Spatiotemporal Distribution Modeling of Cod in Norway Through a Diffusion Model

Blok, Ian (2025) Spatiotemporal Distribution Modeling of Cod in Norway Through a Diffusion Model. Bachelor's Thesis, Artificial Intelligence.

Preview	Text bAI2025IanBlok.pdf Download (2MB) | Preview
	Text Toestemming.pdf Restricted to Registered users only Download (247kB)

Abstract

Cod is a vital species on earth and for Norway. Following the collapse of the Norwegian Spawning Herring, Norway has introduced measures to promote sustainable fishing. Yet these measures do not suffice. Just four out of the fifteen cod populations have a healthy population size. Furthermore, the fishing industry has a significant impact on the climate, having produced 179 million tonnes CO2 worldwide in 2011. Modeling the distribution of cod could reduce these issues, allowing for more efficient and sustainable fishing. Through a diffusion model, this study tries to forecast catch locations of the next day. The model is compared against a baseline (based on a previous paper) that predicts catch data for a given day by summing the catch data from a 15-day window (7 days before, the current day, and 7 days after) from the previous three years. The study shows that the model outperforms the baseline, generating predictions that are closer to the actual catch locations. The model's predictions outperformed the baseline, with shorter mean distances to true catch locations and lower MSE and MAE scores. The findings reveal that catch and environmental data alone cannot reliably predict cod distribution patterns, especially due to the limited temporal resolution and sparse datasets. However, the study also shows the learning capacity of diffusion models despite high sparsity in data, suggesting that given high quality data, the model is able to make more accurate predictions.

Item Type:	Thesis (Bachelor's Thesis)
Supervisor name:	Cardenas Cartagena, J. D.
Degree programme:	Artificial Intelligence
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	29 Jul 2025 05:41
Last Modified:	29 Jul 2025 05:41
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/36572

Actions (login required)

View Item