DISA

When? Friday August 16th 14-16
Where? Onsite: B1009 at Linnaeus University in Växjö and online
Registration: Please sign up for the PhD-seminar via this link https://forms.gle/aYqRMod68hVLv8EW9 by August 14th (especially important if you plan on attending onsite so we have fika for everyone)

Agenda

14.00-14.10 Welcome and practical information from Welf Löwe
14.10-14.55 Presentation and discussion: Improving Non-Indigenous Species Introduction Risk Considering Seasonality and Gravity-informed Deep Learning Models – Amilcar Soares
14.55 – 15.05 Coffee break
15.05 – 15.50 Presentation and discussion – State-of-the-art and ongoing research on the Visualization of Temporal and Multivariate Networks – Claudio Linhares
15.50 -16.00 Sum up and plan for the September seminar

Abstracts

Improving Non-Indigenous Species Introduction Risk Considering Seasonality and Gravity-informed Deep Learning Models – Amilcar Soares

The introduction and spread of aquatic non-indigenous species (NIS) pose significant threats to global biodiversity, disrupt ecosystems, and cause substantial economic damage in agriculture, forestry, and fisheries. The growing complexity of international trade and transportation networks has exacerbated the risk of NIS introduction and spread. In this presentation, I will discuss the common problem of NIS management and the importance of robust risk assessment models to mitigate these threats. First, I will present a study investigating the influence of temporal variability in sea surface temperature and salinity on ballast water risk assessment (BWRA) models. By comparing global ports’ monthly and annual environmental data, the study highlights how seasonal variations can impact the environmental similarity scores between source and recipient locations, which are crucial for predicting NIS survival and establishment. The findings suggest that incorporating monthly data in BWRA models provides a more sensitive and accurate risk assessment than traditional annual average models. Next, I will introduce a novel physics-informed model designed to forecast maritime shipping traffic and assess the risk of NIS spread through global transportation networks. This model, inspired by the gravity model for international trade, integrates factors such as shipping flux density, port distance, trade flow, and centrality measures. By incorporating transformers, the model effectively captures both short- and long-term dependencies, achieving significant improvements in predicting vessel trajectories and traffic flows. The enhanced accuracy of this model aids policymakers and stakeholders in identifying high-risk invasion pathways and prioritizing management actions. Together, these studies advance our understanding of NIS risk assessment and underscore the need for dynamic, data-driven approaches to effectively manage and mitigate NIS’s impacts in a rapidly changing global landscape.

State-of-the-art and ongoing research on the Visualization of Temporal and Multivariate Networks – Claudio Linhares
This presentation will cover an overview of current research on visualizing temporal and multivariate networks, emphasizing the challenges and advancements in representing evolving interactions and diverse attributes. Also, it will discuss ongoing research into temporal network visualization techniques, including static and dynamic approaches, and the incorporation of multivariate attributes, such as node features, edge weights, and temporal dynamics. Furthermore, it will explore the challenges of scalability, interpretability, and interactive exploration.