Technological innovations for predicting fish age represent a paradigm shift from conventional age estimation methods used in fisheries science. Recently developed secondary methods rely on models trained on conventional age estimates, derived from subjective interpretation of growth patterns and a biological property of the fish to predict age. Hence, quantifying the error that propagates from these models and the conventional age estimates on which they are trained is critical to fully account for uncertainty in predicted ages used in fisheries stock assessments. We review the development of three secondary age prediction methods: image analysis, epigenetics and Fourier transform near-infrared spectroscopy (FT-NIRS); the use of artificial intelligence (AI) within each method; and present an approach to examine the effects of ageing error on AI age prediction models. As a case study, we conducted an empirical study, coupled with simulation, to quantify the effects of ageing error on the performance of a multimodal convolutional neural network (MMCNN) model used to predict eastern Bering Sea walleye pollock (Gadus chalcogrammus) ages from otoliths analysed using FT-NIRS. Results indicated repeatability of predicted ages was high between instrument operators, while adding ageing error resulted in a slight decrease in model performance from R 2 = 0.92 and CV = 7.6% to R 2 = 0.87 and CV = 10.2% on test datasets. Our results also suggest the MMCNN model is robust to noise in calibration age data and model performance may be better than performance metrics indicate when the model is evaluated against data with increased error.

Polar cod (Boreogadus saida) is a widespread, cold-adapted fish that plays a central role in Arctic food webs. Rising temperatures, declining sea ice, and associated changes in ecosystem dynamics may threaten this species. We investigated the influence of temperature on the pre-winter size of juvenile polar cod (an important predictor of overwinter survival and recruitment), examining hatch timing and growth across warm and cold years in the Chukchi Sea. Otolith-based estimates of age and growth showed that juveniles collected in warm years were generally larger and older, but detected no significant differences in size-at-age or recent growth (7 days prior to capture). Recent growth was inversely related to mean water column temperature early in the sampling season, a trend that decoupled later in the year. Temperature at catch explained little of the observed variation in growth, suggesting that the expected positive influence of temperature on growth was likely mediated by covarying oceanographic and ecological shifts. Ultimately, multiple environmental factors influenced final pre-winter size, with variation driven more by hatch phenology and an extended growing season, than by accelerated somatic growth rates.

Context Fish daily age information provides insight into growth, phenology and recruitment but is rarely incorporated into the decision-making process for management because of labor-intensive data collection.Aims We tested Fourier transform-near-infrared (FT-NIR) spectroscopy as a rapid method (similar to 1.0 v.similar to 60 min per otolith for microscopic methods) to determine daily age of walleye pollock (pollock; Gadus chalcogrammus), which supports an important fishery.Methods We reared two annual cohorts of pollock over 3 years to accrue a comprehensive dataset for calibration models.Key results FT-NIR spectroscopy provided an improvement over predicting daily age from otolith weight, fish weight or fish length. A calibration model developed using microscopically aged fish had a root mean square error (RMSE) from cross validation (CV) of 12.2 days. Models developed from the full dataset with ages assigned from cohort hatch dates had a test dataset RMSE of 35.4-40.7 days and a CV RMSE of 29.4-36.9 days.Conclusions Spectra were affected by size-at-age, which could affect model efficacy because growth and size are influenced by environmental variability. This can be addressed with robust calibration datasets, model testing and model updating.Implications FT-NIR spectroscopy age prediction is likely to be sufficient to capture population-scale shifts in hatch dates of pollock.

The Pacific cod (Gadus macrocephalus) fishery was closed in 2020 after a rapid decline in biomass caused by the marine heat waves of 2014–2019. Pacific cod are exceptionally thermally sensitive and management of this fishery is now challenged by increasingly unpredictable climate conditions. Fisheries monitoring is critical for climate readiness, but short-term monitoring data may be inadequate for recognizing and anticipating change under rapid climate changes. We propose an interdisciplinary, marine historical ecology framework that looks to long-term records (local and traditional knowledge, history, archaeology, and paleoclimatology) to capture a long range of ecological variability and provide historical context for management. In order to connect to contemporary fisheries management, this framework must be built on a common vocabulary and an understanding of the key metrics used in fisheries stock assessments. Here, we propose metrics derived from Pacific cod stock assessment and synthesize information relevant to understanding the effects of past warming periods on cod populations across the Gulf of Alaska and Bering Sea. This case study provides a framework for thinking about how to use these historical records in the context of fisheries management under rapidly changing climate conditions.

Knowledge of the reproductive biology of fishes is essential for effective fisheries management. Information derived from an understanding of fish reproduction, such as size and age at maturity, is used in models to assess fish stocks and can affect estimates of important ecological processes such as recruitment, abundance, and trophic interactions. Common practices for determining the reproductive status of teleost fishes include macroscopic evaluation of gonads as well as histological analysis. However, macroscopic evaluation can be biased and histological analysis is time-consuming, resulting in limitations to spatial and temporal data availability. Here, we explore Raman spectroscopy of ovaries as a novel approach to rapidly determine the reproductive status of walleye pollock (Gadus chalcogrammus), a commercially and ecologically important species in the North Pacific. We used a two-stage partial least-squares (PLS) regression analysis followed by a linear discriminant analysis (LDA) to classify walleye pollock ovary samples as physiologically mature or immature and to subsequently predict their histologically-determined reproductive stage based on the Raman spectra. Biologically mature samples with visible yolk differentiated from mature and immature samples (non-yolked; 99% accuracy). Non-yolked ovaries that were physiologically mature (either mature non-developing or previously spawned) were further differentiated from physiologically immature ovaries (93% accuracy). In addition, detailed, histologically-determined reproductive stages of yolked samples also differentiated via Raman spectroscopy, but with reduced accuracy (79% - 86% accuracy). Our results indicate that accurate identification of maturity status and the reproductive staging of oocytes of walleye pollock based on spectral data from ovaries is possible. This can provide a fast and efficient way to increase the availability of a key component of reproductive data to inform fisheries research and management.

The northern rockfish ( Sebastes polyspinis) ) is an economically valuable, long-lived species distributed over the continental shelf of the North Pacific Ocean. Ages for this species which can be in excess of 80 years comprise an essential component of models for assessing population status and are crucial for fisheries management. Traditional microscope-based methods of estimating age using otoliths can be time-intensive and prone to reader variability. We explored the application of Fourier transform near infrared (FT-NIR) spectroscopy coupled with multimodal convolutional neural networks (MMCNN) for age prediction. Our study included 2613 FT-NIR scans and associated ages of northern rockfish otoliths from years 2013-2019, with ages ranging from 3 to 66 years. The optimal MMCNN model demonstrated strong performance, yielding an R2 2 of 0.92 and an RMSE of 3.38 for the training set and an R2 2 of 0.89 and an RMSE of 3.74 for the test set. Spectral information in the 11,500 to 4000 cm-1 1 wavenumber range, otolith weight, and other biological/geospatial data contributed to age predictions that were comparable to traditional age estimates. Despite challenges, FT-NIR spectroscopy coupled with MMCNN emerged as a promising alternative for age estimation in long-lived species. This approach, while demonstrating effectiveness for northern rockfish, could be a valuable tool for diverse fish species, supporting sustainable fisheries practices and population monitoring.

Understanding the material transport and mixing processes in the Solar protoplanetary disk provides important constraints on the origin of chemical and isotopic diversities of our planets. The limited extent of radial transport and mixing between the inner and outer Solar System has been suggested based on a fundamental isotopic dichotomy between non-carbonaceous (NC) and carbonaceous (CC) meteorite groups. The limited transport and mixing could be further tested by tracing the formation regions of individual meteoritic components, such as Ca-Al-rich inclusions (CAIs) and chondrules. Here, we show further evidence for the outward transport of CAIs and chondrules from the inner and subsequent thermal processing in the outer region of the protoplanetary disk based on the petrography and combined Cr-Ti-O isotope systematics of chondrules from the Vigarano-like (CV) carbonaceous chondrite Allende. One chondrule studied consists of an olivine core that exhibits NC-like Ti and O, but CC-like Cr isotopic signatures, which is enclosed by a pyroxene igneous rim with CC-like O isotope ratios. These observations indicate that the olivine core formed in the inner Solar System. The olivine core then migrated into the outer Solar System and experienced nebular thermal processing that generated the pyroxene igneous rim. The nebular thermal processing would result in Cr isotope exchange between the olivine core and CC-like materials, but secondary alteration effects on the parent body are also responsible for the CC-like Cr isotope signature. By combining previously reported Cr-Ti-O isotope systematics of CV chondrules, we show that some CV chondrules larger than similar to 1 mm would have formed in the inner Solar System. The accretion of the millimeter-sized, inner Solar System solids onto the CV carbonaceous chondrite parent body would require their very early migration into the outer Solar System within the first 1 million years after the Solar System formation.

Our novel approach for fish age prediction uses quantitative analysis of Fourier transform near-infrared (FT-NIR) spectra of otoliths by means of multimodal convolutional neural networks (MMCNN). We integrate two key data modalities that are related to fish ages: the entire range of wavenumbers of FT-NIR spectra and corresponding biological and geospatial data for nearly 9000 walleye pollock ( Gadus chalcogrammus) specimens. The proposed model extracts informative spectral features automatically and elucidates hidden structural relationships associated with fish growth to improve age predictions. Absorbance associated with 7000 to 4000 cm −1 wavenumbers had the highest influence on model predictions followed by fish length, latitude, depth, and temperature. The optimal model resulted in good overall performance with an R 2 of 0.93 and RMSE of 0.83 for training data set and R 2 of 0.92 and RMSE of 0.91 for test data set. MMCNN's age predictions were comparable to microscope-based ages yielding as good or slightly better precision. Moreover, the model outperformed classical partial least squares analysis of otolith spectra and remedied prediction bias at older ages of fish.

Fourier-transform near infrared (FT-NIR) spectroscopy of ovarian tissue was used to predict maturity status of fish species with variable reproductive strategies collected at limited time periods of their spawning cycle. Reference data were derived from histologically prepared tissue samples from four species: Pacific cod (Gadus macrocephalus), walleye pollock (Gadus chalcogrammus), Greenland turbot (Reinhardtius hippoglossoides), and northern rockfish (Sebastes polyspinis). Each data set was classified into reproductively immature (non-spawning) and reproductively mature (spawning-capable) categories. Principal component analysis of spectral data showed separation between ovarian tissues of spawning-capable and non-spawning females. Multivariate classification using partial least squares discriminant analysis indicated good discrimination based on spawning status with high predictive accuracy. Greenland turbot and northern rockfish showed clear distinction between ovaries of spawning-capable and non-spawning females and a model validation with 100% and 96.6% classification accuracy, respectively. Pacific cod and walleye pollock had more complicated reproductive patterns at time of collection and classification rates were still 96.6% and 92.1%. This study demonstrated the potential application of FT-NIR spectroscopy to predict spawning status from ovarian tissue even for species with complicated spawning patterns and for collections outside of the preferred spawning period. Future work may include the use of this technology to classify distinct oocyte development stages.

There is a paucity of age data for chondrichthyan fishes owing, in large part, to limitations in traditional age estimation methods. Fourier transform near-infrared (FT-NIR) spectroscopy has shown promise as an alternative, more efficient method for acquiring age data from chondrichthyans. However, studies are limited to sharks in the southern hemisphere. We explored FT-NIR spectroscopy to predict age for a batoid species in the northern hemisphere. The longnose skate (Raja rhina) is one of a small number of batoids for which annual band periodicity in vertebral centra has been validated, allowing for traditional age estimation and making it an ideal candidate for this study. We fit a multivariate partial least-square predictive model between FT-NIR spectra collected from vertebral centra and traditional age estimates, and tested model predictive skill by using external validation. Using FT-NIR spectroscopy, we were able to predict age for longnose skates between the ages of 1 and 14 years with precision and bias near equal to those of traditional methods in less than a quarter of the time. These results support potential for FT-NIR spectroscopy to increase the amount of age data available for assessments used to inform the conservation and management of this sensitive group of species.

Identifying changes in fish growth is important for accurate scientific advice used for fisheries management, because environmental change is affecting fish growth and size-at-age is a critical component of contemporary stock assessment methods. Growth-increment biochronologies are time series of growth-increments derived from hard parts of marine organisms that may reveal dynamics of somatic fish growth. Here we use time series of otolith increments of two fish stocks to fit and compare a biologically derived growth model and a generalized statistical model. Both models produced similar trajectories of annual growth trends, but the biologically based one was more precise and predicted smaller interannual fluctuations than the statistical model. The biologically based model strongly indicated covariance between anabolic and catabolic rates among individuals. Otolith size-at-age did not closely match fish length-at-age, and consequently the growth model could not accurately hindcast observed fish length-at-age. For these reasons, fitted growth dynamics from otolith biochronologies may best suited to identify growth rate fluctuations, understand past drivers of growth dynamics, and improve ecological forecast in the face of rapid environmental change.

High-latitude climate warming is expected to have wide-ranging effects on habitats, ecosystems, and the fish species that occupy them. Not all fish species will be able to adapt to increasing temperatures. We investigated oxygen isotope fractionation in fish otoliths and its relationship to environmental temperature and thermal histories of individual fish. Fish from 4 gadid species, Gadus macrocephalus, Boreogadus saida, Eleginus gracilis, and G. chalcogrammus, representing North Pacific and Arctic regions, were reared in a range of controlled temperatures (0-20 degrees C). We estimated 4 new species-specific otolith oxygen isotope fractionation equations, a relationship between otolith delta O-18 and temperature (T) in the form delta O-18(o) - delta O-18(w) = m x T degrees C + b and also in a second form using the fractionation factor alpha: 1000 ln alpha = a x (1000 TK-1) + c, where o is otolith, w is water, and m, b, a and c are regression coefficients. In using the first form, B. saida was the most unique among the 4 species, with the steepest slope (-0.23) and the highest intercept (32.99 parts per thousand Vienna PeeDee Belemnite [VPDB]). G. macrocephalus had the lowest slope (-0.17) and the lowest intercept (31.76 parts per thousand [VPDB]). Results of an ANCOVA test indicated that the 4 fractionation equations were not statistically different (F = 2.25, p > 0.087). However, when we applied the 4 new fractionation equations to delta O-18(o) measured in wild-caught B. saida otoliths, the species-specific fractionation equation resulted in the closest match between measured and predicted water temperatures. These new fractionation equations represent new tools for investigating temperature effects on fish biota and will also improve paleotemperature reconstruction, especially for high-latitude species.

Many rockfish (genus Sebastes) species within multispecies complexes have lacked basic biological data hindering their fisheries management. In this study, we provide essential information for stock assessment decision making for one of these species, Harlequin Rockfish Sebastes variegatus. We analyzed aspects of reproduction, size structure, and growth of Harlequin Rockfish in waters off Alaska using historical survey data and recent field collections. Results are reported primarily from the Gulf of Alaska (GOA) region, but important findings are noted from the Aleutian Islands (AI). Harlequin Rockfish reached an observed maximum age of 76 years in the AI, a new estimate for this species. Females exhibited group-synchronous oocyte development; the parturition period occurred in the spring in both regions, ceasing as late as July in the GOA. Females from the GOA had an estimated length and age at 50% maturity of 187.6 mm (95% confidence limits [CL] = 152.0-215.6 mm; n = 318) and 4.7 years (95% CL = 1.6-6.2 years; n = 188), respectively, indicating that the Harlequin Rockfish is one of the earliest maturing rockfishes. We analyzed a 14-year time series of bottom trawl survey data to examine the relationship between Harlequin Rockfish length and selected covariates. Spatial heterogeneity existed, with size showing a relationship with bottom temperature and ocean color (productivity index) across the GOA. The comparison of Harlequin Rockfish length at age indicated that larger fish were present in the western GOA but variable growth occurred across the region. Our findings contribute to understanding Harlequin Rockfish biology within multispecies management complexes, and our results demonstrate that key life history traits vary spatially, possibly influenced by regional environmental conditions. This study represents the most comprehensive biological examination for Harlequin Rockfish.

AbstractFor decades, age‐structured stock assessments have been a key component to managing fishery resources worldwide. Fisheries management systems have been under increasing demand to generate a greater volume and quality of age estimates. Traditional aging techniques, which require physical preparation followed by microscopic examination of fish otoliths, are labor‐intensive, expensive, and inherently subjective among individual analysts, making repeatability and precision a challenge. Here we investigated an innovative approach to aging fish from their otoliths using Fourier‐transformed near‐infrared spectroscopy and partial least squares regression models. Models were fit to and validated on spectra and used to microscopically estimate ages of Pacific cod from three years of fishery‐independent otolith data out of the Bering sea. Calibrated and validated models for each year, as well as on an ensemble of the three years, yielded high precision for the multiyear model (R2 = 0.869, RMSE = 0.614, PA = 63%, CV = 7.412), and independent year models (R2 = 0.844–0.891, RMSE = 0.555–0.615, PA = 65%, CV = 6.313–6.775). These metrics of model performance were highly comparable to precision from the traditional microscopic aging approach (R2 = 0.763–0.869, RMSE = 0.639–0.737, PA = 63%–70%, CV = 5.671–6.698). In all cases, a two‐sided Kolmogorov–Smirnov test showed no significant difference between reference and model estimated age distributions. Our results illustrate how Fourier‐transformed near‐infrared spectroscopy can be utilized on otoliths to predict age estimates with substantially greater efficiency, good precision, high repeatability, and no loss in data integrity compared to the traditional microscopic method for aging Pacific cod.

AbstractSmall and isolated peripheral populations, which are often remnants of glacial refugia, offer an opportunity to determine the magnitude and direction of fine‐scale connectivity in high gene flow marine species. When located at the equatorial edge of a species’ range, these populations may also harbor genetic diversity related to survival and reproduction at higher temperatures, a critical resource for marine species facing warming ocean temperatures. Pacific cod (Gadus macrocephalus), a marine fish in the North Pacific, has already experienced major shifts in biomass and distribution linked to climate change. We estimated the magnitude and direction of connectivity between peripheral populations of Pacific cod at the southern edge of the species’ range, by conducting restriction site‐associated DNA (RAD) sequencing and individual assignment on fish collected around the Korean Peninsula during the spawning season. Three populations on the western, eastern, and southern Korean coasts were highly differentiated (FST = 0.025–0.042) and relatively small (Ne = 433–1,777). Ten putative dispersers and estimates of contemporary migration rates revealed asymmetrical, west‐to‐east movement around the Korean Peninsula, at a higher rate than predicted by indirect estimates of connectivity (FST). Allele frequencies at 87 RAD loci were decisively correlated with strong marine temperature gradients between the warmer southern coast and the cooler waters of the eastern and western coasts. Despite relatively small sample sizes, our data suggest asymmetrical dispersal and gene flow, potentially involving adaptive alleles, between peripheral populations inhabiting markedly different thermal regimes. Our study emphasizes the conservation value of peripheral populations in high gene flow marine fish species.

Measuring fish population responses to climate change requires timely ecological information, warranting innovative approaches to data collection in fisheries research and management. Fourier transform near-infrared (FT-NIR) spectroscopy is a promising tool to efficiently and cost-effectively obtain multiple types of fisheries data including fish physiological health and energetics that can provide indicators of stock status and environmental change. We tested the applicability of FT-NIR spectroscopy to determine fish physiological state and condition by developing calibration models for morphometric indices of body condition [Fulton’s K and hepatosomatic index (HSI)], biochemical measurements of tissue composition (lipid content and energy density), and a nucleic acid-based index of recent growth (RNA:DNA) of juvenile Pacific cod (Gadus macrocephalus). Calibration models had the best predictive ability for lipid content followed by HSI and energy density, whereas spectral data had weak relationships with Fulton’s K and RNA:DNA. For lipid content, energy density, and HSI, informative spectral regions were primarily associated with carbon-hydrogen bonds in lipid molecules. Additionally, FT-NIR spectroscopy calibration models better predicted lipid content than morphometric measurements that are often used as proxies for measuring energy reserves, indicating that FT-NIR spectroscopy might serve as a more informative index of body condition and energy stores than other rapid methods. Efficient sample analysis by FT-NIR spectroscopy can supplement traditional metrics of body condition and be especially useful for ensuring the availability of monitoring data in support of fisheries research and management.

Bomb-produced radiocarbon (14C) was used to validate age estimates of Greenland halibut (Reinhardtius hippoglossoides) using a stained otolith cross-section method. The Δ14C in eastern Bering Sea (EBS) Greenland halibut otoliths was compared to both EBS and Gulf of Alaska (GOA) Pacific halibut (Hippoglossoides stenolepis) otolith reference chronologies to evaluate which reference chronology was most suitable, and to quantitatively estimate age determination bias. Using Bayesian analysis and a coupled-function model, the Δ14C in the Greenland halibut showed greatest similarities to the Δ14C in the GOA reference chronology. Although the model indicated under ageing, the bias was not large. Assigning an age less than the true age by more than a one year is about 73%, and less than the true age by more than 2 years is only about 25%. When considering the age at which Greenland halibut is only 7.5% of its maximum longevity (50+ years) and that the probability of underageing by 3 years being less than 5%, it is likely that between-age-reader variation will cancel out any systematic bias that exists in the age determination protocols. Prior to the use of stained cross-sections the maximum age was 38 years, now a maximum age of 53 years is supported.

Recent application of Fourier transform near infra-red spectroscopy (FT-NIRS) to predict age in fish otoliths has gained attention among fisheries managers as a potential alternative to costly production ageing of managed species. We assessed the age prediction capability of FT-NIRS scans in whole otoliths from red snapper, Lutjanus campechanus, collected from the US Gulf of Mexico and US Atlantic Ocean (South Atlantic). Otoliths were scanned with an FT-NIR spectrometer and resulting spectral signatures were regressed with traditionally estimated ages via partial least squares regression to produce calibration models, which were validated for predictive capability against test sets of otoliths. Calibration models successfully predicted age with R-2 ranging 0.94-0.95, mean squared error <= 1.8 years, and bias <0.02 years. Percent agreement between FT-NIRS and traditional ages was lower than within-reader agreement for traditional estimates, but average percent error was similar and Kolmogorov-Smirnov tests were not significantly different (p >= 0.06) between traditional and FT-NIRS predicted ages for optimal calibration models. Ages >31 years were not well predicted, possibly due to light attenuation in the thickest otoliths. Our results suggest that FT-NIRS can improve efficiency in production ageing for fisheries management while maintaining data quality standards.

In rockfish (Family Scorpaenidae), age determination is difficult and the annual nature of otolith growth zones must be validated independently. We applied routine age determination to four species of Gulf of Alaska rockfish: two shallower-water species, namely harlequin rockfish (Sebastes variegatus) and redstripe rockfish (Sebastes proriger), and two deep-water species, namely shortspine thornyhead (Sebastolobus alascanus) and shortraker rockfish (Sebastes borealis). The estimated ages (counts of presumed annual growth zones in the otoliths) were then evaluated with bomb-produced radiocarbon (14C) and Bayesian modelling with Markov chain Monte Carlo simulations. This study successfully demonstrated the level of accuracy in estimated ages of redstripe rockfish (a 35% probability of underageing, and ~5% probability of overageing) and harlequin rockfish (a 100% probability that they were underaged by ~3 or 4 years). Measured Δ14C in shortspine thornyhead and shortraker rockfish otoliths was lower and increased later than expected. Hence, incorrect age determination could not be evaluated. This is likely caused by dissimilar environmental and biological availability of 14C between these two species and the Pacific halibut (Hippoglossus stenolepis) reference chronology, or underageing of these two species.

Sustainable management of fishery resources is predicated on a foundational understanding of the biogeography of fish stocks and the delineation of stocks into appropriate management units. Despite notable fluctuations in the catch of commercially valuable Pacific cod (Gadus macrocephalus) in South Korea, relatively little is known regarding the stock structure and migratory trends of this species in the region. Here, otolith microchemistry was used to evaluate the stock structure and ontogenetic migratory trends of Pacific cod from five spawning grounds around the Korean Peninsula. Statistically significant between-region discrimination was evident and resolved Pacific cod around the Korean Peninsula into two distinct stocks. Specimens were classified to region of capture using quadratic discriminant analysis of age-0 and capture elemental signatures with overall accuracies of 71.12 and 79.1% respectively, lending support to the notion that Pacific cod demonstrate natal philopatry. Analysis of the elemental signatures over the first 2 years of life exhibited clear trends indicative of shifts in habitat use, suggesting that trace elements function well as indicators of ontogenetic migration. Our work provides information that is complementary to other direct and indirect methods of monitoring migratory trends and delineating stocks, which are integral components of effective species and ecosystem management plans.