Abstract The recovery of amplifiable DNA from formaldehyde‐fixed (FF) zooplankton samples has long been considered problematic, but advances in degraded‐DNA retrieval have renewed interest in FF samples. To access the information stored in long‐term zooplankton time series, we evaluated methods for extracting amplifiable DNA from community samples preserved for up to 28 yr in formaldehyde at room temperature. A method previously reported as successful in FF zooplankton stored at 4°C proved ineffective here, likely due to the differing storage conditions. In contrast, by adapting two protocols developed for FF museum specimens—a harsher (Hahn–hot alkaline [HHA]) and a gentler (Hahn–proteinase K column [HPC]) extraction approach—we were able to amplify and sequence a subset of samples. As expected, DNA integrity and sample pH decreased with preservation time, and only short DNA fragments were recoverable, ruling out standard ≥ 300 bp metabarcoding markers. While DNA integrity appeared to be a better predictor than DNA yield for amplification success, the presence of a gel band of the expected size did not always guarantee congruence with microscopy assessments. Although amplifiable DNA was recovered from most samples, including some of the oldest, community compositions concordant with microscopy were consistently recovered only from samples preserved for up to 2 yr. Beyond this point, the HHA and HPC methods produced divergent results, reflecting a trade‐off between removing formaldehyde‐induced cross‐linkages and avoiding additional DNA damage. Among the three small universal markers tested (~ 120–170 bp), only the 18S rRNA V9 region consistently amplified. We conclude by providing recommendations aimed at improving the methods assessed here.
Abstract While it is well recognized that suspended particles can bias titration‐based total alkalinity (TA), whether addition of minerals for alkalinity enhancement introduces systematic TA artifacts remains poorly quantified. To address this gap, we quantified the contribution of ultrafine particles (< 0.2 μ m) to TA titration introduced by limestone (0.1, 0.5, 1, and 2 g L −1 ) under two p CO 2 conditions (420 and 1495 ppm). We found that the TA increase is much higher than the theoretical increase determined by [Ca 2+ ] when added limestone is over 0.1 g L −1 . However, this excess TA is negligible when the added mineral particles exclude ultrafine particles. Also, the particulate TA decreases when there is more mineral dissolution under the high p CO 2 treatment. These findings suggest that even after 0.2 μ m filtration, traditional TA titration of mineral‐treated water may still overestimate true mineral dissolution, potentially leading to inflated carbon dioxide removal estimates.
Abstract Accurate measurements of phytoplankton net primary production (NPP) and calcification are essential for quantifying marine carbon cycling, yet in situ calcification rates remain poorly constrained. We compared 13 C‐ and 14 C‐based tracer methods for NPP in laboratory cultures of the diatom Chaetoceros tenuissimus and the coccolithophore Gephyrocapsa huxleyi across two incubation durations (8 and 24 h). Calcification was quantified only in G. huxleyi , for which experiments were conducted across a range of cell concentrations (10 3 –10 5 cells mL −1 ) while diatom incubations were performed at a single concentration (10 5 cells mL −1 ). 13 C‐ and 14 C‐based NPP and calcification estimates converged in 24‐h incubations across all cell concentrations. There was no significant difference between rates measured with either tracer across all incubations, although short (8‐h) incubations showed greater variability in NPP estimates, particularly for G. huxleyi . In situ application of the 13 C method at the Bermuda Atlantic Time‐Series Study site revealed strong vertical structure and seasonal declines in calcification from July to September, while NPP remained comparatively stable. Particulate 13 C enrichments consistently exceeded analytical detection limits in both laboratory and field settings, demonstrating that low biomass in oligotrophic regions would not limit method sensitivity. These results validate 13 C spike incubations as a practical alternative to radiocarbon for simultaneous measurements of NPP and calcification. To support the adoption of this method, we provide a detailed step‐by‐step standard operating procedure for the 13 C spike incubation method in the Supporting Information.
Abstract Accurately tracing organic matter sources in marine systems remains challenging, particularly where natural and anthropogenic inputs coexist. Compound‐specific stable isotope analysis of amino acids (CSIA‐AA) offers strong potential for organic matter source discrimination, but its application is limited by the small number of validated δ 15 N‐AA tracers and the lack of tracer selection tools. Here, we evaluate and expand δ 15 N‐AA tracer use by combining controlled laboratory mixtures with a model‐independent tracer selection framework. Artificial mixtures of known composition, including phytoplankton, zooplankton, and fecal pellets from dominant offshore wind farm fouling species ( Mytilus edulis and Metridium senile ), were used to directly assess tracer performance. Application of the “CI/CR/CTS” framework (Conservativeness Index, CI; Consensus Ranking, CR; and Consistency‐Based Tracer Selection, CTS) enabled systematic pre‐unmixing evaluation of δ 15 N‐AA tracers. Trophic (Glu, Ile, Leu, Pro), intermediate (Ser), and metabolic (Thr) AA consistently showed strong source discrimination and reliable mixing behavior. Using this selected AA‐tracer set in a frequentist unmixing model (FingerPro) resulted in improved model fit, stability, and interpretability. These results demonstrate that artificial mixtures provide an effective platform for validating δ 15 N‐AA tracer selection. The “CI/CR/CTS” framework offers a flexible, model‐independent approach that enhances δ 15 N‐AA selection in complex, multi‐endmember marine systems and supports more robust assessments of organic matter source apportionment.
Abstract Accurate length data underpin many ecological and physiological analyses, from assessing population size structure to parameterizing biomass and bioenergetic models. Stereo‐baited remote underwater video systems (stereo‐BRUVs) provide a non‐intrusive means of collecting these data in situ, supporting management and conservation of motile marine assemblages. These systems are used across a range of habitats; however, their effectiveness in turbid waters is limited when using standard stereo‐BRUV configurations, designed to maintain accuracy at greater distances under good visibility. Under turbid conditions, individuals are usually only visible within the first few meters of the cameras, often falling outside the stereo‐overlap area, preventing measurements. To address this limitation, we characterized how camera horizontal field of view (H‐FOV) affects the minimum distance and area of stereo‐overlap, as well as measurement accuracy, using GoPro Hero 9 cameras across narrow, linear, and wide lens settings between ranges of 0.5–4.0 m. Wider H‐FOVs substantially increased the usable overlap area and improved nearfield coverage, although distortion at image peripheries reduced accuracy with distance. Spatial error mapping showed high accuracy across most of the overlap (< 1% at 4 m), with pronounced error (> 14%) confined to extreme edges. We recommend selecting the wide H‐FOV to increase the likelihood individuals fall within the stereo‐overlap in turbid conditions, while applying spatial constraints to exclude measurements from image peripheries beyond 2.5 m. This approach expands the effective measurement zone under reduced visibility without compromising accuracy, supporting robust length‐based biomass estimates and improving the applicability of stereo‐BRUV data for downstream ecological analyses in turbid environments.
Abstract Flow cytometry measures the optical properties of individual cells and has become essential for enumerating and characterizing marine phytoplankton. Yet systematic comparison of measurements across instruments remains limited, hampering data comparison across laboratories and long‐term time series. Here, we used 10 flow cytometers from seven manufacturers to analyze population concentration and optical properties of polystyrene beads, cultured phytoplankton, and natural phytoplankton. Light‐scatter signals varied by more than two orders of magnitude across flow cytometers, reflecting differences in optical design and detector sensitivity. Light scattering calibration reduced this variability, lowering coefficients of variation from nearly 100% to below 15% in some cases, although small inter‐instrument differences remained. Despite these optical corrections, concentration measurements across instruments were highly variable, with coefficients of variation ranging from 10% to 133%, which can obscure ecological signals in phytoplankton distributions. Population gating contributed minimally to this variability, with no significant difference between expert and novice users (ANOVA, F < 0.001, p > 0.99). Among laboratory culture and field samples, Prochlorococcus concentrations had the highest variability (CV = 56%) and Synechococcus the lowest (CV = 20%). Variance decomposition attributed ~ 80% of concentration variability to hardware characteristics, including optical design, detector sensitivity, and volume‐measurement methodology. These findings show that while optical properties can be standardized through calibration, instrument differences still dominate uncertainty in concentration estimates. With shared calibration tools for both optical and concentration measurements, marine flow cytometry could achieve the consistency needed to enable global, long‐term syntheses of phytoplankton optical properties and abundance.
Tailwaters are ubiquitous and highly managed ecosystems whose food webs often rely disproportionately on autochthonous energy. In situ continuous dissolved oxygen data are increasingly being used to estimate gross primary productivity and ecosystem respiration in rivers, but this approach is complicated in tailwaters, where upriver discontinuities (i.e., dams) violate commonly employed one-station approaches. In such cases, two-station metabolism models can be applied, although substantial diel variation in flow (a common outcome of hydropower production) requires more complex treatment of water parcel travel times. Here, we present a new two-station metabolism model that allows estimation of reach-scale gross primary productivity and ecosystem respiration in streams and rivers that experience within-day variation in flow. Our approach simplifies two-station variable flow model implementation compared to previous efforts. We apply our model to a 6-yr dissolved oxygen time series and use Bayesian inference to estimate daily gross primary productivity, ecosystem respiration, and gas exchange velocity (k(600)) for a similar to 12-km reach of the Colorado River downriver of Glen Canyon Dam. We compare our model's performance to a more mechanistically detailed and computationally intensive Eulerian dynamic flow model and also to a widely-used one-station model that uses assumptions of reach uniformity that are often strongly violated in tailwaters. These comparisons show that our metabolism estimates conform with output from the more detailed dynamic flow model and that the one-station approach deviates substantially from both two-station approaches. Our new stream metabolism model can help resolve a fundamental analytical impediment in tailwater ecology.
Monitoring short-term changes in surface sediment elevation is fundamental to understanding erosion, transport, and deposition dynamics in shallow coastal environments. However, commonly used field approaches, such as horizontal markers, sediment erosion tables, subsurface sediment plates, or erosion pins, are not always cross-validated under both controlled and field conditions, limiting confidence in their comparative performance. This study experimentally evaluates the performance and potential biases of two widely used and cost-effective bed-level monitoring tools: (1) subsurface sedimentation plates (without string, with string, and with string and buoy) and (2) sedimentation bars. Methods were evaluated under controlled conditions in a hydraulic flume using non-cohesive sandy sediment. Plates and bars were subjected to unidirectional and oscillatory flow regimes at two velocity levels (11 and 23 cm s-1) to compare their erosion responses and assess the hydrodynamic interference generated by their structural components, quantified through Reynolds numbers. A complementary field deployment was conducted over 1 year in a Zostera marina meadow in the Bay of Santander estuary (Spain). Across flume flow regimes and field habitats, bed-level change estimates were comparable among methods, with no detectable differences within the resolution of the experimental design (minimum detectable difference approximate to 1.9 mm for current-driven and 2.5 mm for wave-driven conditions). Although the buoy produced localized turbulence, the resulting shear stress was likely below critical erosion thresholds. Together, the results support the use of sedimentation plates (with or without string or buoy) and bars as practical cost-effective tools for monitoring short-term bed-level change in shallow, non-cohesive sandy and vegetated environments.
Anthropogenic noise has increased over the last decades in marine ecosystems, due to diverse human activities (e.g., ships, pile driving, drilling, dredging, underwater explosions, and seismic testing). Nevertheless, there is currently a knowledge gap regarding the effects of noise pollution, especially substrate vibrations, on benthic species. Within this context, we developed a standardized experimental method to study the ecophysiological responses of benthic invertebrates exposed to substrate vibrations, following a dose-response approach. Our protocol enables the exposure of individuals to different treatments of substrate vibrations (three acceleration levels and a control) with several aquariums (i.e., replicates) exposed simultaneously for each treatment. Four independent vibrating structures were built, allowing for the aquarium replicates to be exposed to the different treatments simultaneously without vibration propagation between structures. The objective of the study was therefore to fully characterize the vibrating structures, for frequencies ranging from 5 to 50 Hz, to ensure precise knowledge of the vibration exposure levels inside the aquariums. Especially, intra-aquarium variability of the vibration signal in function of the frequency was estimated as well as long-term signal stability on 48 h. Finally, we presented an example of application of our protocol on great scallops, suggesting that our system can be used to generate dose-response relationships between substrate vibration levels and biological responses.
Assigning taxonomic identity and thus assessing taxonomic diversity in gelatinous zooplankton is challenging due to their fragility and difficult preservation. Formalin fixation compromises DNA integrity, while flash-freezing, ethanol, and acetone preservation affect morphological features. In this study, we developed a methodological protocol to optimize morphological and molecular identification of pelagic cnidarians and tunicates from the Gulf of Naples (Mediterranean Sea). Specimens were identified while fresh and preserved using four methods: liquid nitrogen (N), ethanol (EtOH), ethanol replaced with acetone (EA), and formalin followed by ethanol and acetone (FEA). Genomic DNA was extracted and PCR-amplified targeting the 18S rRNA gene, being subsequently assessed for concentration, integrity, and sequence quality. FEA specimens retained better morphology compared to N, EtOH, or EA samples, while sequencing performance was comparable between methods. DNA concentrations did not significantly differ among treatments (range: 139-1645 ng ind(-1)) and smaller taxa showed higher values per unit size compared to larger ones. DNA integrity was lower in N and FEA samples, and often undetectable in EA and FEA when assessed by gel electrophoresis. Although 18S : gDNA ratio decreased in some FEA samples, this method still produced high-quality reads (> 500 nucleotides). Our findings show that formalin fixation followed by ethanol and acetone does not compromise molecular analysis, even for species with low DNA content. Unlike other methods, this optimized protocol preserves essential morphological features for accurate taxonomic identification of gelatinous zooplankton while substantially reducing molecular processing time. Overall, the adoption of the proposed protocol could significantly enhance biodiversity assessments of gelatinous taxa.
Concentrations of the photosynthetic pigment chlorophyll a are commonly used as a proxy for phytoplankton biomass in aquatic systems. Traditional methods for extracting chlorophyll a in discrete samples limit measurement frequency, while in situ sensor technology provides high frequency chlorophyll a fluorescence measurements at a more appropriate time scale to study processes related to phytoplankton turnover and a more rapid response to harmful algal blooms. Caveats to in situ measurements include that temperature, fluorescent dissolved organic matter, and turbidity can affect accuracy and that chlorophyll a fluorescence does not always correlate well to extracted chlorophyll a. Therefore, practitioners need guidance to interpret fluorescence data considering potential errors and relationships to extracted chlorophyll a to inform management decisions. Before this project, the nature of those relationships had not been tested across a wide gradient in water properties. To address these gaps, a one-year study was conducted across 12 biogeochemically diverse sites to quantify possible fluorescence interferences and test the predictability of extracted chlorophyll a using uncorrected in situ chlorophyll a (in relative fluorescence units) combined with routinely measured water quality parameters. Overall, sensor fluorescence was positively correlated with extracted chlorophyll a, but the strength and drivers of the relationship varied by site. Temperature, turbidity, and fluorescent dissolved organic matter influenced sensor readings independently of phytoplankton biomass. Hence, considerations for in situ sensor implementation depend on monitoring goals and resources. Where feasible, we recommend the simultaneous use of in situ and extractive approaches to track short-term variability and long-term change in chlorophyll a.
Sediment traps, often used in tandem with preservatives or poisons, are widely used for the collection of particulate organic matter (POM), providing insight into the source to sink mechanisms that shape major biogeochemical cycles and sedimentary carbon sequestration. The effectiveness of these treatments has been studied for marine POM, but is poorly constrained for freshwater POM, whose molecular composition differs due to greater terrigenous input. We tested four treatments commonly used in marine sediment traps, along with untreated controls, over a 1-yr period to assess their ability to retain original elemental and isotopic compositions of organic carbon (%TOC and delta 13Corg) and nitrogen (%TN and delta 15N) in POM collected from a freshwater pond (Willow Pond, Michigan, USA). Consistent with marine studies, 0.005% (w/v) mercuric chloride was the most effective in preserving both C and N geochemical signatures across the 1-yr incubation period. Untreated samples also maintained original POM geochemical signals but only for approximately 6 months, providing an inexpensive, nontoxic alternative that can be used for shorter deployment or storage periods.
Carbon (delta 13C) and nitrogen (delta 15N) stable isotope ratios allow the reconstruction of food webs, provided that all trophic levels are properly characterized. In freshwater ecosystems, small-bodied invertebrates often occupy central positions in food webs, with whole-body samples commonly used for their stable isotope analysis, due to insufficient sampling material for single-tissue analysis. However, whole-body samples include multiple tissues, each with its own metabolic routing and turnover rate, thus potentially having different isotopic values. Using two genetic lineages of the Gammarus fossarum complex (Crustacea: Amphipoda) as models, we compared the delta 13C and delta 15N stable isotope values of muscle and whole-body samples using isotopic niche analysis and diet reconstruction via mixing models. When tested, decarbonation of whole-body samples through acidification had a significant effect on the delta 13C values due to the inorganic carbonate in crustaceans' exoskeleton. Thus, decarbonated delta 13C values were used for whole-body samples in the analyses. Whole-body samples had significantly higher delta 15N and a narrower niche than muscle samples. The interspecific niche overlap did not vary among tissues for one of the lineages, but was twice as high using whole-body samples compared to muscle samples for the other lineage. Mixing models for whole-body samples performed better than those for muscle samples, due to the absence of muscle tissue-specific trophic discrimination factors (TDFs) in the literature, highlighting the need for tissue-specific TDFs in macroinvertebrates. Our results reinforce the pre-established caution around the use of whole-body samples and underline biases that can arise from comparing different tissues.
While acidification is widely recognized as a critical step influencing the molecular formula distribution of dissolved organic matter (DOM), pH-dependent molecular fractionation remains largely unexplored. Here, we conducted a comprehensive molecular characterization of DOM by direct analysis of original water samples across a pH gradient (non-acidified, 5, 3, 2, and 1) using online ultrahigh-performance liquid chromatography coupled with Orbitrap mass spectrometry. The results demonstrate that acidification-induced molecules experience a compositional shift from saturated compounds to more oxidized polyphenolic and condensed aromatic species as pH decreases from 5 to 2. However, molecules with higher double-bond equivalents and nitrogen-containing compounds are largely detected when the pH is lowered to 1. This is primarily attributed to the enhanced protonation of DOM molecules under strongly acidic conditions, which reduce their ionization, further improves their retention in reversed-phase systems and facilitates their effective separation from inorganic salts, indicating pH-dependent fractionation of ionizable species. Compared with direct analysis of original water under optimized acidification (pH 1), terrestrially derived and microbial-altered components are largely underestimated in DOM from solid-phase extraction. Given that DOM sources and secondary alterations are key processes governing the fate of DOM, the method developed in this study provides new insights into the composition and transformation of DOM in the original water and offers a novel perspective for exploring its biogeochemical implications in aquatic systems. Future studies should systematically analyze samples from diverse environments to constrain the driving factors of DOM compositional heterogeneity and its environmental implications.
The demand for efficient image sorting methods has increased due to technological advancements that enable more intensive phytoplankton monitoring. Both statistical and machine learning algorithms can misidentify algal taxa in taxonomically diverse samples, in which phytoplankton morphology and image traits can vary. We evaluated the statistical filtering performance of the image processing software of an imaging flow cytometer (FlowCam) for two approaches to image library development; these were applied independently to seven commonly occurring algal shapes in mixed natural samples. The "intrinsic method" used a small selection of images (5-15 images of a target taxon) from the same sample being filtered (i.e., intrinsic), whereas the "compiled method" used a larger selection of images (30-80 images of a target taxon) compiled from multiple samples. Filter performance varied with the type of image library, image library size, and target taxon. The largest image libraries offered the highest recall (> 86% for intrinsic, > 94% for compiled) but lower precision (3-85% for intrinsic, < 1-8% for compiled). Precision was highest for the smallest image libraries, and was higher for the intrinsic method (> 75% for most taxa) than the compiled method (< 20% for most taxa). Statistical filtering performance was higher for larger, solitary-celled taxa with relatively uniform features (e.g., Gyrosigma) compared to small-celled colonial species with more complex or variable shapes (e.g., mucilaginous colonial cyanobacteria, and Scenedesmus). Iteratively using the intrinsic statistical filtering method with manual correction between each iteration can be used to augment manual sample classification and reduce processing time.
Recent technological advancements have rapidly expanded our capacity for collecting image data in the marine environment, but processing images into meaningful ecological metrics remains a manual, time-consuming, and biased process. This is particularly challenging with electro-optical cabled imaging systems which generate images at a rate that makes manual identification impractical. To address this challenge, we have developed a machine learning-assisted method for annotating images. Our approach leverages a pre-trained model based on marine-specific imagery from the FathomNet database (499 classes; some classes to species level). We demonstrated the application of this method on a 1-yr time series of images collected at Southern Hydrate Ridge by a digital still camera on the NSF Ocean Observatories Initiative Regional Cabled Array, which resulted in 92,153 benthic megafaunal (organisms > 2 cm) annotations across 10 morphotaxa classes in 50,840 images. This method annotated the full dataset in 6 weeks, compared to an estimated similar to 5.9 yr required for fully manual annotation, representing approximately a 50-fold increase in efficiency. This process also produced a computer vision model with a precision of 0.75, recall of 0.80, mAP50 of 0.84, and mAP50-95 of 0.65. Our method combines machine learning efficiency with human expertise to create high-quality, verified datasets. The output of this methodology is key to achieving the full potential of sustained ecosystem monitoring via cabled observing systems that can capture both short-term and long-term ecological and environmental dynamics.
Monitoring phytoplankton abundance is essential for understanding ecosystem dynamics and detecting harmful algal blooms (HABs). Traditional microscopy-based single-cell counting, while accurate, is time-consuming and poorly suited to high temporal and spatial resolution monitoring. Automated imaging approaches have therefore been developed to assist cell detection and counting, but they face challenges when dealing with colonial forms. In this study, we investigate transfer learning for automated cell counting in digital images of phytoplankton colonies acquired with a FlowCam system. Several Convolutional Neural Network (CNN) architectures pre-trained on the ImageNet database were fine-tuned using annotated datasets of two ecologically relevant colonial taxa commonly observed in the English Channel and the North Sea: Pseudo-nitzschia and Phaeocystis globosa. Model performance and robustness were evaluated using mean absolute error (MAE) and species-specific accuracy metrics. Across both taxa, DenseNet121 architecture achieved the best performance, with a Top 2 accuracy up to 99% for P.-nitzschia and a Top 20% accuracy exceeding 87% for P. globosa. Data augmentation improved model robustness and generalization, particularly for colonies with higher cell numbers. These results demonstrate the ability of deep learning to capture complex spatial patterns and improve counting accuracy across taxa with different morphologies. Beyond methodological performance, this study raises questions about the adequacy of current HAB alert thresholds based on microscopic counts when transitioning to automated imaging systems. Transfer learning provides a robust, fast, and scalable approach that complements traditional monitoring and supports the development of improved environmental assessment strategies.
Mountain lakes are highly sensitive ecosystems and effective sentinels of environmental change, yet the exposure and magnitude of the human footprint remain poorly quantified. In this study, we develop a simple and non-invasive abiotic index to assess cumulative pressures on mountain lakes. The proposed index integrates eight variables grouped into three categories (potential human impacts, accessibility, and recreational exposure) into a single cumulative score, providing a straightforward but informative measure of human pressure on lakes. We employed a dataset comprising 29 lakes from two Iberian protected areas (Sierra de Gredos and Sanabria region) as a case study for assessing the reliability of the index. We observed strong variability among lakes, highlighting the importance of local pressures that are often overlooked by global indicators such as the Human Footprint Index. Our index provides an accessible first-step diagnostic tool that facilitates the integration of human pressure assessment into site-specific monitoring and conservation actions.
Habitat heterogeneity is a key driver of temporal and spatial variability of subtidal marine benthic biodiversity. However, this makes it a challenging environment in which to measure and quantify the factors driving biodiversity in a consistent manner. Current methodologies are either expensive, logistically challenging, require extensive technical knowledge, or are limited to shallow water environments. To overcome some of these limitations, a novel "SeaPen" device was designed as a cost-effective, safe, easily deployable instrument that can provide repeatable and comparable survey results over the small spatial scales required to describe heterogeneous environments. The device, which comprises a tripod frame holding a steel weighted bar in the center, allows gravity to force a graduated metal spindle into the substratum when it is dropped. A video camera fixed to a tripod leg allows the recording of the nature of the seabed and the depth of penetration of the spindle, from which seabed properties can be described. The device does not need to be recovered to the surface between measurements, which allows it to be repeatedly dropped for multiple replicates in a single deployment. The combination of a penetrating spindle and a camera allows for rapid repeat measures and, most importantly, can be used to measure heterogeneity on finer scales (within a site; m) than other more traditional techniques. The SeaPen was used in a shallow water (<50 m) coastal habitat on the West Antarctic Peninsula. The survey area comprised heterogeneous substrata composed of sediments interspersed between boulders, cobbles, and pebbles, which allowed for proof of concept that the SeaPen could discriminate a wide range of seabed habitat types.
Heart rate is a popular proxy of physiological responses, but the highly complex and variable cardiac data obtained from organisms such as marine invertebrates pose a major challenge to efficient and accurate data processing. To address this, we developed a novel, integrative algorithm for rapid and automated cardiac data processing. This algorithm primarily employs autocorrelation for time series analysis to identify recurrent heartbeats and compute heart rates from their periods. A genetic algorithm framework was used to implement such an autocorrelative analysis to filter out noise and extract meaningful signals, maximizing data utilization. A tracking index is also incorporated to reference previous timepoints and reduce errors associated with complex waveforms. To evaluate its performance, we compared the algorithm estimates to manually obtained heart rates of 33 individuals of marine invertebrates (from nine species of gastropods, bivalves, and crustaceans). The results showed that these features collectively improve data utilization (mean percentage count > 90%) and accuracy (mean absolute percentage error = 3%). By avoiding reliance on any predetermined characteristics, this algorithm can not only accommodate case-by-case variability and thus be applicable to diverse taxa, but also potentially extend to analyze periodicity in other biological time series data such as valvometry, acoustics and movement patterns. As an open-source tool, this algorithm encourages collaborative efforts and further developments that refine and expand its applications, thereby enhancing our capabilities in physiological monitoring and analyses.
