Joint Modeling of Zooplankton Abundance and North Atlantic Right Whale Distribution in Cape Cod Bay using Data Fusion

Core Concepts

Jointly modeling zooplankton abundance and North Atlantic right whale distribution, by fusing multiple data sources for each species, enhances the accuracy and understanding of their relationship in a dynamic marine environment.

Abstract

• Bibliographic Information: Kang, B., Schliep, E. M., Gelfand, A. E., Clark, C. W., Hudak, C. A., Mayo, C. A., ... & Schick, R. S. (2024). Joint Spatiotemporal Modeling of Zooplankton and Whale Abundance in a Dynamic Marine Environment. arXiv preprint arXiv:2411.06001.
• Research Objective: This study aims to develop a novel joint species distribution model (JSDM) to investigate the relationship between North Atlantic right whale (NARW) abundance and the distribution of their zooplankton prey in Cape Cod Bay, Massachusetts.
• Methodology: The researchers developed a Bayesian hierarchical JSDM that fuses four data sources: oblique and surface tow data for zooplankton abundance, and aerial distance sampling and passive acoustic monitoring data for NARW abundance. The model incorporates environmental covariates such as sea surface temperature and bathymetry. The researchers first conducted a simulation study to evaluate the model's performance under various sampling scenarios and then applied the model to real data collected over six days.
• Key Findings: The simulation study demonstrated that the joint model outperforms models using only zooplankton or whale data, providing more accurate abundance estimates and insights into the relationship between the species. The real-data application revealed a positive relationship between NARW abundance and zooplankton concentration.
• Main Conclusions: Jointly modeling species distributions using data fusion from multiple sources can significantly improve the accuracy of abundance estimates and provide valuable insights into the ecological relationships between species. This approach is particularly relevant for studying endangered species like NARWs, where understanding their prey dynamics is crucial for conservation efforts.
• Significance: This research contributes to the field of JSDM by introducing a novel approach that combines geostatistical and point pattern models to link the abundance of a species to the intensity of its prey. This methodology can be applied to other ecological studies where understanding the interplay between species and their environment is critical.
• Limitations and Future Research: The study acknowledges limitations due to the limited spatial and temporal coverage of the data. Future research could explore incorporating additional data sources, such as high-resolution tag data for whales, to improve model accuracy and ecological inference. Further investigation into the influence of environmental factors on the species relationship is also warranted.

Stats

North Atlantic right whale population numbers approximately 372 individuals.
Zooplankton data was collected using oblique and surface tow methods.
Whale data was collected using aerial distance sampling and passive acoustic monitoring.
Oblique tow samples are expected to have higher zooplankton abundance per m3 than surface tows.
The probability of detecting a whale from aerial surveys decays exponentially with distance beyond 0.75km.
The probability of detecting a whale call through passive acoustic monitoring depends on ambient noise levels, distance to the hydrophone, and the whale's call source level.

Quotes

"Thus, it is necessary to concurrently study the distribution of NARW and their prey, because their habitats are changing spatially and temporally and, due to this change, unexpected arrivals of NARW in places like the Gulf of Saint Lawrence have resulted in high mortality of NARW."
"The novelty of our approach is in the joint modeling at the process level of the two latent processes, one capturing zooplankton abundance, the other capturing the spatial point process intensity of NARW."

Key Insights Distilled From

Joint Spatiotemporal Modeling of Zooplankton and Whale Abundance in a Dynamic Marine Environment

by Bokgyeong Ka... at arxiv.org 11-12-2024

https://arxiv.org/pdf/2411.06001.pdf

Deeper Inquiries

How might this joint modeling approach be adapted to incorporate real-time data, such as oceanographic conditions or ship traffic, to improve dynamic management and conservation strategies for North Atlantic right whales?

This joint modeling approach offers a strong foundation for integrating real-time data and enhancing dynamic management strategies for North Atlantic right whales (NARWs). Here's how:
1. Near Real-Time Environmental Data Integration:

Oceanographic Data: Real-time or near real-time oceanographic data, such as sea surface temperature (SST), chlorophyll-a concentration (a proxy for phytoplankton biomass), and ocean currents, can be incorporated as covariates in the model.

SST: Directly influences zooplankton distribution and abundance, which in turn impacts NARW foraging.
Chlorophyll-a: Provides insights into primary productivity, the foundation of the marine food web that supports zooplankton.
Ocean Currents: Influence the movement and aggregation patterns of both zooplankton and NARWs.

Data Assimilation: Techniques like Ensemble Kalman Filtering or particle filtering can be used to assimilate these real-time data streams into the model, updating the model's predictions of zooplankton and NARW distributions dynamically.
2. Incorporating Anthropogenic Factors:

Ship Traffic: Real-time ship location data from Automatic Identification System (AIS) can be incorporated to quantify ship traffic density within NARW habitat. This information can be used to:

Modify NARW Intensity: Adjust the spatial point process intensity for NARWs to reflect areas of high ship traffic, as these areas may be avoided by whales or pose a higher risk of ship strikes.
Inform Management Actions: Trigger dynamic management actions, such as speed restrictions or rerouting of ships, in areas where high ship traffic overlaps with predicted high NARW densities.

Fishing Effort: Similar to ship traffic, real-time data on fishing effort can be used to identify areas of potential entanglement risk and inform management strategies.
3. Dynamic Management and Conservation:

Risk Maps: The model, updated with real-time data, can generate dynamic risk maps highlighting areas of high NARW presence and potential threats (e.g., high ship traffic, low prey availability).
Adaptive Management: These risk maps can guide adaptive management strategies, such as:

Time-Area Closures: Implement temporary closures of specific areas to fishing or shipping activities when the risk to NARWs is elevated.
Targeted Monitoring: Focus monitoring efforts on areas where the model predicts high NARW densities or potential overlap with threats.
Challenges and Considerations:

Data Latency and Resolution: Ensuring timely access to high-resolution real-time data can be challenging.
Model Complexity and Computational Demands: Incorporating real-time data assimilation and dynamic modeling increases computational demands. Efficient algorithms and high-performance computing resources are crucial.
Uncertainty Propagation: It's essential to quantify and communicate the uncertainty associated with model predictions, especially in a dynamic management context.

Could the observed correlation between zooplankton abundance and NARW presence be influenced by other factors not considered in the model, such as the distribution of predators or alternative prey species?

Yes, the observed correlation between zooplankton abundance and NARW presence could be influenced by other ecological factors not explicitly included in the model. Here are some key considerations:

Predator Distribution:

Predator Avoidance: NARWs might avoid areas with high densities of their predators, such as large sharks or killer whales, even if those areas have abundant zooplankton. The model currently doesn't account for predator distribution.
Prey Competition: The presence of other zooplankton-feeding predators could reduce prey availability for NARWs, even in areas with initially high zooplankton concentrations.

Alternative Prey Species:

Prey Switching: NARWs are known to exhibit some degree of prey switching, meaning they might target different zooplankton species depending on their availability and energetic requirements. The model focuses on overall zooplankton abundance, not specific species composition.
Spatial Heterogeneity in Prey: The distribution of preferred prey species within the broader zooplankton community might not be uniform. NARWs could be concentrating in areas with higher densities of their preferred prey, even if overall zooplankton abundance is lower.

Environmental Factors:

Water Column Structure: Factors like water stratification, salinity gradients, and nutrient upwelling can influence the vertical distribution of zooplankton. NARWs might target areas with specific water column characteristics that concentrate their prey, even if overall zooplankton biomass is similar.
Bottom Topography: Seabed features like canyons or seamounts can create localized upwelling zones that enhance zooplankton productivity. NARWs might be attracted to these areas, independent of broader zooplankton patterns.
Incorporating These Factors:

Predator Data: Integrating data on predator distribution (e.g., from tagging studies or acoustic monitoring) could improve the model's predictive accuracy.
Prey Specificity: Incorporating information on NARW diet preferences and the spatial distribution of key prey species would enhance the model's ecological realism.
Oceanographic Modeling: Coupling the joint species distribution model with oceanographic models that simulate water column dynamics and prey production could provide a more comprehensive understanding of NARW habitat use.

If we consider the ocean as a complex adaptive system, how can this research contribute to a broader understanding of the interconnectedness of species and their environment in the face of global change?

This research provides a valuable case study for understanding the interconnectedness of species and their environment within the complex adaptive system of the ocean, particularly in the context of global change. Here's how it contributes:
1. Highlighting Species-Environment Feedbacks:

Dynamic Relationships: The joint modeling approach explicitly links NARW distribution to the abundance of their zooplankton prey, which is itself influenced by environmental factors like SST. This highlights the dynamic feedbacks between species and their environment.
Climate Change Impacts: As climate change alters oceanographic conditions (e.g., warming waters, shifting currents), the model can be used to project how these changes might cascade through the food web, impacting both zooplankton and NARW populations.
2. Emphasizing the Importance of Scale:

Local to Regional Scales: The study focuses on Cape Cod Bay, but the findings have implications for understanding NARW distribution and habitat use across their broader range.
Connecting to Basin-Scale Processes: Changes in zooplankton abundance in Cape Cod Bay can be linked to larger-scale oceanographic processes in the North Atlantic, such as the North Atlantic Oscillation, which influences SST and primary productivity.
3. Informing Ecosystem-Based Management:

Integrated Approach: The research underscores the need for ecosystem-based management approaches that consider the interconnectedness of species and their environment.
Adaptive Strategies: By incorporating real-time data and accounting for dynamic feedbacks, the model can support adaptive management strategies that respond to changing environmental conditions and species interactions.
4. Predicting and Mitigating Global Change Impacts:

Vulnerability Assessments: The model can be used to assess the vulnerability of NARWs and other marine species to climate change and other anthropogenic stressors.
Conservation Planning: By identifying critical habitat areas and potential threats, the research can inform conservation planning efforts aimed at mitigating the impacts of global change on marine ecosystems.
Key Takeaways for a Changing Ocean:

Interconnectedness: Changes in one part of the ocean system can have cascading effects on other components, highlighting the interconnectedness of species and their environment.
Dynamic Responses: Species and ecosystems are not static; they exhibit dynamic responses to environmental change.
Adaptive Management: Effective conservation and management strategies must be adaptive and responsive to the dynamic nature of marine ecosystems in the face of global change.