Search

Zooplankton size is a crucial indicator in marine ecosystems, reflecting demographic structure, species diversity and trophic status. Traditional methods for measuring zooplankton size, which involve direct sampling and microscopic analysis, are laborious and time-consuming. In situ imaging systems are useful sampling tools; however, the variation in angles, orientations, and image qualities presented considerable challenges to early machine learning models tasked with measuring sizes. Our study introduces a novel, efficient, and precise deep learning-based method for zooplankton size measurement. This method employs a deep residual network with an adaptation: replacing the fully connected layer with a convolutional layer. This modification allows for the generation of an accurate predictive heat map for size determination. We validated this automated approach against manual sizing using ImageJ, employing in-situ images from the PlanktonScope. The focus was on three zooplankton groups: copepods, appendicularians, and shrimps. An analysis was conducted on 200 individuals from each of the three groups. Our automated method's performance was closely aligned with the manual process, demonstrating a minimal average discrepancy of just 1.84%. This significant advancement presents a rapid and reliable tool for zooplankton size measurement. By enhancing the capacity for immediate and informed ecosystem-based management decisions, our deep learning-based method addresses previous challenges and opens new avenues for research and monitoring in zooplankton. [ABSTRACT FROM AUTHOR]

Plankton are excellent indicators of ecosystem status and fisheries because of their pivotal role in marine food webs and their core values in the integrated ecosystem assessment (IEA). Monitoring plankton is essential to understand their dynamics and underlying processes. Recent advances in imaging technologies have enabled in situ, high-frequency, real-time observations of plankton in coastal waters. While high-frequency plankton time series have provided unprecedented fundamental information about physical and biological processes, understanding and identifying the underlying mechanisms that influence plankton dynamic remains a major challenge. We use high-frequency plankton data from PlanktonScope as an example to examine the impacts of physical and biological processes on plankton dynamics at different temporal scales. Frequency patterns were identified for both environmental factors and different plankton groups that matched in time. Using logistic regression models on the selected daily peaks for different plankton groups, we found that diurnal cycle, monsoon season, and major episodic events, such as typhoons, had major impacts on the dynamics of plankton, as proxied by our indicators. We further synthesized, across multiple spatiotemporal scales in the study area, the impacts of various processes on plankton with different mobility. Our study demonstrates that the suite of plankton indicators simultaneously generated from PlanktonScope provides a robust holistic view of pelagic ecosystem status over a broad range of spatiotemporal scales. In situ imaging systems like PlanktonScope are promising tools for near real-time plankton monitoring and a deep understanding of plankton dynamics. • Plankton imaging systems can generate a suite of near real-time high-frequency plankton indictors. • High-frequency plankton indicators provide an excellent holistic view of ecosystem status and fisheries. • A combination of highfrequency data and frequency analysis revealed the impact of physical processes at different scales. • A trait-based approach allows a comparison among different plankton groups and their response to physical processes. • Automated deep learning approach can reduce the time lag between observations and data availability. [ABSTRACT FROM AUTHOR]

• An in-situ imaging system captured Noctiluca bloom in Dapeng Bay, P.R. China. • A logistic growth model was applied to simulate the growth and decay of Noctiluca. • We demonstrated an early warning framework using the in-situ observations and model. The frequency and intensity of harmful algal blooms (HABs) are likely to increase with climate change and increases in nutrients due to anthropogenic activities. To manage the environment and health threats to surrounding organisms, it is essential to improve our capability to provide early warning of harmful species. Using Noctiluca scintillans as example, a dinoflagellate that often blooms in Asia and whose presence is rapidly expanding worldwide, we demonstrate a potential way to predict N. scintillans blooms by deploying an in-situ plankton imaging system: PlanktonScope. PlanktonScope is a shadowgraph imaging system and capable of imaging organisms 40 µm – 5 cm. This system was deployed in Shenzhen Bay, China, recording images at 0.5 Hz in April 2016. Using the automated image processing procedure, we found that PlanktonScope recorded nearly a full cycle of a N. scintillans bloom. A simple resource-limited logistic population growth model was then applied to develop a predictive framework for the outburst of N. scintillans. Environmental changes could explain the observed dynamics of N. scintillans ; however, it was difficult to quantify the relationship between the dynamics of N. scintillans and environmental variables. Our study demonstrated that a camera system like PlanktonScope could be useful in developing an early warning system for N. scintillans blooms, which could be applicable to other HAB species for proactive management. [ABSTRACT FROM AUTHOR]