resulting in increased surface runoff, as rainwater lacks natural discharge points and instead accumulates on the surface.
In addition to the lack of natural waterways, increasing urbanization and densification of city infrastructure are key factors that exacerbate these challenges. Urban development and construction lead to greater density and a reduction in green areas, which in turn diminishes the area's ability to sustainably manage rainwater. This creates a dilemma in urban and regional planning: How to encounter the need for urban development while also preserving nature’s functions?
Nature-based solutions refer to how sustainable principles can be integrated into urban environments. The concept aims to reestablish and promote natural processes, such as water infiltration and habitats for biodiversity. Through measures like the restoration of rivers, streams, and wetlands, the implementation of green roofs, rain gardens, and connected green infrastructure, stormwater can be managed in a more effective and holistic way. Nature-based solutions are grounded in natural processes and not only help reduce flood risk but also improve residents' quality of life by creating greener and more attractive neighbourhoods.
The goal of this thesis is to contribute to increased knowledge and highlight the potential for stormwater management through the establishment of green infrastructure and nature-based solutions in urban and regional planning. In this regard, the thesis aims to examine how Norwegian municipalities currently integrate sustainable principles, such as nature-based solutions and stormwater management, into the planning process. The focus is on how these solutions are implemented in the early phases of planning, as well as which specific measures have been realized at the regional and municipal levels
Illegal fishing and human rights violations in Taiwan : examining the role of international fisheries governance
Get PDFThe scale of Illegal, Unreported, and Unregulated (IUU) fishing and the extent of human rights violations in Taiwan’s distant water fishing (DWF) sector call for stronger governance structures with environmental and human rights guarantees. Based on a secondary literature review, this thesis argues that IUU fishing is inherently linked to forced labour and human trafficking and that current governance structures are unable to deal with these problems. The thesis clarifies that Taiwan’s uncertain international status, flags of convenience, and loose labour regulation enable systematic exploitation. It further clarifies how reform is shaped by global pressure, corporate citizenship, and decentralized regulation. The argument hinges on the fact that migrant workers, particularly from Southeast Asia, are at high risk through lawful exclusions and poor working conditions. The study claims that the steps in governance must be more binding, transparent, and participatory, not just an overload of sustainability discourse and enforcement of workers’ rights, particularly on the part of poor workers engaged in offshore and high-seas settings.B-D