Predictive mapping of seafloor biomass reveals resource hotspots and cold zones

Why It Matters

Understanding the distribution and abundance of seafloor biomass is crucial for managing marine ecosystems and resources. These predictive models provide a powerful tool for identifying critical habitats, assessing the impact of environmental changes, and informing sustainable resource management strategies.

Key Finding

Global seafloor biomass is highest in productive coastal and polar areas and lowest in the deep ocean, driven by food availability. Machine learning effectively predicts these patterns.

Key Findings

• Predictive models explained 63% to 88% of the variance in seafloor biomass for major size groups.
• Seafloor biomass is positively correlated with surface primary production and the flux of particulate organic carbon to the seafloor.
• Biomass is highest in polar regions, on continental margins with coastal upwelling, and in equatorial divergence zones.
• Lowest biomass is found on abyssal plains.
• The shift in biomass dominance with depth is influenced by a decrease in average body size, likely due to reduced food quantity and quality.

Research Evidence

Aim: To model and predict global seafloor biomass and abundance using environmental variables and to generate maps illustrating these patterns.

Method: Predictive modelling using a machine learning algorithm (Random Forests).

Procedure: A comprehensive database of seafloor biomass and abundance was compiled from global oceanographic institutions. A Random Forests model was trained using surface primary production, particulate organic matter flux, seafloor relief, and bottom water properties to predict biomass for different size groups (bacteria, meiofauna, macrofauna, megafauna). Global maps of predicted biomass and abundance were then generated.

Context: Marine ecology and oceanography, specifically deep-sea ecosystems.

Design Principle

Predictive spatial modelling of biological resources based on environmental drivers can optimize resource management and conservation efforts.

How to Apply

Use similar machine learning approaches with relevant environmental data to predict the distribution of other biological or mineral resources in various ecosystems.

Limitations

Model accuracy may vary in areas with limited data. Predictions are based on correlations and may not fully capture complex ecological interactions.

Student Guide (IB Design Technology)

Simple Explanation: Scientists used a computer program to guess where the most life is on the ocean floor by looking at things like how much food is available from the surface. They found that the ocean floor is richest near coasts and poles, and emptiest in the deep ocean plains. This helps us understand where to find and protect ocean life.

Why This Matters: This research shows how we can use data and technology to understand and map where valuable natural resources are located, which is important for planning how we use them wisely.

Critical Thinking: How might the accuracy of these predictions be affected by unforeseen ecological shifts or the introduction of invasive species not accounted for in the training data?

IA-Ready Paragraph: This research highlights the utility of machine learning algorithms, such as Random Forests, in predicting the spatial distribution of biological resources. By correlating environmental factors like primary production and organic matter flux with seafloor biomass, the study generated predictive maps that identified resource hotspots and cold zones, offering a robust framework for understanding and managing marine ecosystems.

Project Tips

• Consider using publicly available environmental datasets (e.g., satellite data for primary production) to model the distribution of a resource in a specific area.
• Explore different machine learning algorithms for predictive modelling in your design project.

How to Use in IA

• This study demonstrates the application of predictive modelling for resource assessment, which can be a valuable reference for projects involving spatial analysis and resource mapping.

Examiner Tips

• When discussing predictive models, consider the potential biases introduced by the training data and the limitations of extrapolating predictions to new areas.

Independent Variable: ["Surface primary production","Water-column integrated and export particulate organic matter (POM)","Seafloor relief","Bottom water properties"]

Dependent Variable: ["Seafloor biomass","Seafloor abundance"]

Controlled Variables: ["Major size groups (bacteria, meiofauna, macrofauna, megafauna)"]

Strengths

• Utilizes a large, globally compiled database.
• Employs a robust machine learning algorithm for prediction.
• Generates comprehensive global maps of seafloor biomass.

Critical Questions

• What are the ethical considerations when using predictive models for resource management, especially concerning potential over-exploitation of identified hotspots?
• How can these predictive models be integrated with real-time monitoring data to improve their accuracy and responsiveness to environmental changes?

Extended Essay Application

• An Extended Essay could investigate the application of similar predictive modelling techniques to map the distribution of a specific natural resource (e.g., timber, minerals, agricultural potential) in a chosen region, analyzing the environmental factors that influence its availability.

Source

Global Patterns and Predictions of Seafloor Biomass Using Random Forests · PLoS ONE · 2010 · 10.1371/journal.pone.0015323