GeoBridge Seminar Schedule
Guest and Talk Details
February
06.02.2026
Guest 1: Dr. Emiel van Loon, from the University of Amsterdam
Dr. Emiel is an Associate Professor of Statistical Ecology specializing in dynamic and machine-learning models for ecological systems. His research uses GPS tracking, weather radar, and virtual navigation experiments to study animal memory, population dynamics, and large-scale bird and rodent movement.
Talk: How to interpret map correlations?
This presentation (the first from 2 in fact) emphasizes the limitations of measures of Association. These limitations apply to any type of data (spatial as well as non-spatial), however there are a number of special issues with spatial data that are often overlooked, especially effects of clustering and spatial autocorrelation, invisibility of orientation and the inversion of relationships with scale conversions. Fortunately there are ways to remediate these problems and design meaningful measures of association - these will be demonstrated in part 2.
Guest 2: Augusto C. Lima, from the University of Bergen
Augusto is an early-career researcher in global climate change at the Department of Biological Sciences, UiB. His research focuses on cryosphere monitoring in mountain ranges, from reconstructing glacier-climate dynamics in paleo time scales to apply Earth Observation and GIS techniques to access present and future glacier behaviour. His work also emphasize open science, interdisciplinary collaboration and educational outreach, such as the GeoBridge community.
Talk: GeoBridge Overview
GeoBridge is both a community and a collaborative platform for students and researchers in all career stages who are interested in the application of Geographic Information Systems and Remote Sensing to landscape analysis in a changing world. In this presentation, we are going through a brief history of the GeoBridge idealization and concept, our team members and collaborators, our planned events for Spring 2026, our ongoing research projects, and the next steps with GeoBridge. We are also showcasing past and ongoing GIS/RS research from the Mountain in Motions research team, in a way to disseminate UiB research and promote further collaboration between our community.
20.02.2026
Guest 1: Ide Lie-Nilsen, from the University of Bergen
Ida is a PhD candidate in global change ecology at the Department of Biological Science, UiB. Her PhD project will focus on patterns, drivers, and consequences of vegetation change in mountainous ecosystems. She holds a Master of Science in Ecology, with a specialization in ecological resilience. Ida is particularly interested in integrating remote sensing with field-based studies to improve the mapping and understanding of vegetation change in a changing climate.
Talk: Estimating resilience in Norwegian coastal heathland systems
Climate change is driving biodiversity loss and causing shifts in disturbance regimes worldwide. Coastal heathlands, one of Norway’s most vulnerable ecosystems, hold both high cultural and ecological value, yet understanding how they will respond to future changes in climate remains poorly understood. With changing temperature, precipitation and drought patterns forecast for Norway, understanding the vegetation dynamics of coastal heathlands in response to disturbance is essential. Ecological resilience, the capacity of an ecosystem to resist and recover from disturbances, is an important concept for uncovering more about these responses. This study focuses on resilience and recovery rates within three coastal heathland areas over a latitudinal gradient and investigates how temperature and precipitation affect their stability. Using time series of the Normalized Difference Vegetation Index (NDVI) at 250m resolution, resilience is quantified by estimating post-disturbance recovery rates.
March
06.03.2026
Guest 1: Dr. Richard Ott, from the University of Amsterdam
Dr. Richard is an Earth Scientist whose research aims to better understand the interactions and feedbacks between the tectonic, climatic, anthropogenic, and biological processes that shape topography. My research focuses on the evolution of erosional landscapes and how geomorphic archives can be used to reconstruct and predict past and future landscape dynamics.
Talk: High-resolution mapping of anthropogenic erosion in the Northern Andes
Humans have markedly changed erosion and river sediment fluxes around the world, but by how much? We usually lack natural erosion rates to quantify the extent of human influence. In this talk, I show how millennial time-scale cosmogenic nuclide measurements can track natural erosion and compare them with estimates of modern sediment yields. I show that erosion and sediment transport rates in Colombia were 78% higher than natural conditions from 1980 to 2000, and increased to 111% above baseline between 2000 and 2022. Our data document a doubling of erosion in the Northern Andes due to the joint effects of agriculture, mining, and deforestation, however, the erosional response to land use change varies is modified by environmental conditions such as rainfall erosivity. Additionally, I’ll showcase my recently developed course on Landscape Dynamics, where students perform a geospatial analysis of a real landscape under natural and human influences conditions.
20.03.2026
Guest 1: Dr. Yifang Shi, from the University of Amsterdam
Dr. Yifang is a remote sensing scientist who has specific expertise in LiDAR and multi-source remote sensing for ecosystem monitoring. Her current research involves ecosystem structural dynamics monitoring using multi-temporal LiDAR data and integrating multi-source remote sensing (e.g. satellite imagery, hyperspectral, LiDAR and RADAR data) for earth observations, such as tree species mapping, plant functional traits retrieving, and animal habitat monitoring. She is also interested in high performance computing using AI for large amounts of remote sensing data on cloud infrastructures.
Talk: Multi-source remote sensing for ecosystem dynamics monitoring
This talk showcases how multi-source remote sensing can assist our understanding of the changing planet. A wide range of data types, tools, and services will be demonstrated for a variety of remote sensing and ecological applications. I will also touch upon the state-of-the-art development (e.g. drone based remote sensing, virtual research environment, deep learning) in remote sensing domain and how it can be used for real-life ecological questions.
Guest 2: Dr. Johannes De Groeve, from the University of Amsterdam
Dr. Johannes De Groeve is a data manager at the Institute for Biodiversity and Ecosystem Dynamics (IBED) at the University of Amsterdam and part of the Computational Support Unit. He supports researchers in managing, structuring, and analysing ecological and biodiversity data, with a focus on databases, geospatial analysis, and reproducible workflows in R. His work promotes open and standardized data practices that improve transparency, interoperability, and long-term reuse of scientific data. In addition, he develops and teaches workshops on research data management and coding best practices to support more sustainable and reproducible research. He is also the developer of the tabs R package for reconstructing spatial configurations of altitude-bounded biogeographic systems through time.
Talk: Reconstruct the past with TABS (Temporal Altitudinal Biogeographic Shifts)
Environmental change can strongly reshape the spatial configuration of ecosystems. Climate-driven changes in environmental boundaries such as sea level or forest line position can alter the area, connectivity, and isolation of habitats over time, particularly in systems constrained by elevation including island archipelagos, coastal shelves, and alpine environments. In this seminar, I introduce the tabs (Temporal Altitudinal Biogeographic Shifts) R package, an open-source GIS workflow for reconstructing and analysing spatial configurations of altitude-bounded biogeographic systems through time. Tabs enables researchers to model area and connectivity changes under different environmental scenarios, including sea-level fluctuations and forest-line shifts. The workflow is designed to be user-friendly, requiring only a small set of parameters to quickly explore spatial changes across environmental scenarios. The package focuses on standardized spatial outputs in open formats that can be easily integrated with other GIS tools and analytical pipelines. By relying on open-source software and transparent data structures, tabs encourages reproducible research and improves interoperability and reuse of spatial analyses.
April
10.04.2026
Guest 1: Julien Vadnais, from the University of Bergen
Julien is a PhD candidate with a background in Geomatics and Remote Sensing. In his current research, he develops geospatial deep learning algorithms to map marine oil slicks from satellite remote sensing, primarily SAR. Progress on this notoriously challenging task will enable new analyses of hydrocarbon seepage systems which have important environmental implications at the basin scale (e.g. petroleum exploration, geological storage of carbon).
Talk: An end-to-end, open-source workflow for automatic mapping with drones
With all the buzz surrounding new AI methods for geospatial mapping, it may be difficult to settle on learning one approach over another. In this talk, I will present an automatic mapping workflow that is relatively simple to learn yet powerful. Together, we will go through all the steps to train and use a deep learning model for conducting semantic segmentation, i.e., pixel classification, on any sort of aerial images.This workflow, based on the recent work of Adrien Le Guillou (access his github here), uses fully open-source software! You will need the following: 1) Install miniforge (conda-forge | community-driven packaging for conda), 2) Install QGIS (>= 3.16), 3) Install a CUDA driver versions: 12.6; 12.8; or 13.0 (CUDA Toolkit Archive | NVIDIA Developer), and 4) Have a graphic card in your computer (>=8GB of memory recommended).
24.04.2026
Guest 1: Berend-Christiaan Wijers, from the University of Amsterdam
Berend-Christiaan Wijers is a High Performance Computing expert at the Institute for Biodiversity and Ecosystem Dynamics (IBED) and a Research Software Engineer for the Animal Movement Ecology (AME) group at the University of Amsterdam. He holds Bachelor’s and Master’s degrees in Earth Sciences with a specialization in programming and modelling, and has assisted in statistics and programming courses at the University of Amsterdam. His work focuses on compute efficiency, scaling research workflows to continental scales, and developing robust data processing, data organization and e-infrastructure solutions. Building on his experience with radar aeroecology in the cloud, he is developing a Virtual Lab that abstracts much of the complexity of cloud computing, making continental-scale processing of weather radar data more accessible to the broader AME community. He provides computational support within IBED and collaborates with SURF on advanced research computing services.
Talk: Birds, Beams and Big Data: supercomputing a 150 TB Weather Radar Archive
Weather radar archives hold enormous potential for ecology and remote sensing, but the practicalities of processing 150 TB of data can easily stop a project before it starts. In this talk I will walk through our journey from the first ad-hoc scripts to a production workflow that routinely processes multiple terabytes per hour on national supercomputers. Starting with simple R and Python scripts on “whatever storage we could get,” we gradually converged on one scalable process: standardized storage on S3, strict naming and metadata conventions, and a growing layer of automation using tools such as Kubernetes and Argo. The result is a quality-controlled, 150 TB radar archive accessible via APIs, R packages, command line tools, and a Virtual Lab based on Jupyter notebooks. Along the way I will share concrete lessons on team communication, avoiding duplicated work, and why starting early with an imperfect solution is often the fastest route to a robust one.
Guest 2: Brittany La Fae Hatch, from the University of Bergen
Brittany is a subject consultant and teaching assistant at the Department of Geography in the University of Bergen. She has a master’s degree in physical geography, in which she specialized in remote sensing of glaciers. Her research interests are focused on using GIS and remote sensing methods to analyze glacier fluctuations over time.
Talk: Glacier time series analysis using photogrammetry methods
This talk will present the methods used to generate a time series of Blåskavlen glacier in Lofoten, northern Norway. The primary method is photogrammetry, a technique that enables the creation of 3D models from archival aerial photographs. By combining historical imagery with LiDAR and UAV data, it was produced a high-resolution decadal time series spanning 1952 to 2024. This work examines changes in glacier elevation, mass balance, and surface velocity.
May
08.05.2026
Guest 1: Dr. Benjamin Robson, from the University of Bergen
Benjamin Robson in Associate Professor in Geomatics at the Department of Earth Science, University of Bergen. His research is focused on looking at glacial and periglacial changes using remote sensing, in particular combining aerial, satellite, and UAV based datasets.
Talk: Automated analysis of historical aerial photographs to study national scale glacier change in New Zealand
Despite New Zealand having over 2500 glaciers, only two are measured routinely in the field making it hard to get an overview of how the cryosphere is responding to climate change. New Zealand has an extensive archive of historical aerial photographs that allow changes to be tracked back to the 1940s, however processing these datasets is a time consuming and partially manual process. In this talk I will show preliminary results from using AI image matchers to automate processing of entire historical surveys and look at glacier changes. I will go through the workflow we have developed and show some early results for two glaciers in New Zealand - Tasman Glacier, the largest in the country, and Rollerston glacier, one of the two glaciers that is actively monitored in the field.
22.05.2026
Guest 1: Hanif Kawousi, from the University of Bergen
Hanif Kawousi is a PhD candidate in the BetweenTheFjords research group at the Department of Biological Sciences, University of Bergen. With a background in theoretical behavioural ecology, he now works at the interface of applied ecology, local governance and nature accounting. His research focuses on developing methods that support more sustainable, science-based decision-making at the municipal level.
Talk: Ground truthing techniques for species distribution maps
Norway has signed and ratified the Kunming–Montreal Global Biodiversity Framework, committing to protect 30% of nature, restore 30% of degraded ecosystems, and achieve these targets by 2030. As in many other countries, Norway has introduced a national system for nature accounting to support this process. However, responsibility for implementation largely falls to municipalities, as most land is governed at the municipal level.
This creates a core challenge related to resource allocation. Municipalities—often those with the least intellectual and financial capacity—are responsible for managing the most intact and undeveloped areas. As a result, they are caught between the need to develop land for local economic growth and the obligation to follow national directives for ecosystem management and conservation.
One potential solution lies in more locally adapted ecosystem accounts. A key first step is comprehensive mapping of municipal areas. While remote sensing and modelling approaches allow large areas to be covered efficiently, the uncertainty associated with these methods can be a major barrier for decision-makers. Within the ECOBUDGETS project, I have developed a protocol that explicitly accounts for this uncertainty by working across multiple scales. Through soil sampling, vegetation measurements, and assessment of nature types, the protocol aims to (a) reduce uncertainty in map-based tools regarding existing conditions, and (b) enable these tools to communicate ecosystem condition and restoration potential more clearly.
In this GeoBridge Seminar talk, I will focus on the ground-truthing component of this work. I will share experiences from my first field season and discuss how these insights are shaping the second field season, which I am currently preparing.
Guest 2: Sjur Barndon, from the University of Bergen
Sjur was a research assistant at the Mountains in Motion research lab at the department of Biological Sciences. He has have a background in computer engineering, teaching and geoscience. His master’s theses concerned the dynamics of ice sheets over rough landscapes. I am currently interested in both machine learning methods and numerical modelling of ice sheets and glaciers.
Talk: Spatial overlap quantification
For most glaciated areas, detailed mountain glacier evolution since the last interglacial is largely unknown. Due to limitations of traditional numerical modelling, previous studies have typically operated at a coarse spatial resolution, limited study area size, or focused on major climate events. Here we address these limitations by applying the Instructed Glacier Model (IGM), a deep-learning ice-flow emulator enabling efficient GPU-accelerated transient simulations. We model mountain glacier evolution since 130 ka and up to the present-day at 500 m resolution across eight broad mountain ranges in North America, South America, Eurasia, and Africa. Paleoclimate variables are approximated and regionalised using a combination of global and regional climate proxy datasets. We perform 617 parameter-calibration experiments varying paleoclimate and ice-dynamic parameters, with an average runtime of 21 hours per experiment. Model performance is assessed using combined areal and volumetric validation at two known glacial states; the last glacial maximum and the present-day. We also introduce a reproducible probabilistic model-evaluation framework combining confusion matrix validation score and simulation rank to identify sets of acceptable model parameters rather than a single best-fit solution. Our results show that IGM can model realistic ice-flow patterns, glacier geometries, and transient evolution across full glacial-interglacial cycles, demonstrating that machine-learning models of ice dynamics generalise to new domains and conditions, although performance can decline at coarser spatial resolutions. Together, these results demonstrate the feasibility of global-scale, high-resolution, transient glacier modelling over orbital timescales using a deep learning instructed model, while providing a 100-year interval dataset including glacier extent, ice thickness, and ice flow patterns for the last glacial cycle.
June
05.06.2026
Guest 1: Suzette Flantua
Talk:Alpine biogeography: Insights from remote sensing
Guest 2:
Talk:
19.06.2026
Guest 1: Augusto Lima
Talk: GLACIMONTIS: a global geodatabase of Last Glacial Maximum ice extents
Guest 2:
Talk: