Quantitative Analysis of PlanktoScope Data: a powerful tool for exploring ecosystem function

Technical Summary: Quantitative Analysis of PlanktoScope Data

Problem Statement
Plankton communities are fundamental to aquatic ecosystems, driving primary production and global biogeochemical cycles. However, traditional monitoring methods relying on manual microscopy are time-consuming, labor-intensive, and require high levels of taxonomic expertise. These limitations hinder the ability to conduct large-scale, high-frequency spatio-temporal studies necessary for understanding ecosystem function and predicting responses to climate change and anthropogenic stressors. There is a critical need for efficient, automated, and cost-effective tools to optimize analytical procedures and minimize data processing time.

Methodology
This study evaluated the effectiveness of the PlanktoScope, an open-source, automated imaging system, coupled with the web-based Ecotaxa platform for classifying planktonic organisms. The research was conducted as a proof-of-concept in the coastal waters of La Paz Bay, Mexico, over a three-month period (September to November 2024).

• Sampling: Weekly plankton samples were collected at a single station (El Mogote) using a 20 µm mesh phytoplankton net via 5-minute horizontal tows. Samples were preserved at approximately 10°C and transported to the laboratory within two hours.
• Preprocessing: Samples were filtered through a 300 µm mesh to remove large organisms that could clog the PlanktoScope's fluidic system. A 3 mL subsample was homogenized from the 200 mL cod end.
• Imaging: The PlanktoScope utilized a stop-flow method to capture 200 images per subsample. Images were processed using the Python-based library MorphoCut, which applied background estimation, edge detection, and morphological operations to extract Regions of Interest (ROIs).
• Classification: 32 mathematical image descriptors were calculated for each ROI. Data were uploaded to the Ecotaxa platform, where a pre-trained machine learning model provided preliminary taxonomic labels. These were subsequently reviewed and corrected manually by trained taxonomists to ensure accuracy. Ambiguous images (e.g., blurry or overlapping organisms) were excluded.
• Quantification: Organism counts were extrapolated to a filtered water volume of 10.92 m³ and further normalized to 100 m³. Biovolume was estimated based on cell dimensions (ellipsoid vs. spheroid) and converted to carbon biomass using established conversion parameters.
• Statistical Analysis: Analyses included non-metric multidimensional scaling (nMDS) to assess community composition shifts, Normalized Biomass Size Spectrum (NBSS) analysis using Bayesian models to characterize size distribution, and correlation with satellite-derived Chlorophyll-a (Chl-a) data.

Key Results

• Taxonomic Identification: The system successfully identified 28 distinct taxonomic groups, with 22 assigned to the genus level. The assemblage was dominated by the diatom genus Chaetoceros (Bacillariophyta).
• Temporal Shift: A distinct shift in planktonic composition was detected between the warm season (September–October) and the warm–cold season transition (November).
  • September–October: Bacillariophyta (diatoms) dominated both absolute abundance and carbon biomass.
  • November: A collapse in diatom abundance and biomass occurred. This was accompanied by a drastic increase in the carbon biomass of Arthropoda (specifically Copepoda) and Cyanobacteria (Trichodesmium).
• Size Spectrum: The Normalized Biomass Size Spectrum (NBSS) showed negative slopes throughout the study, indicating a community dominated by smaller organisms. However, the slope became less negative in November, reflecting a shift toward larger-size cells (Arthropoda) and a reduction in the dominance of small-size organisms.
• Satellite Correlation: Satellite Chl-a data revealed a divergence between regional and local scales. While Chl-a concentrations increased regionally in the Bay of La Paz during November, they decreased at the specific El Mogote sampling site, correlating with the observed local collapse of autotrophic Bacillariophyta.

Significance and Claims
The authors claim that the PlanktoScope successfully demonstrated its potential as a high-frequency, cost-effective tool for monitoring plankton community composition and detecting local temporal shifts. The study highlights the system's ability to provide quantitative data on biomass distribution and size spectra, which are critical for understanding food web structure and ecosystem function.

The paper emphasizes that while the PlanktoScope offers a scalable alternative to traditional microscopy, the current taxonomic resolution was limited by the availability of regional training datasets in Ecotaxa. Consequently, the authors conclude that while the findings are promising, further validation over larger spatio-temporal scales and across different oceanographic contexts is necessary to fully realize the tool's potential for global ecosystem monitoring. The study positions the PlanktoScope as a means to lower the entry barrier for institutions with limited laboratory infrastructure, facilitating more consistent and rapid assessments of plankton communities.