A glimpse of food webs from sPACE

This repository contains the data and analysis developed during the Fish-PACE Hackweek 2026.

Project Overview

This project explores the trophic link between ocean color and fish functional diversity by integrating satellite-derived phytoplankton data with NOAA bottom trawl surveys. The project focuses on understanding how phytoplankton functional types (PFTs) cascade through trophic levels to structure fish assemblages.

Folder Structure

• contributor_folders preliminary scripts and data produced during the hackweek.
• final_notebooks (Work in progress) polished, fully documented, and reproducible notebooks that represent the main project output.
• scripts clean scripts and functions.
• data clean datasets.

Collaborators

Name	Role
Josefina Cuesta Núñez	Participant - Project Manager
Maité Barrena	Participant
Patrick Rizk	Participant
Vishwanath Boopathi	Participant *
Jonathan Peake	Project Mentor

Planning

• Ideation jam board
• Project Pitch
• Slack channel: #proj-plankton-fish-guilds-utc-3
• Project Notes
• Final Presentation
• Presentation Slides

Goals

• Goal: reconstruct the trophic link between ocean color and fish assemblages.
• Hypothesis: Different phytoplankton functional types (PFTs) determine the zooplankton size classes, which in turn cascade to support distinct fish guilds.

Datasets

• NOAA's Bottom Trawl Data for Spring and Fall 2024.
• PACE's MOANA & Chl-a data.
• MODIS-AQUA SST.

Workflow/Roadmap

Workflow Overview

Data Acquisition & Pre-processing:

• Biological Data: Processed NOAA bottom trawl surveys (Spring/Fall 2024) for the Northeast US Shelf.
• Satellite Data: Extracted hyperspectral products from PACE (PFTs via MOANA, Chl-a) and MODIS (SST) for matching spatial/temporal windows.

Matrix Construction:

• Species-Abundance (L): Identified a unified "Top 10" most abundant species to build seasonal abundance matrices.
• Functional Traits (Q): Populated a trait matrix (Trophic Level, Body Size, Diet Breadth, Mouth Position) using rfishbase for the dominant species.
• Environmental (R): Conducted a spatiotemporal match-up between trawl stations and satellite pixels, including in-situ bottom temperatures.

Statistical Integration:

• Performed RLQ Multivariate Analysis to link environmental gradients (R) with species composition (L) and functional traits (Q).
• Analyses were conducted seasonally to evaluate how shifting oceanographic conditions and phytoplankton structures reorganize fish functional assemblages.

📅 Project Update: 01-30-2026

✅ Completed Tasks

• Repository Setup: GitHub repository established.
• Study Area Definition: Delimited the spatial scope of the study Latitude: 34.0 – 40.5 Longitude: -75.5 – -71.0
• Data Selection: Adult fish: NOAA bottom trawl surveys + functional traits from fishbase. Phytoplankton: PACE satellite products
• Taxonomic Filtering: Successfully filtered out invertebrate species to focus strictly on fish assemblages.
• Abundance Analysis: - Analyzed species accumulation curves.
  • Spring: ~10 species account for 95% of the total catch.
  • Fall: ~30 species account for 95% of the total catch.
  • Decision: A unified "Top 10" species list has been defined to standardize the analysis across both seasons.
• Species Abundance Matrix:: We built species-abundance matrices for spring and fall, based on the unified top 10 most abundant species.
• Extract satellite data (MOANA, Chl-a, SST) matching station coordinates.
• Environmental Matrix: We merged PACE and MODIS-Aqua data with bottom trawl data (bottom temperature)
• Run RLQ Analysis

⚠️ Methodology Decisions

• Taxonomic Scope (Phytoplankton & Fish): We restricted the analysis to phytoplankton (PACE) and fish (trawl data) only.

  • Reasoning: Ichthyoplankton data was unavailable for the specific spatiotemporal window defined for this project.

• Length Data Removed: We have decided to exclude fish length data from the current analysis scope.

  • Reasoning: Length is not a reliable proxy for age without species-specific validation, which exceeds the current timeline constraints.

• Functional Traits Selection: We prioritized accessible traits from FishBase, explicitly excluding those with low data coverage or limited expected variability (e.g., thermal range, position in water column).- *

  • Reasoning: This ensures dataset completeness and avoids traits that would not provide significant discriminatory power for our analysis.

🚧 Work in Progress (To-Do)

• Documentation: Polish repository.

Results/Findings

• Phytoplankton community structure shapes fish trophic and functional composition.
• There’s a link between primary producer size-classes and the dominance of generalist fish over specialists.
• PACE provides much richer resolution allowing us to “see” the functional base of marine ecosystems.

Lessons Learned

• Extracting and working with PACE products
• Matching them up with in-situ trawl surveys
• Working on shared workflows via GitHub
• Moving between Python and R to leverage the best of both worlds
• Lots of resources and cool ideas to keep exploring

References

• NOAA Fisheries. (2025). Fall Bottom Trawl Survey (2024). Northeast Fisheries Science Center: https://www.fisheries.noaa.gov/inport/item/22560
• NOAA Fisheries. (2025). Spring Bottom Trawl Survey (2024). Northeast Fisheries Science Center: https://www.fisheries.noaa.gov/inport/item/22561
• FishBase. (2025). FISHBASE: A global information system on fishes: https://www.fishbase.se/
• rfishbase v5.02 > Boettiger C, Temple Lang D, Wainwright P (2012). “rfishbase: exploring, manipulating and visualizing FishBase data from R.” Journal of Fish Biology.
• NASA OB. DAAC (2024). PACE Chlorophyll-a concentration products. National Aeronautics and Space Administration. https://oceancolor.gsfc.nasa.gov/ https://pace.gsfc.nasa.gov/
• Brown, O. B., & Minnett, P. J. (1999). MODIS Infrared Sea Surface Temperature Algorithm. Remote Sensing of Environment, 76(1), 3–14.