Nazanin Shamshiri

Topic

A Visualization Design Study for Hydrodynamic Connectivity Analysis in Marine Aquaculture

Department of Computer Science

Date & location

• Wednesday, April 29, 2026

• 9:15 A.M.

• Engineering Computer Science Building

• Room 468

Reviewers

Supervisory Committee

• Dr. Sean Chester, Department of Computer Science, University of Victoria (Supervisor)

• Dr. Miguel Nacenta, Department of Computer Science, UVic (Member) 

External Examiner

• Dr. Lin Cai, Department of Electrical and Computer Engineering, University of Victoria 

Chair of Oral Examination

• Dr. Scott Watson, Department of Political Science, UVic

   

Abstract

Hydrodynamic particle tracking simulations are used in the Fisheries and Oceans Canada study of connectivity between marine finfish aquaculture sites around Vancouver Island [1]. In this context, simulation outputs are summarized as connectivity matrices showing the percentage of particles released from one farm that arrive at another within selected time periods. Although these matrices are the main reporting artifact, domain experts often interpret them together with spatial context, temporal comparisons, and directional relationships. This makes interpretation, exploration, and communication difficult, especially when results must support discussion with non-specialist stakeholders.

This thesis presents a visualization design study conducted in collaboration with ocean scientists working on this aquaculture connectivity application. Through con textual inquiry, affinity-based task characterization, participatory co-design, low- and high-fidelity prototyping, and qualitative evaluation, the study investigates how interactive visualization can better support the interpretation and communication of farm-to-farm connectivity matrices derived from the particle simulations. As part of this process, the study identifies and stabilizes design requirements that capture key conditions for supporting these interpretation and communication tasks.

The resulting prototype combines an interactive connectivity map, a matrix view, temporal exploration controls, threshold filtering, and exploratory clustering. In connectivity exploration mode, coordination was achieved mainly through shared temporal and threshold controls that updated values and visual encodings across views. Stronger selection-based coordination was achieved in clustering mode. Evaluation findings suggest that the prototype supported spatial interpretation of farm-to-farm connectivity, plausibility checking against expected regional flow patterns, and exploratory comparison of candidate farm groupings, while also showing that familiar matrix-based representations remained important for user trust and verification.

Beyond the prototype, the thesis contributes design knowledge for interactive visualization of hydrodynamic connectivity matrices in the aquaculture management context, including the value of combining spatial and matrix-based views, the importance of preserving familiar representations while introducing new analytical views, the need to frame clustering as an exploratory rather than authoritative result, and the importance of aligning interface terminology with the abstraction of aggregated connectivity data.