Predicting Shark Habitats Based on Environmental Conditions Using Machine Learning

Overview

Increasing carbon emissions is resulting in global warming, melting of polar ice, and slowing of ocean currents, putting sharks at risk. Sharks are a keystone species that serve as an indicator of ocean health. Decreases in shark population can hurt coral reefs, seagrass beds, and even local fisheries. Unfortunately almost a third of shark species are threatened or endangered. The goal of this project is to use machine learning to predict shark habitats, based on changes to the environmental conditions. This study used satellite telemetry data of Shortfin Mako sharks in the Atlantic. In order to train the machine learning models, raw positional data was used to find shark habitats and the associated environmental conditions. Using move persistence modeling and K-means clustering, the data were classified into habitats and non-habitats. Then, data on sea surface temperature, salinity, distance to coast, depth, and chlorophyll levels was matched to each location. This was used to train both decision trees and random forest machine learning models to predict shark habitats based on environmental conditions with accuracies of 80% in method 1 and accuracies of 97% and 98% in method 2. Simulated environmental data showed that many habitats shift when environmental conditions change. This showed how climate change will continue to affect shark habitats. This information is vital for conservationists to protect the endangered species and our oceans by allowing them to pass policies and raise funds for conservation efforts.

Competition history

• AJAS 2025 Category not listed

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Earth and Environmental Sciences projects
• AJAS Earth and Environmental Sciences projects
• Every AJAS 2025 project